[XPU] Support V1 Loader in Bf16 (#4746)

* add v1 support for bf16

* update

* update

* update

* update

* update

* update code

.clang-format

@@ -16,7 +16,7 @@
 ---
 Language:        Cpp
 BasedOnStyle:  Google
-IndentWidth:     4
+IndentWidth:     2
 TabWidth:        2
 ContinuationIndentWidth: 4
 AccessModifierOffset: -1  # The private/protected/public has no indent in class

.github/actions/rerun-workflow/action.yml

@@ -0,0 +1,30 @@
+name: 'Rerun Workflow'
+description: 'Re-run GitHub Actions workflow for a given Pull Request'
+inputs:
+  GITHUB_TOKEN:
+    description: 'GitHub token with repo scope'
+    required: true
+  OWNER:
+    description: 'Repository owner'
+    required: true
+  REPO:
+    description: 'Repository name'
+    required: true
+  PR_ID:
+    description: 'Pull Request ID'
+    required: true
+  JOB_NAME:
+    description: 'Job name to rerun'
+    required: true
+
+runs:
+  using: 'composite'
+  steps:
+    - run: bash ./.github/actions/rerun-workflow/rerun.sh
+      shell: bash
+      env:
+        GITHUB_TOKEN: ${{ inputs.GITHUB_TOKEN }}
+        OWNER: ${{ inputs.OWNER }}
+        REPO: ${{ inputs.REPO }}
+        PR_ID: ${{ inputs.PR_ID }}
+        JOB_NAME: ${{ inputs.JOB_NAME }}

.github/actions/rerun-workflow/rerun.sh

@@ -0,0 +1,77 @@
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+set -e
+
+COMMIT_SHA=$(curl -s -H "Authorization: token $GITHUB_TOKEN" \
+  "https://api.github.com/repos/$OWNER/$REPO/pulls/$PR_ID" | jq -r '.head.sha')
+
+echo "Commit SHA: $COMMIT_SHA"
+
+response=$(curl -s -H "Authorization: token $GITHUB_TOKEN" \
+  "https://api.github.com/repos/$OWNER/$REPO/actions/runs?head_sha=$COMMIT_SHA&per_page=100")
+
+echo "Response: $response"
+
+run_ids=$(echo "$response" | jq -r '.workflow_runs[].id')
+
+if [ -n "$run_ids" ]; then
+  echo "Found run_ids for commit $COMMIT_SHA: $run_ids"
+
+  for run_id in $run_ids; do
+    if [ "$JOB_NAME" = "all-failed" ]; then
+      echo "Rerunning all failed jobs for run_id: $run_id"
+
+      rerun_response=$(curl -X POST -s -w "%{http_code}" -o /dev/null \
+        -H "Accept: application/vnd.github.v3+json" \
+        -H "Authorization: Bearer $GITHUB_TOKEN" \
+        "https://api.github.com/repos/$OWNER/$REPO/actions/runs/$run_id/rerun-failed-jobs")
+      if [ "$rerun_response" -eq 201 ]; then
+        echo "Successfully requested rerun for all blocked jobs in run_id: $run_id"
+      else
+        echo "Failed to request rerun for run_id: $run_id with status code $rerun_response"
+      fi
+
+    else
+      jobs_response=$(curl -s -H "Authorization: token $GITHUB_TOKEN" \
+      "https://api.github.com/repos/$OWNER/$REPO/actions/runs/$run_id/jobs")
+
+      echo "Jobs Response for run_id $run_id: $jobs_response"
+
+      # if [[ "$JOB_NAME" == *"bypass"* ]]; then
+        block_jobs=$(echo "$jobs_response" | jq -r --arg job_name "$JOB_NAME" \
+        '.jobs[] | select(.name == $job_name) | .id')
+      # else
+      #   block_jobs=$(echo "$jobs_response" | jq -r --arg job_name "$JOB_NAME" \
+      #   '.jobs[] | select(.name == $job_name and .conclusion != "success") | .id')
+      # fi
+
+      if [ -n "$block_jobs" ]; then
+        echo "Found block jobs for run_id $run_id: $block_jobs"
+
+        for job_id in $block_jobs; do
+          echo "Rerunning job_id: $job_id"
+          curl -X POST -H "Accept: application/vnd.github.v3+json" \
+            -H "Authorization: token $GITHUB_TOKEN" \
+            "https://api.github.com/repos/$OWNER/$REPO/actions/jobs/$job_id/rerun"
+        done
+      else
+        echo "No block jobs found for run_id $run_id with name $JOB_NAME."
+      fi
+    fi
+  done
+else
+  echo "No matching workflow runs found for commit $COMMIT_SHA."
+  exit 1
+fi

.github/pull_request_template.md

@@ -0,0 +1,30 @@
+<!-- TemplateReference: https://github.com/PaddlePaddle/FastDeploy/blob/develop/.github/pull_request_template.md -->
+
+<!-- Thank you for your contribution! Please follow these guidelines to enhance your pull request. If anything is unclear, submit your PR and reach out to maintainers for assistance. -->
+
+## Motivation
+
+<!-- Describe the purpose and goals of this pull request. -->
+
+## Modifications
+
+<!-- Detail the changes made in this pull request. -->
+
+## Usage or Command
+
+<!-- You should provide the usage if this pr is about the new function. -->
+<!-- You should provide the command to run if this pr is about the performance optimization or fixing bug. -->
+
+## Accuracy Tests
+
+<!-- If this pull request affects model outputs (e.g., changes to the kernel or model forward code), provide accuracy test results. -->
+
+## Checklist
+
+- [ ] Add at least a tag in the PR title.
+  - Tag list: [`[FDConfig]`,`[APIServer]`,`[Engine]`, `[Scheduler]`, `[PD Disaggregation]`, `[Executor]`, `[Graph Optimization]`, `[Speculative Decoding]`, `[RL]`, `[Models]`, `[Quantization]`, `[Loader]`, `[OP]`, `[KVCache]`, `[DataProcessor]`, `[BugFix]`, `[Docs]`, `[CI]`, `[Optimization]`, `[Feature]`, `[Benchmark]`, `[Others]`, `[XPU]`, `[HPU]`, `[GCU]`, `[DCU]`, `[Iluvatar]`, `[Metax]`]
+  - You can add new tags based on the PR content, but the semantics must be clear.
+- [ ] Format your code, run `pre-commit` before commit.
+- [ ] Add unit tests. Please write the reason in this PR if no unit tests.
+- [ ] Provide accuracy results.
+- [ ] If the current PR is submitting to the `release` branch, make sure the PR has been submitted to the `develop` branch, then cherry-pick it to the `release` branch with the `[Cherry-Pick]` PR tag.

.github/workflows/_accuracy_test.yml

@@ -44,7 +44,7 @@ jobs:
             FULL_REPO="${{ github.repository }}"
             REPO_NAME="${FULL_REPO##*/}"
             BASE_BRANCH="${{ github.base_ref }}"
-
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -80,12 +80,14 @@ jobs:
           FD_API_PORT=$((42088 + DEVICE_PORT * 100))
           FD_ENGINE_QUEUE_PORT=$((42058 + DEVICE_PORT * 100))
           FD_METRICS_PORT=$((42078 + DEVICE_PORT * 100))
+          FD_CACHE_QUEUE_PORT=$((42098 + DEVICE_PORT * 100))
           echo "Test ENV Parameter:"
           echo "========================================================="
           echo "FLASK_PORT=${FLASK_PORT}"
           echo "FD_API_PORT=${FD_API_PORT}"
           echo "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}"
           echo "FD_METRICS_PORT=${FD_METRICS_PORT}"
+          echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
           echo "DEVICES=${DEVICES}"
           echo "========================================================="
 
@@ -99,7 +101,7 @@ jobs:
             exit 1
           fi
 
-          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT)
+          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT $FD_CACHE_QUEUE_PORT)
           LOG_FILE="./port_cleanup_$(date +%Y%m%d_%H%M%S).log"
           echo "==== LOG_FILE is ${LOG_FILE} ===="
 
@@ -133,6 +135,7 @@ jobs:
           -e "FD_API_PORT=${FD_API_PORT}" \
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
+          -e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
           -e "FLASK_PORT=${FLASK_PORT}" \
           -v "${MODEL_CACHE_DIR}:/MODELDATA" \
           -v "${CACHE_DIR}/gitconfig:/etc/gitconfig:ro" \
@@ -152,11 +155,13 @@ jobs:
           ./llm-deploy-linux-amd64 -python python3.10 \
           -model_name ERNIE-4.5-0.3B-Paddle \
           -model_path /MODELDATA \
-          --skip install
+          --skip install,model
 
           git config --global --add safe.directory /workspace/FastDeploy
           cd FastDeploy
           pushd tests/ce/deploy
+          ps -ef | grep "${FD_CACHE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
+          ps -ef | grep "${FD_ENGINE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
           python3.10 deploy.py > dd.log 2>&1 &
           sleep 3
           curl -X POST http://0.0.0.0:${FLASK_PORT}/start \

.github/workflows/_base_test.yml

@@ -44,7 +44,7 @@ jobs:
             FULL_REPO="${{ github.repository }}"
             REPO_NAME="${FULL_REPO##*/}"
             BASE_BRANCH="${{ github.base_ref }}"
-
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -80,12 +80,14 @@ jobs:
           FD_API_PORT=$((42088 + DEVICE_PORT * 100))
           FD_ENGINE_QUEUE_PORT=$((42058 + DEVICE_PORT * 100))
           FD_METRICS_PORT=$((42078 + DEVICE_PORT * 100))
+          FD_CACHE_QUEUE_PORT=$((42098 + DEVICE_PORT * 100))
           echo "Test ENV Parameter:"
           echo "========================================================="
           echo "FLASK_PORT=${FLASK_PORT}"
           echo "FD_API_PORT=${FD_API_PORT}"
           echo "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}"
           echo "FD_METRICS_PORT=${FD_METRICS_PORT}"
+          echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
           echo "DEVICES=${DEVICES}"
           echo "========================================================="
 
@@ -99,7 +101,7 @@ jobs:
             exit 1
           fi
 
-          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT)
+          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT $FD_CACHE_QUEUE_PORT)
           LOG_FILE="./port_cleanup_$(date +%Y%m%d_%H%M%S).log"
           echo "==== LOG_FILE is ${LOG_FILE} ===="
 
@@ -134,7 +136,7 @@ jobs:
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
           -e "FLASK_PORT=${FLASK_PORT}" \
-          -e "FD_FORCE_CHUNKED_PREFILL=1" \
+          -e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
           -v "${MODEL_CACHE_DIR}:/MODELDATA" \
           -v "${CACHE_DIR}/gitconfig:/etc/gitconfig:ro" \
           -v "${CACHE_DIR}/.cache:/root/.cache" \
@@ -153,11 +155,13 @@ jobs:
           ./llm-deploy-linux-amd64 -python python3.10 \
           -model_name ERNIE-4.5-0.3B-Paddle \
           -model_path /MODELDATA \
-          --skip install
+          --skip install,model
 
           git config --global --add safe.directory /workspace/FastDeploy
           cd FastDeploy
           pushd tests/ce/deploy
+          ps -ef | grep "${FD_CACHE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
+          ps -ef | grep "${FD_ENGINE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
           python3.10 deploy.py > dd.log 2>&1 &
           sleep 3
           curl -X POST http://0.0.0.0:${FLASK_PORT}/start \

.github/workflows/_build_linux.yml

@@ -55,7 +55,7 @@ on:
 jobs:
   fd-build:
     runs-on: [self-hosted, GPU-Build]
-    timeout-minutes: 240
+    timeout-minutes: 360
     outputs:
       wheel_path: ${{ steps.set_output.outputs.wheel_path }}
     steps:
@@ -82,7 +82,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -106,7 +106,12 @@ jobs:
             CARD_ID=$(echo "${runner_name}" | awk -F'-' '{print $NF}')
             gpu_id=$(echo "$CARD_ID" | fold -w1 | paste -sd,)
 
-            CACHE_DIR="${CACHE_DIR:-$(dirname "$(dirname "${{ github.workspace }}")")}"
+            IFS='/' read -ra parts <<< "${GITHUB_WORKSPACE}"
+            len=${#parts[@]}
+            CCACHE_DEFAULT_DIR="/$(IFS=/; echo "${parts[*]:1:$((len-5))}")"
+            echo "$CCACHE_DEFAULT_DIR"
+
+            CACHE_DIR="${CACHE_DIR:-$CCACHE_DEFAULT_DIR}"
             echo "CACHE_DIR is set to ${CACHE_DIR}"
             if [ ! -f "${CACHE_DIR}/gitconfig" ]; then
               touch "${CACHE_DIR}/gitconfig"
@@ -127,6 +132,7 @@ jobs:
             -e "PADDLEVERSION=${PADDLEVERSION}" \
             -e "PADDLE_WHL_URL=${PADDLE_WHL_URL}" \
             -e "BRANCH_REF=${BRANCH_REF}" \
+            -e "CCACHE_MAXSIZE=50G" \
             --gpus "\"device=${gpu_id}\"" ${docker_image} /bin/bash -c '
             if [[ -n "${FD_VERSION}" ]]; then
               export FASTDEPLOY_VERSION=${FD_VERSION}
@@ -134,6 +140,7 @@ jobs:
             fi
 
             git config --global --add safe.directory /workspace/FastDeploy
+            chown -R $(whoami) /workspace/FastDeploy
             cd FastDeploy
             if [[ "${WITH_NIGHTLY_BUILD}" == "ON" ]];then
               GIT_COMMIT_TIME=$(git --no-pager show -s --format=%ci HEAD)

.github/workflows/_ci_gcu.yml

@@ -1,10 +1,10 @@
 name: CI_GCU
 
 on:
-  pull_request:
-    branches:
-      - develop
-      - 'release/*'
+  #pull_request:
+    #branches:
+      #- develop
+      #- 'release/*'
   workflow_dispatch:
 
 concurrency:

.github/workflows/_ci_image_build.yml

@@ -0,0 +1,73 @@
+name: Docker Build
+description: "FastDeploy CI Image Build"
+
+on:
+  workflow_call:
+    inputs:
+      CI_DOCKER_IMAGE_NAME:
+        description: "Build Images"
+        required: true
+        type: string
+        default: "ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:cuda126-py310"
+      FASTDEPLOY_ARCHIVE_URL:
+        description: "URL of the compressed FastDeploy code archive."
+        required: true
+        type: string
+      DOCKER_IMAGE_NAME:
+        description: "Build Images"
+        required: false
+        type: string
+        default: "ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate"
+    outputs:
+      docker_name_precheck:
+        description: "Output path of the generated wheel"
+        value: ${{ jobs.docker_build.outputs.docker_name_precheck }}
+
+jobs:
+  docker_build:
+    runs-on: [self-hosted, Docker-Build]
+    outputs:
+      docker_name_precheck: ${{ steps.docker_build.outputs.docker_name_precheck }}
+    steps:
+      - name: Docker Build
+        id: docker_build
+        shell: bash
+        env:
+          docker_image_name: ${{ inputs.CI_DOCKER_IMAGE_NAME }}
+          docker_image: ${{ inputs.DOCKER_IMAGE_NAME }}
+          fd_archive_url: ${{ inputs.FASTDEPLOY_ARCHIVE_URL }}
+        run: |
+            set -x
+            REPO="https://github.com/${{ github.repository }}.git"
+            FULL_REPO="${{ github.repository }}"
+            REPO_NAME="${FULL_REPO##*/}"
+            BASE_BRANCH="${{ github.base_ref }}"
+
+            # Clean the repository directory before starting
+            docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
+            -e "REPO_NAME=${REPO_NAME}" \
+            ${docker_image} /bin/bash -c '
+              if [ -d ${REPO_NAME} ]; then
+                echo "Directory ${REPO_NAME} exists, removing it..."
+                rm -rf ${REPO_NAME}*
+              fi
+            '
+
+            wget -q --no-proxy ${fd_archive_url}
+            tar -xf FastDeploy.tar.gz
+            rm -rf FastDeploy.tar.gz
+            cd FastDeploy
+            git config --global user.name "FastDeployCI"
+            git config --global user.email "fastdeploy_ci@example.com"
+            git log -n 3 --oneline
+
+            # Docker Build
+            cd tools/dockerfile/
+            set -e
+            cp ../../requirements.txt ./
+            cp ../../scripts/unittest_requirement.txt ./
+            docker build -t ${docker_image_name} -f Dockerfile.ci . \
+                --network host \
+                --no-cache
+            docker push ${docker_image_name}
+            echo "docker_name_precheck=${docker_image_name}" >> $GITHUB_OUTPUT

.github/workflows/_logprob_test_linux.yml

@@ -32,6 +32,7 @@ on:
 jobs:
   run_tests_logprob:
     runs-on: [self-hosted, GPU-h20-1Cards]
+    timeout-minutes: 60
     steps:
       - name: Code Prepare
         shell: bash
@@ -39,6 +40,7 @@ jobs:
           docker_image: ${{ inputs.DOCKER_IMAGE }}
           paddletest_archive_url: ${{ inputs.PADDLETEST_ARCHIVE_URL }}
         run: |
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -46,7 +48,7 @@ jobs:
             ${docker_image} /bin/bash -c '
             rm -rf /workspace/*
             '
-            wget -q ${paddletest_archive_url}
+            wget -q --no-proxy ${paddletest_archive_url}
             tar -xf PaddleTest.tar.gz
             rm -rf PaddleTest.tar.gz
             cd PaddleTest
@@ -70,12 +72,14 @@ jobs:
           FD_API_PORT=$((42088 + DEVICE_PORT * 100))
           FD_ENGINE_QUEUE_PORT=$((42058 + DEVICE_PORT * 100))
           FD_METRICS_PORT=$((42078 + DEVICE_PORT * 100))
+          FD_CACHE_QUEUE_PORT=$((42098 + DEVICE_PORT * 100))
           echo "Test ENV Parameter:"
           echo "========================================================="
           echo "FLASK_PORT=${FLASK_PORT}"
           echo "FD_API_PORT=${FD_API_PORT}"
           echo "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}"
           echo "FD_METRICS_PORT=${FD_METRICS_PORT}"
+          echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
           echo "DEVICES=${DEVICES}"
           echo "========================================================="
 
@@ -89,7 +93,7 @@ jobs:
             exit 1
           fi
 
-          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT)
+          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT $FD_CACHE_QUEUE_PORT)
           LOG_FILE="./port_cleanup_$(date +%Y%m%d_%H%M%S).log"
           echo "==== LOG_FILE is ${LOG_FILE} ===="
 
@@ -114,7 +118,6 @@ jobs:
             echo "Removing stale container: ${runner_name}"
             docker rm -f ${runner_name} || true
           fi
-
           docker run --rm --ipc=host --pid=host --net=host \
           --name ${runner_name} \
           -v $(pwd):/workspace \
@@ -123,6 +126,7 @@ jobs:
           -e "FD_API_PORT=${FD_API_PORT}" \
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
+          -e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
           -e "FLASK_PORT=${FLASK_PORT}" \
           -v "${MODEL_CACHE_DIR}:/MODELDATA" \
           -v "${CACHE_DIR}/gitconfig:/etc/gitconfig:ro" \
@@ -141,9 +145,11 @@ jobs:
           ./llm-deploy-linux-amd64 -python python3.10 \
           -model_name ERNIE-4.5-0.3B-Paddle \
           -model_path /MODELDATA \
-          --skip install
+          --skip install,model
 
           cd PaddleTest/framework/ServeTest
+          ps -ef | grep "${FD_CACHE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
+          ps -ef | grep "${FD_ENGINE_QUEUE_PORT}" | grep -v grep | awk "{print \$2}" | xargs -r kill -9
           python3.10 deploy.py > dd.log 2>&1 &
           sleep 3
           curl -X POST http://0.0.0.0:${FLASK_PORT}/start \

.github/workflows/_pre_ce_test.yml

@@ -46,7 +46,7 @@ jobs:
             FULL_REPO="${{ github.repository }}"
             REPO_NAME="${FULL_REPO##*/}"
             BASE_BRANCH="${{ github.base_ref }}"
-
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -57,7 +57,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -81,12 +81,17 @@ jobs:
           FD_API_PORT=$((42088 + DEVICE_PORT * 100))
           FD_ENGINE_QUEUE_PORT=$((42058 + DEVICE_PORT * 100))
           FD_METRICS_PORT=$((42078 + DEVICE_PORT * 100))
+          FD_CACHE_QUEUE_PORT=$((42098 + DEVICE_PORT * 100))
+          FD_ZMQ_RECV_REQUEST_SERVER_PORT=$((42048 + DEVICE_PORT * 100))
+          FD_ZMQ_SEND_RESPONSE_SERVER_PORT=$((42038 + DEVICE_PORT * 100))
+          FD_ZMQ_CONTROL_CMD_SERVER_PORTS=$((42028 + DEVICE_PORT * 100))
           echo "Test ENV Parameter:"
           echo "========================================================="
           echo "FLASK_PORT=${FLASK_PORT}"
           echo "FD_API_PORT=${FD_API_PORT}"
           echo "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}"
           echo "FD_METRICS_PORT=${FD_METRICS_PORT}"
+          echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
           echo "DEVICES=${DEVICES}"
           echo "========================================================="
 
@@ -96,7 +101,7 @@ jobs:
             touch "${CACHE_DIR}/gitconfig"
           fi
 
-          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT)
+          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT $FD_CACHE_QUEUE_PORT)
           LOG_FILE="./port_cleanup_$(date +%Y%m%d_%H%M%S).log"
           echo "==== LOG_FILE is ${LOG_FILE} ===="
 
@@ -134,6 +139,7 @@ jobs:
           -e "FD_API_PORT=${FD_API_PORT}" \
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
+          -e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
           -e "FLASK_PORT=${FLASK_PORT}" \
           -e "fd_wheel_url=${fd_wheel_url}" \
           --gpus "\"device=${DEVICES}\"" ${docker_image} /bin/bash -c '

.github/workflows/_stable_test.yml

@@ -44,7 +44,7 @@ jobs:
             FULL_REPO="${{ github.repository }}"
             REPO_NAME="${FULL_REPO##*/}"
             BASE_BRANCH="${{ github.base_ref }}"
-
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -55,7 +55,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy

.github/workflows/_unit_test_coverage.yml

@@ -41,7 +41,7 @@ jobs:
 
   run_tests_with_coverage:
     runs-on: [self-hosted, GPU-h1z1-2Cards]
-    timeout-minutes: 60
+    timeout-minutes: 90
     needs: check_cov_skip
     if: needs.check_cov_skip.outputs.can-skip != 'true'
     outputs:
@@ -60,7 +60,7 @@ jobs:
             FULL_REPO="${{ github.repository }}"
             REPO_NAME="${FULL_REPO##*/}"
             BASE_BRANCH="${{ github.base_ref }}"
-
+            docker pull ${docker_image}
             # Clean the repository directory before starting
             docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
             -e "REPO_NAME=${REPO_NAME}" \
@@ -71,7 +71,7 @@ jobs:
               fi
             '
 
-            wget -q ${fd_archive_url}
+            wget -q --no-proxy ${fd_archive_url}
             tar -xf FastDeploy.tar.gz
             rm -rf FastDeploy.tar.gz
             cd FastDeploy
@@ -102,12 +102,14 @@ jobs:
           FD_API_PORT=$((42088 + DEVICE_PORT * 100))
           FD_ENGINE_QUEUE_PORT=$((42058 + DEVICE_PORT * 100))
           FD_METRICS_PORT=$((42078 + DEVICE_PORT * 100))
+          FD_CACHE_QUEUE_PORT=$((42098 + DEVICE_PORT * 100))
           echo "Test ENV Parameter:"
           echo "========================================================="
           echo "FLASK_PORT=${FLASK_PORT}"
           echo "FD_API_PORT=${FD_API_PORT}"
           echo "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}"
           echo "FD_METRICS_PORT=${FD_METRICS_PORT}"
+          echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
           echo "DEVICES=${DEVICES}"
           echo "========================================================="
 
@@ -117,7 +119,7 @@ jobs:
             touch "${CACHE_DIR}/gitconfig"
           fi
 
-          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT)
+          PORTS=($FLASK_PORT $FD_API_PORT $FD_ENGINE_QUEUE_PORT $FD_METRICS_PORT $FD_CACHE_QUEUE_PORT)
           LOG_FILE="./port_cleanup_$(date +%Y%m%d_%H%M%S).log"
           echo "==== LOG_FILE is ${LOG_FILE} ===="
 
@@ -156,6 +158,7 @@ jobs:
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \
           -e "FLASK_PORT=${FLASK_PORT}" \
+          -e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
           -e TZ="Asia/Shanghai" \
           -e "fd_wheel_url=${fd_wheel_url}" \
           -e "BASE_REF=${BASE_REF}" \
@@ -168,10 +171,7 @@ jobs:
           python -m pip install --pre paddlepaddle-gpu -i https://www.paddlepaddle.org.cn/packages/nightly/cu126/
           pip config set global.extra-index-url https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple
 
-          python -m pip install coverage
-          python -m pip install diff-cover
-          python -m pip install pytest-cov
-          python -m pip install jsonschema aistudio_sdk==0.3.5
+          python -m pip install -r scripts/unittest_requirement.txt
           python -m pip install ${fd_wheel_url}
           rm -rf fastdeploy
           # coverage subprocess use
@@ -194,51 +194,104 @@ jobs:
           coverage xml -o python_coverage_all.xml
           COVERAGE_EXIT_CODE=0
           if [[ "$IS_PR" == "true" ]]; then
+            echo "Running diff coverage for PR..."
             diff-cover python_coverage_all.xml --diff-file=diff.txt --fail-under=80 --json-report diff_coverage.json || COVERAGE_EXIT_CODE=9
             python scripts/generate_diff_coverage_xml.py diff.txt python_coverage_all.xml
           else
-            echo "Not a PR, skipping diff-cover"
+            echo "Running full coverage"
+            coverage report -m > full_coverage_report.txt
+            python scripts/generate_full_coverage_csv.py full_coverage_report.txt full_coverage_report.csv
           fi
           echo "COVERAGE_EXIT_CODE=${COVERAGE_EXIT_CODE}" >> exit_code.env
           '
           if [ -f FastDeploy/exit_code.env ]; then
             cat FastDeploy/exit_code.env >> $GITHUB_ENV
           fi
-      - name: Upload unit resule and diff coverage to bos
+      - name: Upload coverage and unit test results to BOS
         id: cov_upload
         shell: bash
+        env:
+          IS_PR: ${{ github.event_name == 'pull_request' }}
+          GITHUB_SHA: ${{ github.sha }}
+          BRANCH: ${{ github.ref_name }}
+          PR_COMMIT_SHA: ${{ github.event.pull_request.head.sha }}
+          PR_NUMBER: ${{ github.event.pull_request.number }}
         run: |
           cd FastDeploy
-          commit_id=${{ github.event.pull_request.head.sha }}
-          pr_num=${{ github.event.pull_request.number }}
-          target_path=paddle-github-action/PR/FastDeploy/${pr_num}/${commit_id}/SM${compile_arch//,/_}
-          wget  -q --no-proxy --no-check-certificate https://paddle-qa.bj.bcebos.com/CodeSync/develop/PaddlePaddle/PaddleTest/tools/bos_tools.py -O bos_tools.py
+          python -m pip install -q bce-python-sdk==0.9.29
+          wget -q --no-proxy --no-check-certificate \
+            https://paddle-qa.bj.bcebos.com/CodeSync/develop/PaddlePaddle/PaddleTest/tools/bos_tools.py \
+            -O bos_tools.py
           push_file=$(realpath bos_tools.py)
-          python -m pip install bce-python-sdk==0.9.29
-          diff_cov_file="diff_coverage.xml"
-          if [ -f ${diff_cov_file} ];then
-            python ${push_file} ${diff_cov_file} ${target_path}/CoverageData
+
+          if [[ "$IS_PR" == "true" ]]; then
+            commit_id=${PR_COMMIT_SHA}
+            pr_num=${PR_NUMBER}
+            target_path=paddle-github-action/PR/FastDeploy/${pr_num}/${commit_id}/SM${compile_arch//,/_}
+          elif [[ "${{ github.ref_type }}" == "tag" ]]; then
+            commit_id=${{ github.sha }}
+            tag_name=${{ github.ref_name }}
+            target_path=paddle-github-action/TAG/FastDeploy/${tag_name}/${commit_id}/SM${compile_arch//,/_}
+            target_path_latest=paddle-github-action/TAG/FastDeploy/${tag_name}/latest/SM${compile_arch//,/_}
+            target_path_stripped_latest="${target_path_latest#paddle-github-action/}"
+          else
+            commit_id=${{ github.sha }}
+            branch_name=${{ github.ref_name }}
+            target_path=paddle-github-action/BRANCH/FastDeploy/${branch_name}/${commit_id}/SM${compile_arch//,/_}
+            target_path_latest=paddle-github-action/BRANCH/FastDeploy/${branch_name}/latest/SM${compile_arch//,/_}
+            target_path_stripped_latest="${target_path_latest#paddle-github-action/}"
+          fi
+
           target_path_stripped="${target_path#paddle-github-action/}"
+
+          if [[ "$IS_PR" == "true" ]]; then
+            diff_cov_file="diff_coverage.xml"
+            if [ -f ${diff_cov_file} ]; then
+              python ${push_file} ${diff_cov_file} ${target_path}/CoverageData
               DIFF_COV_FILE_URL=https://paddle-github-action.bj.bcebos.com/${target_path_stripped}/CoverageData/${diff_cov_file}
               echo "diff_cov_file_url=${DIFF_COV_FILE_URL}" >> $GITHUB_OUTPUT
               echo "diff_cov_file_url=${DIFF_COV_FILE_URL}" >> $GITHUB_ENV
             fi
+
             diff_cov_result_json="diff_coverage.json"
-          if [ -f ${diff_cov_result_json} ];then
+            if [ -f ${diff_cov_result_json} ]; then
               python ${push_file} ${diff_cov_result_json} ${target_path}/CoverageData
-            target_path_stripped="${target_path#paddle-github-action/}"
               DIFF_COV_JSON_URL=https://paddle-github-action.bj.bcebos.com/${target_path_stripped}/CoverageData/${diff_cov_result_json}
               echo "diff_cov_result_json_url=${DIFF_COV_JSON_URL}" >> $GITHUB_OUTPUT
               echo "diff_cov_result_json_url=${DIFF_COV_JSON_URL}" >> $GITHUB_ENV
             fi
+          fi
+
+          HAS_FAILED_TESTS=false
           unittest_result="failed_tests.log"
-          if [ -s ${unittest_result} ];then
+          if [ -s ${unittest_result} ]; then
+            HAS_FAILED_TESTS=true
             python ${push_file} ${unittest_result} ${target_path}/UnitTestResult
-            target_path_stripped="${target_path#paddle-github-action/}"
             UNIT_TEST_RESULT_URL=https://paddle-github-action.bj.bcebos.com/${target_path_stripped}/UnitTestResult/${unittest_result}
             echo "unittest_failed_url=${UNIT_TEST_RESULT_URL}" >> $GITHUB_OUTPUT
             echo "unittest_failed_url=${UNIT_TEST_RESULT_URL}" >> $GITHUB_ENV
           fi
+
+          if [[ "$IS_PR" != "true" ]]; then
+            full_cov_file="full_coverage_report.txt"
+            full_cov_csv="full_coverage_report.csv"
+
+            if [ -f ${full_cov_file} ]; then
+              python ${push_file} ${full_cov_file} ${target_path}/CoverageData
+              python ${push_file} ${full_cov_file} ${target_path_latest}/CoverageData
+              FULL_COV_FILE_URL=https://paddle-github-action.bj.bcebos.com/${target_path_stripped}/CoverageData/${full_cov_file}
+              echo "full_coverage_report_url=${FULL_COV_FILE_URL}" >> $GITHUB_OUTPUT
+              echo "full_coverage_report_url=${FULL_COV_FILE_URL}" >> $GITHUB_ENV
+            fi
+
+            if [ "$HAS_FAILED_TESTS" = false ] && [ -f ${full_cov_csv} ]; then
+              python ${push_file} ${full_cov_csv} ${target_path}/CoverageData
+              python ${push_file} ${full_cov_csv} ${target_path_latest}/CoverageData
+              FULL_COV_CSV_URL=https://paddle-github-action.bj.bcebos.com/${target_path_stripped}/CoverageData/${full_cov_csv}
+              echo "full_coverage_csv_url=${FULL_COV_CSV_URL}" >> $GITHUB_OUTPUT
+              echo "full_coverage_csv_url=${FULL_COV_CSV_URL}" >> $GITHUB_ENV
+            fi
+          fi
       - name: Check Unit Test Success
         shell: bash
         run: |
@@ -291,6 +344,7 @@ jobs:
     needs: run_tests_with_coverage
     if: always()
     runs-on: ubuntu-latest
+    timeout-minutes: 15
     env:
       fd_archive_url: ${{ inputs.FASTDEPLOY_ARCHIVE_URL }}
     steps:
@@ -299,7 +353,7 @@ jobs:
         env:
           diff_cov_file_url: ${{ needs.run_tests_with_coverage.outputs.diff_cov_file_url }}
         run: |
-          wget ${fd_archive_url}
+          wget -q --no-proxy ${fd_archive_url}
           tar -xf FastDeploy.tar.gz
           cd FastDeploy
           if [ -z "${diff_cov_file_url}" ]; then

.github/workflows/ce_job.yml

@@ -9,7 +9,7 @@ on:
 permissions: read-all
 
 concurrency:
-  group: ${{ github.ref }}-${{ github.sha }}
+  group: CE-Job-${{ github.ref }}-${{ github.sha }}
   cancel-in-progress: true
 
 jobs:
@@ -154,6 +154,7 @@ jobs:
       COMPILE_ARCH: "80,90"
       WITH_NIGHTLY_BUILD: OFF
       FD_VERSION: 0.0.0
+      PADDLE_WHL_URL: ${{ needs.ce_job_pre_check.outputs.compile_use_paddle_whl_url }}
 
   build_sm8689:
     name: BUILD_SM8689
@@ -166,6 +167,7 @@ jobs:
       COMPILE_ARCH: "86,89"
       WITH_NIGHTLY_BUILD: OFF
       FD_VERSION: 0.0.0
+      PADDLE_WHL_URL: ${{ needs.ce_job_pre_check.outputs.compile_use_paddle_whl_url }}
 
   ce_upload_sm8090:
     environment: CodeSync
@@ -175,14 +177,13 @@ jobs:
     env:
       AK: ${{ secrets.BOS_AK }}
       SK: ${{ secrets.BOS_SK }}
-      FASTDEPLOY_WHEEL_URL: ${{ needs.build.outputs.wheel_path }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
       COMPILE_ARCH: "80,90"
     steps:
       - uses: actions/setup-python@v5
         with:
           python-version: '3.10'
       - name: Wheel Info Show and Upload
-        if: github.ref_name == 'develop' || github.ref_type == 'tag'
         run: |
           echo "The wheel is located at: ${{ needs.build_sm8090.outputs.wheel_path }}"
           wget -q --no-check-certificate ${{ needs.build_sm8090.outputs.wheel_path }}
@@ -190,7 +191,7 @@ jobs:
 
           commit_id=${{ github.sha }}
           branch_name=${{ github.ref_name }}
-          target_path=paddle-qa/paddle-pipeline/FastDeploy_ActionCE${COMPILE_ARCH//,/_}/${branch_name}/${commit_id}
+          target_path=paddle-qa/paddle-pipeline/FastDeploy_ActionCE/SM${COMPILE_ARCH//,/_}/${branch_name}/${commit_id}
 
           wget  -q --no-proxy --no-check-certificate https://paddle-qa.bj.bcebos.com/CodeSync/develop/PaddlePaddle/PaddleTest/tools/bos_tools.py
           push_file=$(realpath bos_tools.py)
@@ -198,12 +199,14 @@ jobs:
           ls
           python ${push_file} ${filename} ${target_path}
           target_path_stripped="${target_path#paddle-qa/}"
-          WHEEL_PATH=https://paddle-qa.bj.bcebos.com/${target_path_stripped}/${fd_wheel_name}
+          WHEEL_PATH=https://paddle-qa.bj.bcebos.com/${target_path_stripped}/${filename}
 
-          target_path_latest=paddle-qa/paddle-pipeline/FastDeploy_ActionCE${COMPILE_ARCH//,/_}/${branch_name}/latest
+          target_path_latest=paddle-qa/paddle-pipeline/FastDeploy_ActionCE/SM${COMPILE_ARCH//,/_}/${branch_name}/latest
           python ${push_file} ${filename} ${target_path_latest}
           target_path_stripped_latest="${target_path_latest#paddle-qa/}"
-          WHEEL_PATH_LATEST=https://paddle-qa.bj.bcebos.com/${target_path_stripped_latest}/${fd_wheel_name}
+          WHEEL_PATH_LATEST=https://paddle-qa.bj.bcebos.com/${target_path_stripped_latest}/${filename}
+          echo "commit wheel url is ${WHEEL_PATH}"
+          echo "latest wheel url is ${WHEEL_PATH_LATEST}"
 
   ce_upload_sm8689:
     environment: CodeSync
@@ -213,22 +216,21 @@ jobs:
     env:
       AK: ${{ secrets.BOS_AK }}
       SK: ${{ secrets.BOS_SK }}
-      FASTDEPLOY_WHEEL_URL: ${{ needs.build.outputs.wheel_path }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8689.outputs.wheel_path }}
       COMPILE_ARCH: "86,89"
     steps:
       - uses: actions/setup-python@v5
         with:
           python-version: '3.10'
       - name: Wheel Info Show and Upload
-        if: github.ref_name == 'develop' || github.ref_type == 'tag'
         run: |
-          echo "The wheel is located at: ${{ needs.build_sm8090.outputs.wheel_path }}"
-          wget -q --no-check-certificate ${{ needs.build_sm8090.outputs.wheel_path }}
-          filename=$(basename ${{ needs.build_sm8090.outputs.wheel_path }})
+          echo "The wheel is located at: ${{ needs.build_sm8689.outputs.wheel_path }}"
+          wget -q --no-check-certificate ${{ needs.build_sm8689.outputs.wheel_path }}
+          filename=$(basename ${{ needs.build_sm8689.outputs.wheel_path }})
 
           commit_id=${{ github.sha }}
           branch_name=${{ github.ref_name }}
-          target_path=paddle-qa/paddle-pipeline/FastDeploy_ActionCE${COMPILE_ARCH//,/_}/${branch_name}/${commit_id}
+          target_path=paddle-qa/paddle-pipeline/FastDeploy_ActionCE/SM${COMPILE_ARCH//,/_}/${branch_name}/${commit_id}
 
           wget  -q --no-proxy --no-check-certificate https://paddle-qa.bj.bcebos.com/CodeSync/develop/PaddlePaddle/PaddleTest/tools/bos_tools.py
           push_file=$(realpath bos_tools.py)
@@ -236,9 +238,11 @@ jobs:
           ls
           python ${push_file} ${filename} ${target_path}
           target_path_stripped="${target_path#paddle-qa/}"
-          WHEEL_PATH=https://paddle-qa.bj.bcebos.com/${target_path_stripped}/${fd_wheel_name}
+          WHEEL_PATH=https://paddle-qa.bj.bcebos.com/${target_path_stripped}/${filename}
 
-          target_path_latest=paddle-qa/paddle-pipeline/FastDeploy_ActionCE${COMPILE_ARCH//,/_}/${branch_name}/latest
+          target_path_latest=paddle-qa/paddle-pipeline/FastDeploy_ActionCE/SM${COMPILE_ARCH//,/_}/${branch_name}/latest
           python ${push_file} ${filename} ${target_path_latest}
           target_path_stripped_latest="${target_path_latest#paddle-qa/}"
-          WHEEL_PATH_LATEST=https://paddle-qa.bj.bcebos.com/${target_path_stripped_latest}/${fd_wheel_name}
+          WHEEL_PATH_LATEST=https://paddle-qa.bj.bcebos.com/${target_path_stripped_latest}/${filename}
+          echo "commit wheel url is ${WHEEL_PATH}"
+          echo "latest wheel url is ${WHEEL_PATH_LATEST}"

.github/workflows/ci_iluvatar.yml

@@ -28,18 +28,22 @@ jobs:
           REPO="https://github.com/${{ github.repository }}.git"
           FULL_REPO="${{ github.repository }}"
           REPO_NAME="${FULL_REPO##*/}"
+          BASE_BRANCH="${{ github.base_ref }}"
           # Clean the repository directory before starting
           docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
           -e "REPO_NAME=${REPO_NAME}" \
+          -e "BASE_BRANCH=${BASE_BRANCH}" \
           ${docker_image} /bin/bash -c '
             if [ -d ${REPO_NAME} ]; then
               echo "Directory ${REPO_NAME} exists, removing it..."
               rm -rf ${REPO_NAME}
             fi
           '
+          git config --global http.proxy "http://61.151.249.150:33128"
+          git config --global https.proxy "http://61.151.249.150:33128"
           git config --global user.name "FastDeployCI"
           git config --global user.email "fastdeploy_ci@example.com"
-          git clone ${REPO} ${REPO_NAME}
+          git clone --recursive ${REPO} ${REPO_NAME} -b ${BASE_BRANCH}
           cd FastDeploy
           if [ "${{ github.event_name }}" = "pull_request" ]; then
             git fetch origin pull/${{ github.event.pull_request.number }}/head:pr/${{ github.event.pull_request.number }}

.github/workflows/ci_image_update.yml

@@ -0,0 +1,174 @@
+name: CI Images Build
+
+on:
+  workflow_dispatch:
+  schedule:
+    - cron: '0 18 * * *'   # 2:00 AM China Standard Time (UTC+8)
+
+permissions: read-all
+
+concurrency:
+  group: CI-Images-Build-${{ github.ref }}-${{ github.sha }}
+  cancel-in-progress: true
+
+
+jobs:
+  clone:
+    environment: CodeSync
+    name: FD-Clone-Linux
+    runs-on: ubuntu-latest
+    outputs:
+      repo_archive_url: ${{ steps.set_output.outputs.repo_archive_url }}
+    steps:
+      - name: Clone FastDeploy
+        uses: actions/checkout@v4
+        with:
+          ref: ${{ github.ref_name }}
+          submodules: 'recursive'
+          fetch-depth: 1000
+
+      - name: Python Setup
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.10'
+      - name: Code Info Show and Upload
+        id: set_output
+        env:
+          AK: ${{ secrets.BOS_AK }}
+          SK: ${{ secrets.BOS_SK }}
+        run: |
+          git config --unset http.https://github.com/.extraheader
+          git submodule foreach --recursive sh -c "git config --local --unset-all 'http.https://github.com/.extraheader'"
+          git submodule foreach --recursive sh -c "git config remote.origin.fetch '+refs/heads/*:refs/remotes/origin/*'"
+          echo "Current HEAD Log:"
+          git log --oneline -n 5
+          ls
+          cd ..
+          tar -zcf FastDeploy.tar.gz FastDeploy
+          if [[ "${{ github.ref_type }}" == "tag" ]]; then
+            commit_id=${{ github.sha }}
+            tag_name=${{ github.ref_name }}
+            target_path=paddle-qa/TAG/FastDeploy/${tag_name}/${commit_id}
+          else
+            commit_id=${{ github.sha }}
+            branch_name=${{ github.ref_name }}
+            target_path=paddle-qa/BRANCH/FastDeploy/${branch_name}/${commit_id}
+          fi
+          wget  -q --no-proxy --no-check-certificate https://paddle-qa.bj.bcebos.com/CodeSync/develop/PaddlePaddle/PaddleTest/tools/bos_tools.py
+          push_file=$(realpath bos_tools.py)
+          python -m pip install bce-python-sdk==0.9.29
+          ls
+          python ${push_file} FastDeploy.tar.gz ${target_path}
+          target_path_stripped="${target_path#paddle-qa/}"
+          REPO_ARCHIVE_URL=https://paddle-qa.bj.bcebos.com/${target_path_stripped}/FastDeploy.tar.gz
+          echo "repo_archive_url=${REPO_ARCHIVE_URL}" >> $GITHUB_OUTPUT
+
+  resultshow:
+    name: Show Code Archive Output
+    needs: clone
+    runs-on: ubuntu-latest
+    steps:
+      - name: Print wheel path
+        run: |
+          echo "The code archive is located at: ${{ needs.clone.outputs.repo_archive_url }}"
+
+  ci_image_build:
+    name: CI Images Build
+    needs: clone
+    uses: ./.github/workflows/_ci_image_build.yml
+    with:
+      CI_DOCKER_IMAGE_NAME: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate-precheck
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+
+
+  build_sm8090:
+    name: BUILD_SM8090
+    needs: [clone, ci_image_build]
+    uses: ./.github/workflows/_build_linux.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      COMPILE_ARCH: "90"
+      WITH_NIGHTLY_BUILD: ${{ needs.publish_pre_check.outputs.with_nightly_build }}
+      FD_VERSION: ${{ needs.publish_pre_check.outputs.fd_version }}
+      PADDLEVERSION: ${{ needs.publish_pre_check.outputs.compile_use_paddle_version }}
+      PADDLE_WHL_URL: ${{ needs.publish_pre_check.outputs.compile_use_paddle_whl_url }}
+
+
+  unittest_coverage:
+    name: Run FastDeploy Unit Tests and Coverage
+    needs: [clone,build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_unit_test_coverage.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+    secrets:
+      github-token: ${{ secrets.GITHUB_TOKEN }}
+
+  logprob_test:
+    name: Run FastDeploy LogProb Tests
+    needs: [build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_logprob_test_linux.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      PADDLETEST_ARCHIVE_URL: "https://xly-devops.bj.bcebos.com/PaddleTest/PaddleTest.tar.gz"
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+  pre_ce_test:
+    name: Extracted partial CE model tasks to run in CI.
+    needs: [clone,build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_pre_ce_test.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+  base_test:
+    name: Run Base Tests
+    needs: [clone,build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_base_test.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+  accuracy_test:
+    name: Run Accuracy Tests
+    needs: [clone,build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_accuracy_test.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+  stable_test:
+    name: Run Stable Tests
+    needs: [clone,build_sm8090,ci_image_build]
+    uses: ./.github/workflows/_stable_test.yml
+    with:
+      DOCKER_IMAGE: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+
+  publish_pre_check:
+    name: Publish Docker Images Pre Check
+    needs: [ci_image_build, unittest_coverage,logprob_test,pre_ce_test,base_test,accuracy_test,stable_test]
+    runs-on: [self-hosted, Docker-Build]
+    steps:
+      - name: Images Uploading
+        env:
+          images_name: ${{ needs.ci_image_build.outputs.docker_name_precheck }}
+          ci_image_name: "ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate"
+        run: |
+          echo "images_name=${images_name}"
+          docker images ${ci_image_name}
+          docker tag ${images_name} ${ci_image_name}
+          docker push ${ci_image_name}

.github/workflows/ci_xpu.yml

@@ -24,7 +24,7 @@ jobs:
 
       - name: Code Checkout
         env:
-          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.1.0
+          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.2.0
         run: |
           REPO="https://github.com/${{ github.repository }}.git"
           FULL_REPO="${{ github.repository }}"
@@ -55,7 +55,7 @@ jobs:
 
       - name: Run CI unittest
         env:
-          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.1.0
+          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.2.0
         run: |
           runner_name="${{ runner.name }}"
           last_char="${runner_name: -1}"
@@ -77,7 +77,7 @@ jobs:
           -e "MODEL_PATH=/ssd3/model" \
           -e "http_proxy=$(git config --global --get http.proxy)" \
           -e "https_proxy=$(git config --global --get https.proxy)" \
-          -e "no_proxy=bcebos.com" \
+          -e "no_proxy=bcebos.com,mirrors.tuna.tsinghua.edu.cn,127.0.0.1,localhost" \
           -e "FD_API_PORT=${FD_API_PORT}" \
           -e "FD_ENGINE_QUEUE_PORT=${FD_ENGINE_QUEUE_PORT}" \
           -e "FD_METRICS_PORT=${FD_METRICS_PORT}" \

.github/workflows/pr_build_and_test.yml

@@ -21,7 +21,7 @@ jobs:
     with:
       DOCKER_IMAGE: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate
       FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
-      COMPILE_ARCH: "89,90"
+      COMPILE_ARCH: "90"
       WITH_NIGHTLY_BUILD: "OFF"
       FD_VERSION: "0.0.0"
 

.github/workflows/publish_job.yml

@@ -13,7 +13,7 @@ on:
 permissions: read-all
 
 concurrency:
-  group: ${{ github.ref }}-${{ github.sha }}
+  group: Publish-Job-${{ github.ref }}-${{ github.sha }}
   cancel-in-progress: true
 
 
@@ -268,6 +268,56 @@ jobs:
           ls
           python ${push_file} ${filename} ${target_path}
 
+  images_build:
+    name: Run FD Image Build
+    needs: [clone, publish_pre_check, build_sm8090]
+    runs-on: [self-hosted, Docker-Build]
+    if: |
+      github.event.repository.fork == false &&
+      (
+        (github.event_name == 'push' && github.ref_type == 'tag') ||
+        (github.event_name == 'workflow_dispatch' && github.ref_type == 'tag')
+      )
+    env:
+      FD_VERSION: ${{ needs.publish_pre_check.outputs.fd_version }}
+      PADDLEVERSION: ${{ needs.publish_pre_check.outputs.compile_use_paddle_version }}
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+    steps:
+      - name: Images Build
+        shell: bash
+        env:
+          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate
+          fd_archive_url: ${FASTDEPLOY_ARCHIVE_URL}
+        run: |
+          set -x
+          FULL_REPO="${{ github.repository }}"
+          REPO_NAME="${FULL_REPO##*/}"
+
+          # Clean the repository directory before starting
+          docker run --rm --net=host -v $(pwd):/workspace -w /workspace \
+          -e "REPO_NAME=${REPO_NAME}" \
+          ${docker_image} /bin/bash -c '
+            if [ -d ${REPO_NAME} ]; then
+              echo "Directory ${REPO_NAME} exists, removing it..."
+              rm -rf ${REPO_NAME}*
+            fi
+          '
+          wget -q --no-proxy ${fd_archive_url}
+          tar -xf FastDeploy.tar.gz
+          rm -rf FastDeploy.tar.gz
+          cd FastDeploy
+          git config --global user.name "FastDeployCI"
+          git config --global user.email "fastdeploy_ci@example.com"
+          git log -n 3 --oneline
+
+          PRODUCT_NAME=ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-cuda-12.6:${FD_VERSION}
+          docker build --no-cache -t ${PRODUCT_NAME} -f Dockerfile.gpu . \
+            --network host \
+            --build-arg PADDLE_VERSION=${PADDLEVERSION} \
+            --build-arg FD_VERSION=${FD_VERSION}
+
+          docker push ${PRODUCT_NAME}
+
   unittest_coverage:
     name: Run FastDeploy Unit Tests and Coverage
     needs: [clone,build_sm8090]
@@ -319,3 +369,13 @@ jobs:
       FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
       FASTDEPLOY_WHEEL_URL: ${{ needs.build_sm8090.outputs.wheel_path }}
       MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"
+
+  stable_test:
+    name: Run Stable Tests
+    needs: [clone,build_sm8090]
+    uses: ./.github/workflows/_stable_test.yml
+    with:
+      DOCKER_IMAGE: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:fastdeploy-ciuse-cuda126-dailyupdate
+      FASTDEPLOY_ARCHIVE_URL: ${{ needs.clone.outputs.repo_archive_url }}
+      FASTDEPLOY_WHEEL_URL: ${{ needs.build.outputs.wheel_path }}
+      MODEL_CACHE_DIR: "/ssd2/actions-runner/ModelData"

.github/workflows/rerun.yml

@@ -0,0 +1,157 @@
+name: Re-run
+
+on:
+  issue_comment:
+    types: [created]
+
+jobs:
+  re-run:
+    if: ${{ github.event.issue.pull_request && contains(github.event.comment.body, '/re-run') && github.event.comment.user.login == github.event.issue.user.login }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Cleanup
+        run: |
+          rm -rf * .[^.]*
+
+      - name: Checkout code
+        uses: actions/checkout@v5
+
+      - name: Rerun all failed jobs
+        if: ${{ contains(github.event.comment.body, 'all-failed') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'all-failed'
+
+      - name: Rerun Approval
+        if: ${{ contains(github.event.comment.body, 'approval') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Approval'
+
+      - name: Rerun CI_ILUVATAR
+        if: ${{ contains(github.event.comment.body, 'ci_iluvatar') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'CI_ILUVATAR'
+
+      - name: Rerun CI_XPU
+        if: ${{ contains(github.event.comment.body, 'ci_xpu') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'CI_XPU'
+
+      - name: Rerun Codestyle-check
+        if: ${{ contains(github.event.comment.body, 'codestyle') || contains(github.event.comment.body, 'pre_commit') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Pre Commit'
+
+      - name: Rerun Clone
+        if: ${{ contains(github.event.comment.body, 'clone') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'FD-Clone-Linux / code-clone'
+
+      - name: Rerun Build
+        if: ${{ contains(github.event.comment.body, 'build') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'FD-Build-Linux / fd-build'
+
+      - name: Rerun run_ce_cases
+        if: ${{ contains(github.event.comment.body, 'run_ce_cases') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Extracted partial CE model tasks to run in CI. / run_ce_cases'
+
+      - name: Rerun accuracy_tests
+        if: ${{ contains(github.event.comment.body, 'accuracy_tests') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run Accuracy Tests / accuracy_tests'
+
+      - name: Rerun base_tests
+        if: ${{ contains(github.event.comment.body, 'base_tests') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run Base Tests / base_tests'
+
+      - name: Rerun run_tests_logprob
+        if: ${{ contains(github.event.comment.body, 'run_tests_logprob') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run FastDeploy LogProb Tests / run_tests_logprob'
+
+      - name: Rerun run_tests_with_coverage
+        if: ${{ contains(github.event.comment.body, 'run_tests_with_coverage') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run FastDeploy Unit Tests and Coverage / run_tests_with_coverage'
+
+      - name: Rerun diff_coverage_report
+        if: ${{ contains(github.event.comment.body, 'diff_coverage_report') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run FastDeploy Unit Tests and Coverage / diff_coverage_report'
+
+      - name: Rerun stable_tests
+        if: ${{ contains(github.event.comment.body, 'stable_tests') }}
+        uses: ./.github/actions/rerun-workflow
+        with:
+          PR_ID: ${{ github.event.issue.number }}
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          OWNER: ${{ github.repository_owner }}
+          REPO: ${{ github.event.repository.name }}
+          JOB_NAME: 'Run Stable Tests / stable_tests'

.gitmodules

@@ -0,0 +1,10 @@
+[submodule "custom_ops/third_party/DeepGEMM"]
+	path = custom_ops/third_party/DeepGEMM
+	url = https://github.com/deepseek-ai/DeepGEMM.git
+	ignore = all
+[submodule "custom_ops/third_party/cutlass"]
+	path = custom_ops/third_party/cutlass
+	url = https://github.com/NVIDIA/cutlass.git
+[submodule "custom_ops/third_party/nlohmann_json"]
+	path = custom_ops/third_party/nlohmann_json
+	url = https://github.com/nlohmann/json.git

.pre-commit-config.yaml

@@ -1,3 +1,7 @@
+exclude: |
+  (?x)^(
+    dockerfiles/.+
+  )$
 default_install_hook_types:
   - pre-commit
   - commit-msg
@@ -27,6 +31,15 @@ repos:
   hooks:
   - id: ruff
     args: [--output-format, github, --fix, --line-length=120, --config, pyproject.toml]
+# For C++ files
+- repo: local
+  hooks:
+    - id: clang-format
+      name: clang-format
+      description: Format files with ClangFormat.
+      entry: clang-format -i
+      language: system
+      files: \.(c|cc|cxx|cpp|cu|h|cuh|hpp|hxx|xpu|kps)$
 # # 拼写检查
 # - repo: https://github.com/codespell-project/codespell
 #   rev: v2.4.1

README.md

@@ -26,6 +26,8 @@ English | [简体中文](README_CN.md)
 # FastDeploy : Inference and Deployment Toolkit for LLMs and VLMs based on PaddlePaddle
 
 ## News
+**[2025-09] 🔥 FastDeploy v2.2 is newly released!** It now offers compatibility with models in the HuggingFace ecosystem, has further optimized performance, and newly adds support for [baidu/ERNIE-21B-A3B-Thinking](https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Thinking)!
+
 **[2025-08] 🔥 Released FastDeploy v2.1:** A brand-new KV Cache scheduling strategy has been introduced, and expanded support for PD separation and CUDA Graph across more models. Enhanced hardware support has been added for platforms like Kunlun and Hygon, along with comprehensive optimizations to improve the performance of both the service and inference engine.
 
 **[2025-07] The FastDeploy 2.0 Inference Deployment Challenge is now live!** Complete the inference deployment task for the ERNIE 4.5 series open-source models to win official FastDeploy 2.0 merch and generous prizes! 🎁 You're welcome to try it out and share your feedback! 📌[Sign up here](https://www.wjx.top/vm/meSsp3L.aspx#) 📌[Event details](https://github.com/PaddlePaddle/FastDeploy/discussions/2728)
@@ -41,7 +43,7 @@ English | [简体中文](README_CN.md)
 - 🤝 **OpenAI API Server and vLLM Compatible**: One-command deployment with [vLLM](https://github.com/vllm-project/vllm/) interface compatibility.
 - 🧮 **Comprehensive Quantization Format Support**: W8A16, W8A8, W4A16, W4A8, W2A16, FP8, and more.
 - ⏩ **Advanced Acceleration Techniques**: Speculative decoding, Multi-Token Prediction (MTP) and Chunked Prefill.
-- 🖥️ **Multi-Hardware Support**: NVIDIA GPU, Kunlunxin XPU, Hygon DCU, Ascend NPU, Iluvatar GPU, Enflame GCU, MetaX GPU etc.
+- 🖥️ **Multi-Hardware Support**: NVIDIA GPU, Kunlunxin XPU, Hygon DCU, Ascend NPU, Iluvatar GPU, Enflame GCU, MetaX GPU, Intel Gaudi etc.
 
 ## Requirements
 
@@ -57,8 +59,10 @@ FastDeploy supports inference deployment on **NVIDIA GPUs**, **Kunlunxin XPUs**,
 - [Iluvatar GPU](./docs/get_started/installation/iluvatar_gpu.md)
 - [Enflame GCU](./docs/get_started/installation/Enflame_gcu.md)
 - [Hygon DCU](./docs/get_started/installation/hygon_dcu.md)
+- [MetaX GPU](./docs/get_started/installation/metax_gpu.md)
+- [Intel Gaudi](./docs/get_started/installation/intel_gaudi.md)
 
-**Note:** We are actively working on expanding hardware support. Additional hardware platforms including Ascend NPU and MetaX GPU are currently under development and testing. Stay tuned for updates!
+**Note:** We are actively working on expanding hardware support. Additional hardware platforms including Ascend NPU are currently under development and testing. Stay tuned for updates!
 
 ## Get Started
 
@@ -68,20 +72,12 @@ Learn how to use FastDeploy through our documentation:
 - [ERNIE-4.5-VL Multimodal Model Deployment](./docs/get_started/ernie-4.5-vl.md)
 - [Offline Inference Development](./docs/offline_inference.md)
 - [Online Service Deployment](./docs/online_serving/README.md)
-- [Full Supported Models List](./docs/supported_models.md)
 - [Best Practices](./docs/best_practices/README.md)
 
 ## Supported Models
 
-| Model | Data Type | PD Disaggregation | Chunked Prefill | Prefix Caching |  MTP | CUDA Graph | Maximum Context Length |
-|:--- | :------- | :---------- | :-------- | :-------- | :----- | :----- | :----- |
-|ERNIE-4.5-300B-A47B | BF16/WINT4/WINT8/W4A8C8/WINT2/FP8 | ✅| ✅ | ✅|✅| ✅ |128K |
-|ERNIE-4.5-300B-A47B-Base| BF16/WINT4/WINT8 | ✅| ✅ | ✅|❌| ✅ | 128K |
-|ERNIE-4.5-VL-424B-A47B | BF16/WINT4/WINT8 | WIP | ✅ | WIP | ❌ | WIP |128K |
-|ERNIE-4.5-VL-28B-A3B | BF16/WINT4/WINT8 | ❌ | ✅ | WIP | ❌ | WIP |128K |
-|ERNIE-4.5-21B-A3B | BF16/WINT4/WINT8/FP8  |  ❌ |  ✅ |  ✅ | ✅ | ✅|128K |
-|ERNIE-4.5-21B-A3B-Base | BF16/WINT4/WINT8/FP8  |  ✅ |  ✅ |  ✅ | ❌  | ✅|128K |
-|ERNIE-4.5-0.3B | BF16/WINT8/FP8  |  ✅ |  ✅ |  ✅ | ❌ | ✅| 128K |
+Learn how to download models, enable using the torch format, and more:
+- [Full Supported Models List](./docs/supported_models.md)
 
 ## Advanced Usage
 

README_CN.md

@@ -26,7 +26,9 @@
 # FastDeploy ：基于飞桨的大语言模型与视觉语言模型推理部署工具包
 
 ## 最新活动
-**[2025-08] 🔥 FastDeploy v2.1 全新发布:** 全新的KV Cache调度策略，更多模型支持PD分离和CUDA Graph，昆仑、海光等更多硬件支持增强，全方面优化服务和推理引擎的性能。
+**[2025-09] 🔥 FastDeploy v2.2 全新发布**: HuggingFace生态模型兼容，性能进一步优化，更新增对[baidu/ERNIE-21B-A3B-Thinking](https://huggingface.co/baidu/ERNIE-4.5-21B-A3B-Thinking)支持!
+
+**[2025-08] FastDeploy v2.1 发布**:全新的KV Cache调度策略，更多模型支持PD分离和CUDA Graph，昆仑、海光等更多硬件支持增强，全方面优化服务和推理引擎的性能。
 
 **[2025-07] 《FastDeploy2.0推理部署实测》专题活动已上线!** 完成文心4.5系列开源模型的推理部署等任务，即可获得骨瓷马克杯等FastDeploy2.0官方周边及丰富奖金！🎁 欢迎大家体验反馈～ 📌[报名地址](https://www.wjx.top/vm/meSsp3L.aspx#)   📌[活动详情](https://github.com/PaddlePaddle/FastDeploy/discussions/2728)
 
@@ -39,7 +41,7 @@
 - 🤝 **OpenAI API服务与vLLM兼容**：单命令部署，兼容[vLLM](https://github.com/vllm-project/vllm/)接口
 - 🧮 **全量化格式支持**：W8A16、W8A8、W4A16、W4A8、W2A16、FP8等
 - ⏩ **高级加速技术**：推测解码、多令牌预测（MTP）及分块预填充
-- 🖥️ **多硬件支持**：NVIDIA GPU、昆仑芯XPU、海光DCU、昇腾NPU、天数智芯GPU、燧原GCU、沐曦GPU等
+- 🖥️ **多硬件支持**：NVIDIA GPU、昆仑芯XPU、海光DCU、昇腾NPU、天数智芯GPU、燧原GCU、沐曦GPU、英特尔Gaudi等
 
 ## 要求
 
@@ -55,8 +57,10 @@ FastDeploy 支持在**英伟达（NVIDIA）GPU**、**昆仑芯（Kunlunxin）XPU
 - [天数 CoreX](./docs/zh/get_started/installation/iluvatar_gpu.md)
 - [燧原 S60](./docs/zh/get_started/installation/Enflame_gcu.md)
 - [海光 DCU](./docs/zh/get_started/installation/hygon_dcu.md)
+- [沐曦 GPU](./docs/zh/get_started/installation/metax_gpu.md)
+- [英特尔 Gaudi](./docs/zh/get_started/installation/intel_gaudi.md)
 
-**注意:** 我们正在积极拓展硬件支持范围。目前，包括昇腾（Ascend）NPU 和 沐曦（MetaX）GPU 在内的其他硬件平台正在开发测试中。敬请关注更新！
+**注意:** 我们正在积极拓展硬件支持范围。目前，包括昇腾（Ascend）NPU 等其他硬件平台正在开发测试中。敬请关注更新！
 
 ## 入门指南
 
@@ -66,20 +70,12 @@ FastDeploy 支持在**英伟达（NVIDIA）GPU**、**昆仑芯（Kunlunxin）XPU
 - [ERNIE-4.5-VL 部署](./docs/zh/get_started/ernie-4.5-vl.md)
 - [离线推理](./docs/zh/offline_inference.md)
 - [在线服务](./docs/zh/online_serving/README.md)
-- [模型支持列表](./docs/zh/supported_models.md)
 - [最佳实践](./docs/zh/best_practices/README.md)
 
 ## 支持模型列表
 
-| Model | Data Type | PD Disaggregation | Chunked Prefill | Prefix Caching |  MTP | CUDA Graph | Maximum Context Length |
-|:--- | :------- | :---------- | :-------- | :-------- | :----- | :----- | :----- |
-|ERNIE-4.5-300B-A47B | BF16/WINT4/WINT8/W4A8C8/WINT2/FP8 | ✅| ✅ | ✅|✅| ✅ |128K |
-|ERNIE-4.5-300B-A47B-Base| BF16/WINT4/WINT8 | ✅| ✅ | ✅|❌| ✅ | 128K |
-|ERNIE-4.5-VL-424B-A47B | BF16/WINT4/WINT8 | WIP | ✅ | WIP | ❌ | WIP |128K |
-|ERNIE-4.5-VL-28B-A3B | BF16/WINT4/WINT8 | ❌ | ✅ | WIP | ❌ | WIP |128K |
-|ERNIE-4.5-21B-A3B | BF16/WINT4/WINT8/FP8  |  ❌ |  ✅ |  ✅ | ✅ | ✅|128K |
-|ERNIE-4.5-21B-A3B-Base | BF16/WINT4/WINT8/FP8  |  ✅ |  ✅ |  ✅ | ❌  | ✅|128K |
-|ERNIE-4.5-0.3B | BF16/WINT8/FP8  |  ✅ |  ✅ |  ✅ | ❌ | ✅| 128K |
+通过我们的文档了解如何下载模型，如何支持torch格式等：
+- [模型支持列表](./docs/zh/supported_models.md)
 
 ## 进阶用法
 

benchmarks/backend_request_func.py

@@ -58,10 +58,12 @@ class RequestFuncOutput:
     """Output for requesting LLMs via API"""
 
     no: int = 0
+    request_id: str = ""
     generated_text: str = ""
     reasoning_content: str = ""
     success: bool = False
     latency: float = 0.0
+    end_timestamp: float = 0.0  # 模型完全返回的时间戳（秒, perf_counter基准）
     output_tokens: int = 0
     ttft: float = 0.0  # Time to first token
     arrival_time: list = field(default_factory=list)  # arrival_time
@@ -110,12 +112,14 @@ async def async_request_eb_openai_chat_completions(
         output = RequestFuncOutput()
         output.prompt_len = 0
         output.no = request_func_input.no
+        request_id = "None"
 
         ttft = 0.0
         st = time.perf_counter()
         most_recent_timestamp = st
         try:
             async with session.post(url=api_url, json=payload, headers=headers) as response:
+                data = {}
                 if response.status == 200:
                     async for chunk_bytes in response.content:
                         chunk_bytes = chunk_bytes.strip()
@@ -124,10 +128,13 @@ async def async_request_eb_openai_chat_completions(
 
                         chunk = chunk_bytes.decode("utf-8").removeprefix("data: ")
                         if chunk != "[DONE]":
-                            # print("####chunk:", chunk, type(chunk))
+                            #print("####chunk:", chunk, type(chunk))
                             timestamp = time.perf_counter()
                             data = json.loads(chunk)
 
+                            if request_id == "None" and "id" in data:
+                                request_id = data["id"]
+                            
                             if choices := data.get("choices"):
                                 content = choices[0]["delta"].get("content")
                                 reason_content = choices[0]["delta"].get("reasoning_content")
@@ -136,9 +143,12 @@ async def async_request_eb_openai_chat_completions(
                                     ttft = timestamp - st
                                     output.ttft = ttft
                                     # cached_tokens
+                                    if data["usage"] and data["usage"].get("prompt_tokens_details", {}):
                                         output.prompt_len = (
                                             data["usage"].get("prompt_tokens_details", {}).get("cached_tokens", 0)
                                         )
+                                    else:
+                                        output.prompt_len = 0
 
                                 # Decoding phase
                                 else:
@@ -151,9 +161,12 @@ async def async_request_eb_openai_chat_completions(
                                 output.output_tokens = usage.get("completion_tokens", 0)
                                 output.prompt_tokens = usage.get("prompt_tokens", 0)
                             
+
                             most_recent_timestamp = timestamp
 
                     # output.generated_text = generated_text
+                    # 在流式结束时，记录最后一个 chunk 收到的时间戳
+                    output.end_timestamp = most_recent_timestamp
                     if output.generated_text.strip() == "":
                         output.success = False
                         output.error = "No generated text found!"
@@ -175,6 +188,8 @@ async def async_request_eb_openai_chat_completions(
             exc_info = sys.exc_info()
             output.error = "".join(traceback.format_exception(*exc_info))
 
+        output.request_id = request_id
+
         # 保存失败请求结果
         if not output.success:
             with open("error_output.txt", "a") as f:

benchmarks/benchmark_mtp.py

@@ -98,7 +98,7 @@ def main(args):
         raise ValueError("--max_concurrency should be same length as --s_itl_base_model")
 
     for max_concurrency, s_itl in zip(args.max_concurrency, args.s_itl_base_model):
-        # Wramup
+        # Warmup
         print("Starting warmup...")
         with open(os.devnull, "w") as f:
             with contextlib.redirect_stdout(f):

benchmarks/benchmark_serving.py

@@ -150,7 +150,7 @@ async def get_request(
 
 
 def calculate_metrics(
-    input_requests: list[SampleRequest],
+    # input_requests: list[SampleRequest],
     outputs: list[RequestFuncOutput],
     dur_s: float,
     selected_percentiles: list[float],
@@ -177,7 +177,7 @@ def calculate_metrics(
             output_len = outputs[i].output_tokens
 
             if not output_len:
-                print("no output_len")
+                print("no output_len", outputs[i])
                 # We use the tokenizer to count the number of output tokens
                 # for some serving backends instead of looking at
                 # len(outputs[i].itl) since multiple output tokens may be
@@ -395,6 +395,7 @@ async def benchmark(
     print(f"Traffic request rate: {request_rate}")
     print(f"Burstiness factor: {burstiness} ({distribution})")
     print(f"Maximum request concurrency: {max_concurrency}")
+    print(f"Drop ratio: {args.drop_ratio}")
 
     pbar = None if disable_tqdm else tqdm(total=len(input_requests))
 
@@ -443,6 +444,8 @@ async def benchmark(
         tasks.append(asyncio.create_task(limited_request_func(request_func_input=request_func_input, pbar=pbar)))
     outputs: list[RequestFuncOutput] = await asyncio.gather(*tasks)
 
+    outputs.sort(key=lambda x: x.end_timestamp)
+
     if profile:
         print("Stopping profiler...")
         profile_input = RequestFuncInput(
@@ -460,12 +463,35 @@ async def benchmark(
     if pbar is not None:
         pbar.close()
 
+    benchmark_outputs = outputs
+    drop_ratio = args.drop_ratio
+    if 0.0 < drop_ratio < 1:
+        # 按drop_ratio头尾各舍弃一半请求，不计入benchmark统计
+        n = len(outputs)
+        drop_count = int(n * drop_ratio)
+        half = drop_count // 2
+        if half > 0:
+            benchmark_outputs = outputs[half : n - half]
+
+        # 先过滤掉 end_timestamp == 0.0 的请求（失败请求）
+        benchmark_outputs = [o for o in benchmark_outputs if o.end_timestamp != 0.0]
+
+        # 根据收到最后一个chunk的时间戳计算总时长
+        if len(benchmark_outputs) >= 2:
+            benchmark_duration = benchmark_outputs[-1].end_timestamp - benchmark_outputs[0].end_timestamp
+        else:
+            benchmark_duration = 0.0
+
+        print(f"丢弃前数量: {n}")
+        print(f"丢弃后数量: {len(benchmark_outputs)}")
+        print(f"benchmark_duration: {benchmark_duration} 秒")
+    else:
         benchmark_duration = time.perf_counter() - benchmark_start_time
-    print("benchmark_duration:", benchmark_duration)
+        print(f"benchmark_duration: {benchmark_duration} 秒")
 
     metrics, actual_output_lens = calculate_metrics(
-        input_requests=input_requests,
-        outputs=outputs,
+        # input_requests=input_requests,
+        outputs=benchmark_outputs,
         dur_s=benchmark_duration,
         # tokenizer=tokenizer,
         selected_percentiles=selected_percentiles,
@@ -494,7 +520,7 @@ async def benchmark(
         "total_token_throughput": metrics.total_token_throughput,
         "input_lens": [output.prompt_len for output in outputs],
         "infer_input_lens": [output.prompt_tokens for output in outputs],
-        "output_lens": actual_output_lens,
+        "output_lens": [output.output_tokens for output in outputs],
         "ttfts": [output.ttft for output in outputs],
         "itls": [output.itl for output in outputs],
         "input_texts": [input.prompt for input in input_requests],
@@ -609,7 +635,7 @@ def benchmark_metrics(
     goodput_config_dict = check_goodput_args(args)
 
     metrics, actual_output_lens = calculate_metrics(
-        input_requests=input_requests,
+        # input_requests=input_requests,
         outputs=outputs,
         dur_s=benchmark_duration,
         selected_percentiles=selected_percentiles,
@@ -965,7 +991,7 @@ if __name__ == "__main__":
     parser.add_argument(
         "--backend",
         type=str,
-        default="vllm",
+        default="openai-chat",
         choices=list(ASYNC_REQUEST_FUNCS.keys()),
     )
     parser.add_argument(
@@ -1081,6 +1107,12 @@ if __name__ == "__main__":
         action="store_true",
         help="shuffle dataset",
     )
+    parser.add_argument(
+        "--drop-ratio",
+        type=float,
+        default=0.0,
+        help="Drop ratio of the outputs. [0, 1)",
+    )
     parser.add_argument(
         "--trust-remote-code",
         action="store_true",

benchmarks/yaml/GLM45-air-32k-bf16.yaml

@@ -0,0 +1,5 @@
+max_model_len: 32768
+max_num_seqs: 128
+tensor_parallel_size: 4
+use_cudagraph: True
+load_choices: "default_v1"

benchmarks/yaml/GLM45-air-32k-wfp8afp8.yaml

@@ -0,0 +1,6 @@
+max_model_len: 32768
+max_num_seqs: 128
+tensor_parallel_size: 4
+use_cudagraph: True
+load_choices: "default_v1"
+quantization: wfp8afp8

benchmarks/yaml/deepseek-32k-tp8-wint4.yaml

@@ -0,0 +1,9 @@
+quantization: wint4
+load_choices: "default_v1"
+graph_optimization_config:
+  use_cudagraph: True
+  use_unique_memory_pool: True
+enable_prefix_caching: False
+max_num_seqs: 256
+max_model_len: 32768
+tensor_parallel_size: 8

benchmarks/yaml/eb45-128k-wint4-a800-tp8.yaml

@@ -6,3 +6,4 @@ tensor_parallel_size: 8
 max_num_batched_tokens: 4096
 max_num_partial_prefills: 3
 max_long_partial_prefills: 3
+quantization: wint4

benchmarks/yaml/eb45-128k-wint4-tp1-plas.yaml

@@ -0,0 +1,6 @@
+tensor_parallel_size: 1
+max_model_len: 131072
+max_num_seqs: 32
+quantization: wint4
+max_num_batched_tokens: 8192
+plas_attention_config: '{"plas_encoder_top_k_left": 50, "plas_encoder_top_k_right": 60, "plas_decoder_top_k_left": 100, "plas_decoder_top_k_right": 120}'

benchmarks/yaml/eb45-128k-wint8-a800-tp8.yaml

@@ -6,3 +6,4 @@ tensor_parallel_size: 8
 max_num_batched_tokens: 4096
 max_num_partial_prefills: 3
 max_long_partial_prefills: 3
+quantization: wint8

benchmarks/yaml/eb45-32k-wint2-tp4.yaml

@@ -0,0 +1,5 @@
+max_model_len: 32768
+max_num_seqs: 256
+kv_cache_ratio: 0.75
+tensor_parallel_size: 4
+gpu_memory_utilization: 0.9

benchmarks/yaml/eb45-32k-wint4-a800-tp4.yaml

@@ -1,6 +1,6 @@
 max_model_len: 32768
 max_num_seqs: 96
-gpu_memory_utilization: 0.9
+gpu_memory_utilization: 0.85
 kv_cache_ratio: 0.71
 tensor_parallel_size: 4
 quantization: wint4

benchmarks/yaml/eb45-32k-wint4-mtp-h800-tp4.yaml

@@ -1,6 +1,6 @@
 max_model_len: 32768
 max_num_seqs: 96
-gpu_memory_utilization: 0.9
+gpu_memory_utilization: 0.85
 kv_cache_ratio: 0.71
 tensor_parallel_size: 4
 quantization: wint4

benchmarks/yaml/eb45-32k-wint4-tp4_decode.yaml

@@ -13,3 +13,4 @@ pd_comm_port: "2334"
 max_num_batched_tokens: 384
 max_num_partial_prefills: 3
 max_long_partial_prefills: 3
+quantization: wint4

benchmarks/yaml/eb45-32k-wint4-tp4_prefill.yaml

@@ -10,3 +10,4 @@ engine_worker_queue_port: 6677
 cache_transfer_protocol: "rdma,ipc"
 rdma_comm_ports: "7675,7676,7677,7678"
 pd_comm_port: "2333"
+quantization: wint4

benchmarks/yaml/eb45-32k-wint8-a800-tp8.yaml

@@ -1,6 +1,6 @@
 max_model_len: 32768
 max_num_seqs: 96
-gpu_memory_utilization: 0.9
+gpu_memory_utilization: 0.85
 kv_cache_ratio: 0.71
 tensor_parallel_size: 8
 quantization: wint8

benchmarks/yaml/eb45-vl-128k-wint4-h800-tp8.yaml

@@ -0,0 +1,11 @@
+enable_mm: True
+max_model_len: 131072
+max_num_seqs: 56
+gpu_memory_utilization: 0.8
+kv_cache_ratio: 0.8
+tensor_parallel_size: 8
+quantization: wint4
+limit_mm_per_prompt: '{"image": 100, "video": 100}'
+enable_chunked_prefill: True
+max_num_batched_tokens: 384
+reasoning_parser: ernie-45-vl

benchmarks/yaml/eb45-vl-32k-wint8-a800-tp8.yaml

@@ -1,7 +1,7 @@
 enable_mm: True
 max_model_len: 32768
 max_num_seqs: 36
-gpu_memory_utilization: 0.95
+gpu_memory_utilization: 0.9
 kv_cache_ratio: 0.8
 tensor_parallel_size: 8
 quantization: wint8

benchmarks/yaml/eb45-vl-32k-wint8-h800-tp8.yaml

@@ -1,7 +1,7 @@
 enable_mm: True
 max_model_len: 32768
 max_num_seqs: 36
-gpu_memory_utilization: 0.8
+gpu_memory_utilization: 0.85
 kv_cache_ratio: 0.8
 tensor_parallel_size: 8
 quantization: wint8

benchmarks/yaml/eb45-vl-lite-32k-bf16-a800-tp1.yaml

@@ -0,0 +1,9 @@
+enable_mm: True
+max_model_len: 32768
+max_num_seqs: 128
+gpu_memory_utilization: 0.9
+kv_cache_ratio: 0.71
+tensor_parallel_size: 1
+enable_chunked_prefill: True
+max_num_batched_tokens: 384
+reasoning_parser: ernie-45-vl

benchmarks/yaml/eb45-vl-lite-32k-wint4-a800-tp1.yaml

@@ -0,0 +1,10 @@
+enable_mm: True
+max_model_len: 32768
+max_num_seqs: 128
+gpu_memory_utilization: 0.9
+kv_cache_ratio: 0.71
+tensor_parallel_size: 1
+enable_chunked_prefill: True
+max_num_batched_tokens: 384
+quantization: wint4
+reasoning_parser: ernie-45-vl

benchmarks/yaml/eb45-vl-lite-32k-wint8-a800-tp1.yaml

@@ -0,0 +1,10 @@
+enable_mm: True
+max_model_len: 32768
+max_num_seqs: 128
+gpu_memory_utilization: 0.9
+kv_cache_ratio: 0.71
+tensor_parallel_size: 1
+enable_chunked_prefill: True
+max_num_batched_tokens: 384
+quantization: wint8
+reasoning_parser: ernie-45-vl

benchmarks/yaml/request_yaml/GLM-32k.yaml

@@ -2,7 +2,7 @@ top_p: 0.95
 temperature: 0.6
 metadata:
   min_tokens: 1
-max_tokens: 65535
+max_tokens: 12288
 repetition_penalty: 1.0
 frequency_penalty: 0
 presence_penalty: 0

benchmarks/yaml/request_yaml/deepseek-32k.yaml

@@ -1,11 +1,10 @@
-top_p: 1.0
-temperature: 1.0
-metadata:
-  min_tokens: 1
-max_tokens: 30721
+temperature: 0.8
+top_p: 0.8
+presence_penalty: 0
 repetition_penalty: 1.0
 frequency_penalty: 0
-presence_penalty: 0
-skip_special_tokens: false
+max_tokens: 12288
+metadata:
+  min_tokens: 1
 chat_template_kwargs:
-  enable_thinking: true
+  enable_thinking: false

benchmarks/yaml/request_yaml/eb45-vl-128k.yaml

@@ -0,0 +1 @@
+max_tokens: 131071

benchmarks/yaml/request_yaml/eb45-vl-32k.yaml

@@ -0,0 +1 @@
+max_tokens: 12288

benchmarks/yaml/request_yaml/request.yaml

@@ -0,0 +1,11 @@
+top_p: 0.8
+temperature: 0.8
+max_tokens: 12288
+repetition_penalty: 1.0
+frequency_penalty: 0
+presence_penalty: 0
+metadata:
+  enable_thinking: false
+  min_tokens: 1
+chat_template_kwargs:
+  enable_thinking: false

benchmarks/yaml/request_yaml/x1-128k.yaml

@@ -0,0 +1,8 @@
+top_p: 0.95
+temperature: 0.6
+metadata:
+  min_tokens: 1
+max_tokens: 131071
+repetition_penalty: 1.0
+frequency_penalty: 0
+presence_penalty: 0

benchmarks/yaml/x1-64k-w4a8c8-tp4.yaml

@@ -1,10 +1,10 @@
-reasoning-parser: ernie_x1
-tool_call_parser: ernie_x1
+reasoning-parser: ernie-x1
+tool_call_parser: ernie-x1
 tensor_parallel_size: 4
 max_model_len: 65536
 max_num_seqs: 128
 enable_prefix_caching: True
 enable_chunked_prefill: True
 gpu_memory_utilization: 0.85
-use_cudagraph: True
-enable_custom_all_reduce: True
+graph_optimization_config:
+  use_cudagraph: True

benchmarks/yaml/x1-a3b-128k-wint8-h800-tp1.yaml

@@ -0,0 +1,7 @@
+tensor_parallel_size: 1
+max_model_len: 131072
+max_num_seqs: 32
+reasoning_parser: ernie-x1
+tool_call_parser: ernie-x1
+load_choices: "default_v1"
+quantization: wint8

build.sh

@@ -128,6 +128,12 @@ function copy_ops(){
       echo -e "MACA ops have been copy to fastdeploy"
       return
     fi
+    is_intel_hpu=`$python -c "import paddle; print(paddle.is_compiled_with_custom_device('intel_hpu'))"`
+    if [ "$is_intel_hpu" = "True" ]; then
+      DEVICE_TYPE="intel-hpu"
+      echo -e "intel_hpu ops have been copy to fastdeploy"
+      return
+    fi
 
     DEVICE_TYPE="cpu"
     cd ../../../../
@@ -143,9 +149,9 @@ function build_and_install_ops() {
   TMP_DIR_REAL_PATH=`readlink -f ${OPS_TMP_DIR}`
   is_xpu=`$python -c "import paddle; print(paddle.is_compiled_with_xpu())"`
   if [ "$is_xpu" = "True" ]; then
-    cd xpu_ops/src
+    cd xpu_ops
     bash build.sh ${TMP_DIR_REAL_PATH}
-    cd ../..
+    cd ..
   elif [ "$FD_CPU_USE_BF16" == "true" ]; then
     if [ "$FD_BUILDING_ARCS" == "" ]; then
       FD_CPU_USE_BF16=True ${python} setup_ops.py install --install-lib ${OPS_TMP_DIR}
@@ -159,7 +165,9 @@ function build_and_install_ops() {
     else
       FD_BUILDING_ARCS=${FD_BUILDING_ARCS} ${python} setup_ops.py install --install-lib ${OPS_TMP_DIR}
     fi
+    if [ -d "${OPS_TMP_DIR}" ]; then
       find ${OPS_TMP_DIR} -type f -name "*.o" -exec rm -f {} \;
+    fi
   else
       echo "Error: Invalid parameter '$FD_CPU_USE_BF16'. Please use true or false."
       exit 1

custom_ops/cpu_ops/get_padding_offset.cc

@@ -83,10 +83,7 @@ std::vector<paddle::Tensor> GetPaddingOffset(const paddle::Tensor &input_ids,
                  cum_offsets_out.data<int>(),
                  seq_length,
                  bsz);
-    return {x_remove_padding,
-            padding_offset,
-            cu_seqlens_q,
-            cu_seqlens_k};
+  return {x_remove_padding, padding_offset, cu_seqlens_q, cu_seqlens_k};
 }
 
 std::vector<std::vector<int64_t>> GetPaddingOffsetInferShape(
@@ -104,18 +101,13 @@ std::vector<paddle::DataType> GetPaddingOffsetInferDtype(
     const paddle::DataType &cum_offsets_dtype,
     const paddle::DataType &token_num_dtype,
     const paddle::DataType &seq_len_dtype) {
-    return {input_ids_dtype,
-            seq_len_dtype,
-            seq_len_dtype,
-            seq_len_dtype};
+  return {input_ids_dtype, seq_len_dtype, seq_len_dtype, seq_len_dtype};
 }
 
 PD_BUILD_STATIC_OP(get_padding_offset_cpu)
     .Inputs({"input_ids", "cum_offsets", "token_num", "seq_len"})
-    .Outputs({"x_remove_padding",
-              "padding_offset",
-              "cu_seqlens_q",
-              "cu_seqlens_k"})
+    .Outputs(
+        {"x_remove_padding", "padding_offset", "cu_seqlens_q", "cu_seqlens_k"})
     .SetKernelFn(PD_KERNEL(GetPaddingOffset))
     .SetInferShapeFn(PD_INFER_SHAPE(GetPaddingOffsetInferShape))
     .SetInferDtypeFn(PD_INFER_DTYPE(GetPaddingOffsetInferDtype));

custom_ops/cpu_ops/rebuild_padding.cc

@@ -19,7 +19,6 @@
 #define PD_BUILD_STATIC_OP(name) PD_BUILD_OP(static_op_##name)
 #endif
 
-
 template <typename T>
 void RebuildPaddingCPUImpl(T *output_data,
                            const T *input_data,
@@ -66,12 +65,10 @@ void RebuildAppendPaddingCPUImpl(T *output_data,
                                  const int64_t output_elem_nums) {
   for (int i = 0; i < output_elem_nums; ++i) {
     int out_token_id = i / dim_embed;
-        int ori_token_id =
-            out_token_id + output_padding_offset_data[out_token_id];
+    int ori_token_id = out_token_id + output_padding_offset_data[out_token_id];
     int bi = ori_token_id / max_input_length;
     if (seq_len_this_time_data[bi] == 0 ||
-            (seq_lens_decoder_data[bi] == 0 &&
-            seq_lens_encoder_data[bi] == 0)) {
+        (seq_lens_decoder_data[bi] == 0 && seq_lens_encoder_data[bi] == 0)) {
       continue;
     }
     int seq_id = 0;
@@ -118,8 +115,7 @@ std::vector<paddle::Tensor> RebuildPaddingCPU(
     int need_delete_token_num = 0;
     for (int i = 0; i < bsz; ++i) {
       if (seq_lens_encoder_cpu.data<int>()[i] > 0) {
-                need_delete_token_num +=
-                    seq_lens_encoder_cpu.data<int>()[i] - 1;
+        need_delete_token_num += seq_lens_encoder_cpu.data<int>()[i] - 1;
       }
     }
     int output_token_num = token_num - need_delete_token_num;

custom_ops/cpu_ops/set_value_by_flags.cc

@@ -14,7 +14,7 @@
 
 #include "paddle/extension.h"
 
-void set_value_by_flag_and_id(const bool *stop_flags,
+void set_value_by_flags_and_idx(const bool *stop_flags,
                                 int64_t *pre_ids_all,
                                 const int64_t *input_ids,
                                 const int *seq_lens_encoder,
@@ -50,7 +50,7 @@ void SetValueByFlagsAndIdx(const paddle::Tensor &pre_ids_all,
   int length = pre_ids_all_shape[1];
   int length_input_ids = input_ids.shape()[1];
 
-    set_value_by_flag_and_id(stop_flags.data<bool>(),
+  set_value_by_flags_and_idx(stop_flags.data<bool>(),
                              const_cast<int64_t *>(pre_ids_all.data<int64_t>()),
                              input_ids.data<int64_t>(),
                              seq_lens_encoder.data<int>(),

custom_ops/cpu_ops/simd_sort.cc

@@ -40,8 +40,8 @@ void probs_sort(const float *probs,
 std::vector<paddle::Tensor> SimdSort(const paddle::Tensor &probs) {
   const int bsz = probs.shape()[0];
   const int vocab_size = probs.shape()[1];
-    auto sorted_indices = paddle::empty(
-        {bsz, vocab_size}, paddle::DataType::INT64, probs.place());
+  auto sorted_indices =
+      paddle::empty({bsz, vocab_size}, paddle::DataType::INT64, probs.place());
   auto sorted_probs = paddle::empty(
       {bsz, vocab_size}, paddle::DataType::FLOAT32, probs.place());
   probs_sort(probs.data<float>(),

custom_ops/cpu_ops/simd_sort_fake.cc

@@ -18,8 +18,8 @@
 std::vector<paddle::Tensor> SimdSort(const paddle::Tensor &probs) {
   const int bsz = probs.shape()[0];
   const int vocab_size = probs.shape()[1];
-    auto sorted_indices = paddle::empty(
-        {bsz, vocab_size}, paddle::DataType::INT64, probs.place());
+  auto sorted_indices =
+      paddle::empty({bsz, vocab_size}, paddle::DataType::INT64, probs.place());
   auto sorted_probs = paddle::empty(
       {bsz, vocab_size}, paddle::DataType::FLOAT32, probs.place());
   return {sorted_indices, sorted_probs};

custom_ops/cpu_ops/stop_generation_multi_ends.cc

@@ -18,6 +18,10 @@
 #include <stdio.h>
 #include "paddle/extension.h"
 
+#ifndef PD_BUILD_STATIC_OP
+#define PD_BUILD_STATIC_OP(name) PD_BUILD_OP(static_op_##name)
+#endif
+
 bool is_in_end(const int64_t id, const int64_t *end_ids, int length) {
   bool flag = false;
   for (int i = 0; i < length; i++) {
@@ -49,6 +53,8 @@ void set_value_by_flags(bool *stop_flags,
     }
     if (!beam_search && is_in_end(topk_ids[bi], end_ids, end_length)) {
       stop_flags[bi] = true;
+      topk_ids[bi] = end_ids[0];
+      next_tokens[bi] = end_ids[0];
     }
   }
 }

custom_ops/cpu_ops/token_penalty_multi_scores.cc

@@ -75,12 +75,10 @@ void update_value_by_repeat_times(const int *repeat_times,
       int times = repeat_times_now[i];
       float logit_now = static_cast<float>(logits_now[i]);
       if (times == 0) {
-                logits_now[i] =
-                    static_cast<float>(logit_now / temperatures[bi]);
+        logits_now[i] = static_cast<float>(logit_now / temperatures[bi]);
       }
       logit_now = logit_now < 0 ? logit_now * alpha : logit_now / alpha;
-            logits_now[i] =
-                static_cast<float>(logit_now - times * beta - gamma);
+      logits_now[i] = static_cast<float>(logit_now - times * beta - gamma);
     }
   }
 }
@@ -94,8 +92,7 @@ void ban_bad_words(float *logits,
     float *logits_now = logits + bi * length;
     for (int bwid = 0; bwid < bad_words_length; ++bwid) {
       const int64_t bad_words_token_id = bad_words_list[bwid];
-            if (bad_words_token_id >= length || bad_words_token_id < 0)
-                continue;
+      if (bad_words_token_id >= length || bad_words_token_id < 0) continue;
       logits_now[bad_words_token_id] = -1e10;
     }
   }

custom_ops/cpu_ops/update_inputs.cc

@@ -34,9 +34,9 @@ void update_inputs_kernel(bool *not_need_stop,
     auto seq_len_encoder = seq_lens_encoder[bi];
     auto seq_len_decoder = seq_lens_decoder[bi];
     seq_lens_decoder[bi] =
-            stop_flag_now ? 0
-                          : (seq_len_decoder == 0 ? seq_len_encoder
-                                                  : seq_len_decoder + 1);
+        stop_flag_now
+            ? 0
+            : (seq_len_decoder == 0 ? seq_len_encoder : seq_len_decoder + 1);
     seq_lens_this_time[bi] = stop_flag_now ? 0 : 1;
     seq_lens_encoder[bi] = 0;
     int64_t *input_ids_now = input_ids + bi * input_ids_stride;
@@ -46,7 +46,7 @@ void update_inputs_kernel(bool *not_need_stop,
   not_need_stop[0] = stop_sum < stop_nums[0];
 }
 
-void UpdateInputes(const paddle::Tensor &stop_flags,
+void UpdateInputs(const paddle::Tensor &stop_flags,
                   const paddle::Tensor &not_need_stop,
                   const paddle::Tensor &seq_lens_this_time,
                   const paddle::Tensor &seq_lens_encoder,
@@ -90,4 +90,4 @@ PD_BUILD_STATIC_OP(update_inputs_cpu)
                     {"seq_lens_encoder", "seq_lens_encoder_out"},
                     {"seq_lens_decoder", "seq_lens_decoder_out"},
                     {"input_ids", "input_ids_out"}})
-    .SetKernelFn(PD_KERNEL(UpdateInputes));
+    .SetKernelFn(PD_KERNEL(UpdateInputs));

custom_ops/gpu_ops/append_attention.cu

@@ -59,7 +59,6 @@ void AppendAttentionKernel(
     const paddle::Tensor& decoder_tile_ids_per_batch,
     const paddle::Tensor& decoder_num_blocks,
     const paddle::Tensor& set_max_lengths,
-    const paddle::Tensor& max_len_kv,
     paddle::Tensor& fmha_out,
     const paddle::optional<paddle::Tensor>& rotary_embs,
     const paddle::optional<paddle::Tensor>& attn_mask,
@@ -73,10 +72,10 @@ void AppendAttentionKernel(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const paddle::optional<paddle::Tensor>& out_linear_shifts,
     const paddle::optional<paddle::Tensor>& out_linear_smooths,
-    const paddle::optional<paddle::Tensor>& mask_offset,
     const paddle::optional<paddle::Tensor>& kv_signal_data,
     const paddle::optional<paddle::Tensor>& q_norm_weight,
     const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const paddle::optional<paddle::Tensor>& sinks,
     const float rms_norm_eps,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
@@ -91,7 +90,8 @@ void AppendAttentionKernel(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   typedef PDTraits<D> traits_;
   typedef typename traits_::DataType DataType_;
   typedef typename traits_::data_t data_t;
@@ -103,6 +103,7 @@ void AppendAttentionKernel(
   int max_dec_len_this_time = set_max_lengths.data<int>()[2];
   int max_enc_dec_len_this_time = set_max_lengths.data<int>()[3];
   int max_just_dec_len_this_time = set_max_lengths.data<int>()[4];
+  int max_kv_len_this_time = set_max_lengths.data<int>()[8];
 
   auto main_stream = qkv.stream();
   static cudaEvent_t main_event;
@@ -140,12 +141,13 @@ void AppendAttentionKernel(
           key_cache,
           value_cache,
           attn_mask,
-          cache_k_dequant_scales,
-          cache_v_dequant_scales,
+          cache_quant_type_str == "block_wise_fp8" ? cache_k_quant_scales : cache_k_dequant_scales,
+          cache_quant_type_str == "block_wise_fp8" ? cache_v_quant_scales : cache_v_dequant_scales,
           cache_k_zp,
           cache_v_zp,
           out_linear_shifts,
           out_linear_smooths,
+          sinks,
           seq_lens_this_time,
           seq_lens_decoder,
           seq_lens_encoder,
@@ -169,7 +171,8 @@ void AppendAttentionKernel(
           lambda_is_decoder,
           lambda_enable_prefill,
           lambda_stream,
-          &fmha_out);
+          &fmha_out,
+          sliding_window);
   };
 
   if (max_enc_len_this_time > 0) {
@@ -245,7 +248,6 @@ void AppendAttentionKernel(
 
   if (max_just_dec_len_this_time > 0) {
     int decoder_num_blocks_data = decoder_num_blocks.data<int>()[0];
-    int max_len_kv_data = max_len_kv.data<int>()[0];
 
     cudaStream_t exec_stream;
     if (max_enc_len_this_time > 0) {
@@ -273,11 +275,15 @@ void AppendAttentionKernel(
             cache_v_zp,
             cache_quant_type_str,
             use_neox_rotary_style,
+            rope_3d,
             max_input_length,
             exec_stream,
             &qkv_out,
             const_cast<paddle::Tensor*>(&key_cache),
-            const_cast<paddle::Tensor*>(&value_cache));
+            const_cast<paddle::Tensor*>(&value_cache),
+            q_norm_weight,
+            k_norm_weight,
+            rms_norm_eps);
       } else {
         SpeculateWriteCacheWithRoPEKernel<data_t, data_t>(
             meta_data,
@@ -296,11 +302,15 @@ void AppendAttentionKernel(
             cache_v_zp,
             cache_quant_type_str,
             use_neox_rotary_style,
+            rope_3d,
             max_input_length,
             exec_stream,
             &qkv_out,
             const_cast<paddle::Tensor*>(&key_cache),
-            const_cast<paddle::Tensor*>(&value_cache));
+            const_cast<paddle::Tensor*>(&value_cache),
+            q_norm_weight,
+            k_norm_weight,
+            rms_norm_eps);
       }
     } else {
       if (qkv_out_scales) {
@@ -309,7 +319,6 @@ void AppendAttentionKernel(
             qkv,  // [token_num, num_heads, head_dim]
             seq_lens_decoder,
             seq_lens_encoder,
-            batch_id_per_token,
             cu_seqlens_q,
             block_tables,
             rotary_embs,
@@ -336,7 +345,6 @@ void AppendAttentionKernel(
             qkv_out,  // [token_num, num_heads, head_dim]
             seq_lens_decoder,
             seq_lens_encoder,
-            batch_id_per_token,
             cu_seqlens_q,
             block_tables,
             rotary_embs,
@@ -365,20 +373,20 @@ void AppendAttentionKernel(
         case paddle::DataType::INT8:{
         int8_t tmp;
         dispatch_CascadeAppendAttentionKernel(tmp, decoder_batch_ids, decoder_tile_ids_per_batch, decoder_num_blocks_data,
-        decoder_block_shape_q, max_len_kv_data, !speculate_decoder, !speculate_decoder, exec_stream);
+        decoder_block_shape_q, max_kv_len_this_time, !speculate_decoder, !speculate_decoder, exec_stream);
           break;
         }
         case paddle::DataType::FLOAT8_E4M3FN:{
         phi::dtype::float8_e4m3fn tmp;
         dispatch_CascadeAppendAttentionKernel(tmp, decoder_batch_ids, decoder_tile_ids_per_batch, decoder_num_blocks_data,
-        decoder_block_shape_q, max_len_kv_data, !speculate_decoder, !speculate_decoder, exec_stream);
+        decoder_block_shape_q, max_kv_len_this_time, !speculate_decoder, !speculate_decoder, exec_stream);
           break;
         }
       }
     } else {
         data_t tmp;
         dispatch_CascadeAppendAttentionKernel(tmp, decoder_batch_ids, decoder_tile_ids_per_batch, decoder_num_blocks_data,
-        decoder_block_shape_q, max_len_kv_data, !speculate_decoder, !speculate_decoder, exec_stream);
+        decoder_block_shape_q, max_kv_len_this_time, !speculate_decoder, !speculate_decoder, exec_stream);
     }
     if (max_enc_len_this_time > 0) {
       cudaEventRecord(decoder_event, exec_stream);
@@ -407,7 +415,6 @@ std::vector<paddle::Tensor> AppendAttention(
     const paddle::Tensor& decoder_tile_ids_per_batch,
     const paddle::Tensor& decoder_num_blocks,
     const paddle::Tensor& set_max_lengths,
-    const paddle::Tensor& max_len_kv,
     const paddle::optional<paddle::Tensor>& rotary_embs,
     const paddle::optional<paddle::Tensor>& attn_mask,
     const paddle::optional<paddle::Tensor>& qkv_bias,
@@ -424,6 +431,7 @@ std::vector<paddle::Tensor> AppendAttention(
     const paddle::optional<paddle::Tensor>& kv_signal_data,
     const paddle::optional<paddle::Tensor>& q_norm_weight,
     const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const paddle::optional<paddle::Tensor>& sinks,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -439,7 +447,8 @@ std::vector<paddle::Tensor> AppendAttention(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   AppendAttnMetaData meta_data;
 
   const auto& qkv_dims = qkv.dims();
@@ -488,12 +497,12 @@ std::vector<paddle::Tensor> AppendAttention(
   paddle::Tensor fmha_out;
   if (out_linear_in_scale > 0.0) {
     if (fabs(quant_max_bound - 127.0f) < 0.000001) {
-      fmha_out = GetEmptyTensor(
+      fmha_out = paddle::zeros(
         {meta_data.token_nums, meta_data.q_num_heads * meta_data.head_dims},
         paddle::DataType::INT8,
         qkv.place());
     } else if (fabs(quant_max_bound - 448.0f) < 0.000001) {
-      fmha_out = GetEmptyTensor(
+      fmha_out = paddle::zeros(
         {meta_data.token_nums, meta_data.q_num_heads * meta_data.head_dims},
         paddle::DataType::FLOAT8_E4M3FN,
         qkv.place());
@@ -501,7 +510,7 @@ std::vector<paddle::Tensor> AppendAttention(
       PD_THROW("Only supported attr of quant_max_bound in ['127', '448'].");
     }
   } else {
-    fmha_out = GetEmptyTensor(
+    fmha_out = paddle::zeros(
         {meta_data.token_nums, meta_data.q_num_heads * meta_data.head_dims},
         dtype_id,
         qkv.place());
@@ -533,7 +542,6 @@ std::vector<paddle::Tensor> AppendAttention(
           decoder_tile_ids_per_batch,
           decoder_num_blocks,
           set_max_lengths,
-          max_len_kv,
           fmha_out,
           rotary_embs,
           attn_mask,
@@ -547,10 +555,10 @@ std::vector<paddle::Tensor> AppendAttention(
           cache_v_zp,
           out_linear_shifts,
           out_linear_smooths,
-          mask_offset,
           kv_signal_data,
           q_norm_weight,
           k_norm_weight,
+          sinks,
           rms_norm_eps,
           cache_quant_type_str,
           use_neox_rotary_style,
@@ -565,7 +573,8 @@ std::vector<paddle::Tensor> AppendAttention(
           encoder_max_partition_size,
           speculate_max_draft_token_num,
           causal,
-          speculate_decoder);
+          speculate_decoder,
+          sliding_window);
   };
 
 
@@ -590,7 +599,7 @@ std::vector<paddle::Tensor> AppendAttention(
   return {paddle::Tensor{}};
 }
 
-void AppendAttentionWithOutput(
+std::vector<paddle::Tensor> AppendAttentionWithOutput(
     const paddle::Tensor& qkv,
     const paddle::Tensor& key_cache,
     const paddle::Tensor& value_cache,
@@ -610,7 +619,6 @@ void AppendAttentionWithOutput(
     const paddle::Tensor& decoder_tile_ids_per_batch,
     const paddle::Tensor& decoder_num_blocks,
     const paddle::Tensor& set_max_lengths,
-    const paddle::Tensor& max_len_kv,
     paddle::Tensor& fmha_out,
     const paddle::optional<paddle::Tensor>& rotary_embs,
     const paddle::optional<paddle::Tensor>& attn_mask,
@@ -628,6 +636,7 @@ void AppendAttentionWithOutput(
     const paddle::optional<paddle::Tensor>& kv_signal_data,
     const paddle::optional<paddle::Tensor>& q_norm_weight,
     const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const paddle::optional<paddle::Tensor>& sinks,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -643,7 +652,8 @@ void AppendAttentionWithOutput(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   AppendAttnMetaData meta_data;
 
   const auto& qkv_dims = qkv.dims();
@@ -689,7 +699,6 @@ void AppendAttentionWithOutput(
           decoder_tile_ids_per_batch,
           decoder_num_blocks,
           set_max_lengths,
-          max_len_kv,
           fmha_out,
           rotary_embs,
           attn_mask,
@@ -703,10 +712,10 @@ void AppendAttentionWithOutput(
           cache_v_zp,
           out_linear_shifts,
           out_linear_smooths,
-          mask_offset,
           kv_signal_data,
           q_norm_weight,
           k_norm_weight,
+          sinks,
           rms_norm_eps,
           cache_quant_type_str,
           use_neox_rotary_style,
@@ -721,7 +730,8 @@ void AppendAttentionWithOutput(
           encoder_max_partition_size,
           speculate_max_draft_token_num,
           causal,
-          speculate_decoder);
+          speculate_decoder,
+          sliding_window);
   };
 
   phi::dtype::float16 fp16_dtype;
@@ -755,6 +765,8 @@ void AppendAttentionWithOutput(
       break;
     }
   }
+
+  return {fmha_out};
 }
 
 
@@ -778,7 +790,6 @@ std::vector<std::vector<int64_t>> AppendAttentionInferShape(
     const std::vector<int64_t>& decoder_tile_ids_per_batch_shape,
     const std::vector<int64_t>& decoder_num_blocks_shape,
     const std::vector<int64_t>& set_max_lengths_shape,
-    const std::vector<int64_t>& max_len_kv_shape,
     const paddle::optional<std::vector<int64_t>>& rotary_embs_shape,
     const paddle::optional<std::vector<int64_t>>& attn_mask_shape,
     const paddle::optional<std::vector<int64_t>>& qkv_bias_shape,
@@ -795,6 +806,7 @@ std::vector<std::vector<int64_t>> AppendAttentionInferShape(
     const paddle::optional<std::vector<int64_t>>& kv_signal_data_shape,
     const paddle::optional<std::vector<int64_t>>& q_norm_weight_shape,
     const paddle::optional<std::vector<int64_t>>& k_norm_weight_shape,
+    const paddle::optional<std::vector<int64_t>>& sinks_shape,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -810,7 +822,8 @@ std::vector<std::vector<int64_t>> AppendAttentionInferShape(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   const int token_num = qkv_shape[0];
   const int kv_num_heads = key_cache_shape[1];
   int head_dim = key_cache_shape[3];
@@ -842,7 +855,6 @@ std::vector<paddle::DataType> AppendAttentionInferDtype(
     const paddle::DataType& decoder_tile_ids_per_batch_dtype,
     const paddle::DataType& decoder_num_blocks_dtype,
     const paddle::DataType& set_max_lengths_dtype,
-    const paddle::DataType& max_len_kv_dtype,
     const paddle::optional<paddle::DataType>& rotary_embs_dtype,
     const paddle::optional<paddle::DataType>& attn_mask_dtype,
     const paddle::optional<paddle::DataType>& qkv_bias_dtype,
@@ -859,6 +871,7 @@ std::vector<paddle::DataType> AppendAttentionInferDtype(
     const paddle::optional<paddle::DataType>& kv_signal_data_dtype,
     const paddle::optional<paddle::DataType>& q_norm_weight_dtype,
     const paddle::optional<paddle::DataType>& k_norm_weight_dtype,
+    const paddle::optional<paddle::DataType>& sinks_dtype,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -874,7 +887,8 @@ std::vector<paddle::DataType> AppendAttentionInferDtype(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   if (compute_dtype == "bf16") {
     if (out_linear_in_scale > 0.0) {
       if (fabs(quant_max_bound - 127.0f) < 0.000001) {
@@ -924,7 +938,6 @@ std::vector<std::vector<int64_t>> AppendAttentionWithOutputInferShape(
     const std::vector<int64_t>& decoder_tile_ids_per_batch_shape,
     const std::vector<int64_t>& decoder_num_blocks_shape,
     const std::vector<int64_t>& set_max_lengths_shape,
-    const std::vector<int64_t>& max_len_kv_shape,
     const std::vector<int64_t>& fmha_out_shape,
     const paddle::optional<std::vector<int64_t>>& rotary_embs_shape,
     const paddle::optional<std::vector<int64_t>>& attn_mask_shape,
@@ -942,6 +955,7 @@ std::vector<std::vector<int64_t>> AppendAttentionWithOutputInferShape(
     const paddle::optional<std::vector<int64_t>>& kv_signal_data_shape,
     const paddle::optional<std::vector<int64_t>>& q_norm_weight_shape,
     const paddle::optional<std::vector<int64_t>>& k_norm_weight_shape,
+    const paddle::optional<std::vector<int64_t>>& sinks_shape,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -957,7 +971,8 @@ std::vector<std::vector<int64_t>> AppendAttentionWithOutputInferShape(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   return {fmha_out_shape};
 }
 
@@ -981,7 +996,6 @@ std::vector<paddle::DataType> AppendAttentionWithOutputInferDtype(
     const paddle::DataType& decoder_tile_ids_per_batch_dtype,
     const paddle::DataType& decoder_num_blocks_dtype,
     const paddle::DataType& set_max_lengths_dtype,
-    const paddle::DataType& max_len_kv_dtype,
     const paddle::DataType& fmha_out_dtype,
     const paddle::optional<paddle::DataType>& rotary_embs_dtype,
     const paddle::optional<paddle::DataType>& attn_mask_dtype,
@@ -999,6 +1013,7 @@ std::vector<paddle::DataType> AppendAttentionWithOutputInferDtype(
     const paddle::optional<paddle::DataType>& kv_signal_data_dtype,
     const paddle::optional<paddle::DataType>& q_norm_weight_dtype,
     const paddle::optional<paddle::DataType>& k_norm_weight_dtype,
+    const paddle::optional<paddle::DataType>& sinks_dtype,
     const float rms_norm_eps,
     const std::string& compute_dtype,
     const std::string& cache_quant_type_str,
@@ -1014,7 +1029,8 @@ std::vector<paddle::DataType> AppendAttentionWithOutputInferDtype(
     const int encoder_max_partition_size,
     const int speculate_max_draft_token_num,
     const bool causal,
-    const bool speculate_decoder) {
+    const bool speculate_decoder,
+    const int sliding_window) {
   return {fmha_out_dtype};
 }
 
@@ -1040,7 +1056,6 @@ PD_BUILD_STATIC_OP(append_attention)
              "decoder_tile_ids_per_batch",
              "decoder_num_blocks",
              "set_max_lengths",
-             "max_len_kv",
              paddle::Optional("rotary_embs"),
              paddle::Optional("attn_mask"),
              paddle::Optional("qkv_bias"),
@@ -1056,10 +1071,9 @@ PD_BUILD_STATIC_OP(append_attention)
              paddle::Optional("mask_offset"),
              paddle::Optional("kv_signal_data"),
              paddle::Optional("q_norm_weight"),
-             paddle::Optional("k_norm_weight")})
-    .Outputs({"fmha_out", "key_cache_out", "value_cache_out"})
-    .SetInplaceMap({{"key_cache", "key_cache_out"},
-                    {"value_cache", "value_cache_out"}})
+             paddle::Optional("k_norm_weight"),
+             paddle::Optional("sinks")})
+    .Outputs({"fmha_out"})
     .Attrs({"rms_norm_eps: float",
             "compute_type: std::string",
             "cache_quant_type: std::string",
@@ -1076,6 +1090,7 @@ PD_BUILD_STATIC_OP(append_attention)
             "speculate_max_draft_token_num: int",
             "causal: bool",
             "speculate_decoder: bool",
+            "sliding_window: int",
             })
     .SetKernelFn(PD_KERNEL(AppendAttention))
     .SetInferShapeFn(PD_INFER_SHAPE(AppendAttentionInferShape))
@@ -1101,7 +1116,6 @@ PD_BUILD_STATIC_OP(append_attention_with_output)
              "decoder_tile_ids_per_batch",
              "decoder_num_blocks",
              "set_max_lengths",
-             "max_len_kv",
              "fmha_out",
              paddle::Optional("rotary_embs"),
              paddle::Optional("attn_mask"),
@@ -1118,11 +1132,10 @@ PD_BUILD_STATIC_OP(append_attention_with_output)
              paddle::Optional("mask_offset"),
              paddle::Optional("kv_signal_data"),
              paddle::Optional("q_norm_weight"),
-             paddle::Optional("k_norm_weight")})
-    .Outputs({"fmha_out_out", "qkv_out", "key_cache_out", "value_cache_out"})
-    .SetInplaceMap({{"fmha_out", "fmha_out_out"},
-                    {"key_cache", "key_cache_out"},
-                    {"value_cache", "value_cache_out"}})
+             paddle::Optional("k_norm_weight"),
+             paddle::Optional("sinks")})
+    .Outputs({"fmha_out_out"})
+    .SetInplaceMap({{"fmha_out", "fmha_out_out"}})
     .Attrs({"rms_norm_eps: float",
             "compute_type: std::string",
             "cache_quant_type: std::string",
@@ -1139,6 +1152,7 @@ PD_BUILD_STATIC_OP(append_attention_with_output)
             "speculate_max_draft_token_num: int",
             "causal: bool",
             "speculate_decoder: bool",
+            "sliding_window: int",
             })
     .SetKernelFn(PD_KERNEL(AppendAttentionWithOutput))
     .SetInferShapeFn(PD_INFER_SHAPE(AppendAttentionWithOutputInferShape))

custom_ops/gpu_ops/append_attn/append_attention_func.cuh

@@ -77,6 +77,14 @@ struct prefill_softmax_state_t {
 
   __device__ __forceinline__ void normalize() {
     const T d_t = static_cast<T>(d);
+#pragma unroll
+    for (size_t i = 0; i < vec_size; ++i) {
+      o[i] /= d_t;
+    }
+  }
+
+  __device__ __forceinline__ void normalize(float current_sink) {
+    const T d_t = static_cast<T>(d + __expf(current_sink - m));
 #pragma unroll
     for (size_t i = 0; i < vec_size; ++i) {
       o[i] /= d_t;
@@ -384,6 +392,113 @@ __device__ __forceinline__ void produce_v_blockwise_c8(
   }
 }
 
+template<uint32_t block_size,
+         uint32_t num_frags_z,
+         uint32_t NUM_WARP_Q,
+         typename T>
+__device__ __forceinline__ void produce_k_dynamic_scale(
+  T* k_smem_scale,
+  T* cache_k_reg,
+  const int* block_table_now,
+  const T* cache_k_scale,
+  const uint32_t kv_idx,
+  const uint32_t kv_num_heads,
+  const uint32_t kv_head_idx,
+  const uint32_t chunk_end
+) {
+  const uint32_t tx = threadIdx.x, ty = threadIdx.y;
+  if constexpr (NUM_WARP_Q == 4) {
+    // 4 warps shared block_size
+    const uint32_t tid = ty * 32 + tx;
+    int block_id = __ldg(&block_table_now[kv_idx / block_size]);
+    if (block_id < 0) block_id = 0;
+    const T* cache_k_scale_now = cache_k_scale + block_id * kv_num_heads * block_size + kv_head_idx * block_size;
+    if (tid < block_size) {
+      k_smem_scale[tid] = cache_k_scale_now[tid];
+    }
+    __syncthreads();
+    const uint32_t row_id = tx / 4;
+    for (uint32_t fz = 0; fz < num_frags_z; fz++) {
+      cache_k_reg[fz * 2] = k_smem_scale[fz * 16 + row_id];
+      cache_k_reg[fz * 2 + 1] = k_smem_scale[fz * 16 + row_id + 8];
+    }
+  } else {
+    // 1 warp 32 tokens
+    const uint32_t kv_idx_now = kv_idx + block_size * ty / 2;
+    int block_id = __ldg(&block_table_now[kv_idx_now / block_size]);
+    if (block_id < 0) block_id = 0;
+    const T* cache_k_scale_now = cache_k_scale + block_id * kv_num_heads * block_size + kv_head_idx * block_size;
+    const int kv_idx_this_thread = kv_idx + ty * 32 + tx;
+    if (kv_idx_this_thread < chunk_end) {
+      k_smem_scale[ty * 32 + tx] = cache_k_scale_now[(ty  % 2) * 32 + tx];
+    } else {
+      k_smem_scale[ty * 32 + tx] = 0;
+    }
+    __syncwarp();
+    const uint32_t row_id = tx / 4;
+    for (uint32_t fz = 0; fz < num_frags_z; fz++) {
+      cache_k_reg[fz * 2] = k_smem_scale[ty * 32 + fz * 16 + row_id];
+      cache_k_reg[fz * 2 + 1] = k_smem_scale[ty * 32 + fz * 16 + row_id + 8];
+    }
+  }
+}
+
+template<uint32_t block_size,
+         uint32_t num_frags_z,
+         uint32_t NUM_WARP_Q,
+         typename T>
+__device__ __forceinline__ void produce_v_dynamic_scale(
+  T* v_smem_scale,
+  T* cache_v_reg,
+  const int* block_table_now,
+  const T* cache_v_scale,
+  const uint32_t kv_idx,
+  const uint32_t kv_num_heads,
+  const uint32_t kv_head_idx,
+  const uint32_t chunk_end
+) {
+  const uint32_t tx = threadIdx.x, ty = threadIdx.y;
+
+  if constexpr (NUM_WARP_Q == 4) {
+    // 4 warps shared block_size
+    const uint32_t tid = ty * 32 + tx;
+    int block_id = __ldg(&block_table_now[kv_idx / block_size]);
+    if (block_id < 0) block_id = 0;
+    const T* cache_v_scale_now = cache_v_scale + block_id * kv_num_heads * block_size + kv_head_idx * block_size;
+    if (tid < block_size) {
+      v_smem_scale[tid] = cache_v_scale_now[tid];
+    }
+    __syncthreads();
+    const uint32_t row_id = tx % 4 * 2;
+    for (uint32_t fz = 0; fz < num_frags_z; fz++) {
+      cache_v_reg[fz * 4] = v_smem_scale[fz * 16 + row_id];
+      cache_v_reg[fz * 4 + 1] = v_smem_scale[fz * 16 + row_id + 1];
+      cache_v_reg[fz * 4 + 2] = v_smem_scale[fz * 16 + row_id + 8];
+      cache_v_reg[fz * 4 + 3] = v_smem_scale[fz * 16 + row_id + 9];
+    }
+  } else {
+    // 1 warp 32 tokens
+    const uint32_t kv_idx_now = kv_idx + block_size * ty / 2;
+    int block_id = __ldg(&block_table_now[kv_idx_now / block_size]);
+    if (block_id < 0) block_id = 0;
+    const T* cache_v_scale_now = cache_v_scale + block_id * kv_num_heads * block_size + kv_head_idx * block_size;
+    const int kv_idx_this_thread = kv_idx + ty * 32 + tx;
+    if (kv_idx_this_thread < chunk_end) {
+      v_smem_scale[ty * 32 + tx] = cache_v_scale_now[(ty  % 2) * 32 + tx];
+    } else {
+      v_smem_scale[ty * 32 + tx] = 0;
+    }
+    __syncwarp();
+    const uint32_t row_id = tx % 4 * 2;
+    for (uint32_t fz = 0; fz < num_frags_z; fz++) {
+      cache_v_reg[fz * 4] = v_smem_scale[ty * 32 + fz * 16 + row_id];
+      cache_v_reg[fz * 4 + 1] = v_smem_scale[ty * 32 + fz * 16 + row_id + 1];
+      cache_v_reg[fz * 4 + 2] = v_smem_scale[ty * 32 + fz * 16 + row_id + 8];
+      cache_v_reg[fz * 4 + 3] = v_smem_scale[ty * 32 + fz * 16 + row_id + 9];
+    }
+  }
+}
+
 template <SharedMemFillMode fill_mode,
           uint32_t num_warps,
           uint32_t block_size,
@@ -816,7 +931,8 @@ template <uint32_t num_frags_x,
           typename T,
           typename CacheT,
           bool is_scale_channel_wise = false,
-          bool IsFP8=false>
+          bool IsFP8 = false,
+          bool IsDynamicC8 = false>
 __device__ __forceinline__ void compute_qk_c8(smem_t* q_smem,
                                               uint32_t* q_smem_offset_r,
                                               smem_t* k_smem,
@@ -860,6 +976,7 @@ __device__ __forceinline__ void compute_qk_c8(smem_t* q_smem,
         convert_c8<T,IsFP8>(b_frag_dq_T, b_frag[fy * 2]);
         convert_c8<T,IsFP8>(b_frag_dq_T + 4, b_frag[fy * 2 + 1]);
         // scale zp
+        if constexpr (!IsDynamicC8) {
           if constexpr (is_scale_channel_wise) {
             const int scale_col = (ky * 2 + fy) * 4;
             b_frag_dq_T[0] *= cache_k_scale[scale_col];
@@ -876,6 +993,12 @@ __device__ __forceinline__ void compute_qk_c8(smem_t* q_smem,
               b_frag_dq_T[b_i] *= cache_k_scale[0];
             }
           }
+        } else {
+#pragma unroll
+          for (uint32_t b_i = 0; b_i < 8; ++b_i) {
+            b_frag_dq_T[b_i] *= cache_k_scale[fz * 2 + b_i / 4];
+          }
+        }
 #pragma unroll
         for (uint32_t fx = 0; fx < num_frags_x; ++fx) {
           if (ky == 0 && fy == 0) {
@@ -913,7 +1036,8 @@ __device__ __forceinline__ void mask_s(const bool* attn_mask,
                                        const uint32_t chunk_end,
                                        const uint32_t attn_mask_len,
                                        float (*s_frag)[num_frags_z][8],
-                                       const int *mask_offset = nullptr) {
+                                       const int *mask_offset = nullptr,
+                                       const int sliding_window = 0) {
   const uint32_t tx = threadIdx.x;
 #pragma unroll
   for (uint32_t fx = 0; fx < num_frags_x; ++fx) {
@@ -929,8 +1053,18 @@ __device__ __forceinline__ void mask_s(const bool* attn_mask,
                                   8 * (reg_id / 4) + reg_id % 2;
           bool out_of_boundary;
           if (mask_offset) {
-            out_of_boundary = q_idx < qo_len ? (kv_idx > mask_offset[q_idx]) : true;
-          } else {
+            out_of_boundary = q_idx < qo_len ? (kv_idx >= mask_offset[q_idx * 2 + 1] || kv_idx < mask_offset[q_idx * 2]) : true;
+          }
+          else if (sliding_window > 0)
+          {
+            bool out_of_window = int(kv_idx) <= (int)kv_len + (int)q_idx - (int)qo_len - sliding_window;
+            out_of_boundary =
+              (causal
+                  ? (kv_idx > kv_len + q_idx - qo_len || out_of_window || (kv_idx >= chunk_end))
+                  : kv_idx >= chunk_end);
+          }
+          else
+          {
             out_of_boundary =
               (causal
                   ? (kv_idx > kv_len + q_idx - qo_len || (kv_idx >= chunk_end))
@@ -949,7 +1083,7 @@ __device__ __forceinline__ void mask_s(const bool* attn_mask,
             s_frag[fx][fz][reg_id] =
                 out_of_boundary ? -3.0e+30f : s_frag[fx][fz][reg_id];
           }
-          // printf("tid: %d. qk[%u,%u] = %f, mask: %d \n ", threadIdx.x, kv_idx, q_idx, static_cast<float>(s_frag[fx][fz][reg_id]), int(out_of_boundary));
+
         } else {
           const uint32_t q_idx = qo_idx_base,
                          kv_idx = kv_idx_base + fz * 16 + 2 * (tx % 4) +
@@ -1093,7 +1227,9 @@ template <uint32_t num_frags_x,
           uint32_t block_size,
           typename T,
           typename CacheT,
-          bool is_scale_channel_wise = false, bool IsFP8=false>
+          bool is_scale_channel_wise = false,
+          bool IsFP8 = false,
+          bool IsDynamicC8 = false>
 __device__ __forceinline__ void compute_sfm_v_c8(
     smem_t* v_smem,
     uint32_t* v_smem_offset_r,
@@ -1135,6 +1271,7 @@ __device__ __forceinline__ void compute_sfm_v_c8(
         convert_c8<T,IsFP8>(b_frag_dq_T, b_frag[fz * 2]);
         convert_c8<T,IsFP8>(b_frag_dq_T + 4, b_frag[fz * 2 + 1]);
         // scale zp
+        if constexpr (!IsDynamicC8) {
           if constexpr (is_scale_channel_wise) {
 #pragma unroll
             for (uint32_t b_i = 0; b_i < 8; ++b_i) {
@@ -1146,6 +1283,17 @@ __device__ __forceinline__ void compute_sfm_v_c8(
               b_frag_dq_T[b_i] *= cache_v_scale[0];
             }
           }
+        } else {
+          const int scale_col = (kz * 2 + fz) * 4;
+          b_frag_dq_T[0] *= cache_v_scale[scale_col];
+          b_frag_dq_T[1] *= cache_v_scale[scale_col + 1];
+          b_frag_dq_T[2] *= cache_v_scale[scale_col + 2];
+          b_frag_dq_T[3] *= cache_v_scale[scale_col + 3];
+          b_frag_dq_T[4] *= cache_v_scale[scale_col];
+          b_frag_dq_T[5] *= cache_v_scale[scale_col + 1];
+          b_frag_dq_T[6] *= cache_v_scale[scale_col + 2];
+          b_frag_dq_T[7] *= cache_v_scale[scale_col + 3];
+        }
 #pragma unroll
         for (uint32_t fx = 0; fx < num_frags_x; ++fx) {  // m: num_frags_x * 16
           mma_sync_m16n16k16_row_col_f16f16f32<T>(
@@ -1171,7 +1319,9 @@ template <uint32_t num_frags_x,
           uint32_t block_size,
           typename T,
           typename CacheT,
-          bool is_scale_channel_wise = false, bool IsFP8=false>
+          bool is_scale_channel_wise = false,
+          bool IsFP8 = false,
+          bool IsDynamicC8 = false>
 __device__ __forceinline__ void compute_sfm_v_c8_iter_sq_bvec(
     smem_t* v_smem,
     uint32_t* v_smem_offset_r,
@@ -1215,6 +1365,7 @@ __device__ __forceinline__ void compute_sfm_v_c8_iter_sq_bvec(
         convert_c8<T,IsFP8>(b_frag_dq_T, b_frag[fz * 2]);
         convert_c8<T,IsFP8>(b_frag_dq_T + 4, b_frag[fz * 2 + 1]);
         // scale zp
+        if constexpr (!IsDynamicC8) {
           if constexpr (is_scale_channel_wise) {
 #pragma unroll
             for (uint32_t b_i = 0; b_i < 8; ++b_i) {
@@ -1226,6 +1377,17 @@ __device__ __forceinline__ void compute_sfm_v_c8_iter_sq_bvec(
               b_frag_dq_T[b_i] *= cache_v_scale[0];
             }
           }
+        } else {
+          const int scale_col = (kz * 2 + fz) * 4;
+          b_frag_dq_T[0] *= cache_v_scale[scale_col];
+          b_frag_dq_T[1] *= cache_v_scale[scale_col + 1];
+          b_frag_dq_T[2] *= cache_v_scale[scale_col + 2];
+          b_frag_dq_T[3] *= cache_v_scale[scale_col + 3];
+          b_frag_dq_T[4] *= cache_v_scale[scale_col];
+          b_frag_dq_T[5] *= cache_v_scale[scale_col + 1];
+          b_frag_dq_T[6] *= cache_v_scale[scale_col + 2];
+          b_frag_dq_T[7] *= cache_v_scale[scale_col + 3];
+        }
 #pragma unroll
         for (uint32_t fx = 0; fx < num_frags_x; ++fx) {  // m: num_frags_x * 16
           mma_sync_m16n16k16_row_col_f16f16f32<T>(
@@ -1315,6 +1477,33 @@ __device__ __forceinline__ void normalize_d(float (*o_frag)[num_frags_y][8],
   }
 }
 
+template <uint32_t num_frags_x, uint32_t num_frags_y>
+__device__ __forceinline__ void normalize_d(float (*o_frag)[num_frags_y][8],
+                                            float (*d)[2],
+                                            float (*m)[2],
+                                            float (*current_sinks)[2]) {
+  float d_rcp[num_frags_x][2];
+#pragma unroll
+  for (uint32_t fx = 0; fx < num_frags_x; ++fx) {
+#pragma unroll
+    for (uint32_t j = 0; j < 2; ++j) {
+      d_rcp[fx][j] = 1.f / (d[fx][j] + __expf(current_sinks[fx][j] - m[fx][j]));
+    }
+  }
+
+#pragma unroll
+  for (uint32_t fx = 0; fx < num_frags_x; ++fx) {
+#pragma unroll
+    for (uint32_t fy = 0; fy < num_frags_y; ++fy) {
+#pragma unroll
+      for (uint32_t reg_id = 0; reg_id < 8; ++reg_id) {
+        o_frag[fx][fy][reg_id] =
+            o_frag[fx][fy][reg_id] * d_rcp[fx][(reg_id % 4) / 2];
+      }
+    }
+  }
+}
+
 template <uint32_t num_frags_x,
           uint32_t num_frags_y,
           uint32_t NUM_WARPS,
@@ -2128,6 +2317,7 @@ __global__ void merge_multi_chunks_decoder_kernel(
     const int *__restrict__ cu_seqlens_q,
     const T *__restrict__ shift_bias,     // [q_num_heads * HEAD_DIM]
     const T *__restrict__ smooth_weight,  // [q_num_heads * HEAD_DIM]
+    const T *__restrict__ sinks, // [q_num_heads]
     OutT *__restrict__ out,
     const float quant_max_bound,
     const float quant_min_bound,
@@ -2165,17 +2355,11 @@ __global__ void merge_multi_chunks_decoder_kernel(
   using LoadT = AlignedVector<T, vec_size>;
   LoadT load_vec;
   LoadT res_vec;
-  if constexpr (std::is_same<T, half>::value) {
-#pragma unroll
-    for (int i = 0; i < vec_size / 2; ++i) {
-      *((half2 *)(&res_vec) + i) = make_half2(0, 0);
-    }
-  } else {
-#pragma unroll
-    for (int i = 0; i < vec_size / 2; ++i) {
-      *((nv_bfloat162 *)(&res_vec) + i) = make_bfloat162(0, 0);
-    }
+
+  for (int i = 0; i < vec_size; ++i) {
+    res_vec[i] = T(0.f);
   }
+
   float m;
   float d = 1.f;
   if constexpr (std::is_same<T, half>::value) {
@@ -2191,8 +2375,7 @@ __global__ void merge_multi_chunks_decoder_kernel(
     const float m_now = multi_m[offset];
     const float d_now = multi_d[offset];
     m = max(m_prev, m_now);
-    offset = (bid * num_chunks * num_heads + i * num_heads + hid) * head_dim +
-             vid * vec_size;
+    offset = offset * head_dim + vid * vec_size;
     Load<T, vec_size>(&multi_out[offset], &load_vec);
     const float scale1 = __expf(m_prev - m), scale2 = __expf(m_now - m);
     const T scale1_T = static_cast<T>(scale1),
@@ -2218,7 +2401,12 @@ __global__ void merge_multi_chunks_decoder_kernel(
       const float m_tmp = md_smem[2 * i], d_tmp = md_smem[2 * i + 1];
       st.merge(load_vec, m_tmp, d_tmp);
     }
+    if (sinks) {
+      float current_sink = static_cast<float>(sinks[hid]);
+      st.normalize(current_sink);
+    } else {
       st.normalize();
+    }
 
     const uint32_t shift_smooth_offset = hid * head_dim + vid * vec_size;
     AlignedVector<T, vec_size> shift_bias_vec;
@@ -2258,6 +2446,7 @@ __global__ void merge_multi_chunks_v2_kernel(
     const int *__restrict__ cu_seqlens_q,
     const T *__restrict__ shift_bias,     // [q_num_heads * HEAD_DIM]
     const T *__restrict__ smooth_weight,  // [q_num_heads * HEAD_DIM]
+    const T *__restrict__ sinks,  // [q_num_heads]
     OutT *__restrict__ out,
     const float quant_max_bound,
     const float quant_min_bound,
@@ -2275,6 +2464,9 @@ __global__ void merge_multi_chunks_v2_kernel(
   __shared__ float md_smem[bdy * 2];
   for (int qid = blockIdx.x; qid < token_num; qid += gridDim.x) {
     const uint32_t bid = batch_id_per_token[qid];
+    if(bid == -1){
+      continue;
+    }
     const uint32_t local_seq_id = qid - cu_seqlens_q[bid];
     const int seq_len_q = seq_lens_q[bid];
     if (seq_len_q == 0) continue;
@@ -2294,6 +2486,8 @@ __global__ void merge_multi_chunks_v2_kernel(
     const int num_chunks_this_seq = div_up(seq_len_kv, chunk_size);
     if (num_chunks_this_seq <= 1) {
       continue;
+    }else if (!ENABLE_PREFILL){
+      continue;
     }
 
     using LoadT = AlignedVector<T, vec_size>;
@@ -2370,7 +2564,13 @@ __global__ void merge_multi_chunks_v2_kernel(
         const float m_tmp = md_smem[2 * i], d_tmp = md_smem[2 * i + 1];
         st.merge(load_vec, m_tmp, d_tmp);
       }
+
+      if (sinks) {
+        float current_sink = static_cast<float>(sinks[hid]);
+        st.normalize(current_sink);
+      } else {
         st.normalize();
+      }
 
       const uint32_t shift_smooth_offset = hid * head_dim + vid * vec_size;
       AlignedVector<T, vec_size> shift_bias_vec;

custom_ops/gpu_ops/append_attn/append_attention_kernel.h

@@ -15,141 +15,9 @@
 
 #include "helper.h"
 #include "utils.cuh"
-
-template <typename T, typename OutT>
-void CascadeAppendAttentionC16Kernel(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);
-
-template <typename T, typename OutT, bool IsFP8 = false>
-void CascadeAppendAttentionC8Kernel(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);
-
-template <typename T, typename OutT>
-void CascadeAppendAttentionC4Kernel(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);
+#include "append_attention_c16_impl.cuh"
+#include "append_attention_c8_impl.cuh"
+#include "append_attention_c4_impl.cuh"
 
 template <typename T, typename OutT>
 void CascadeAppendAttentionKernel(
@@ -172,6 +40,8 @@ void CascadeAppendAttentionKernel(
         shift_bias,  // [num_kv_heads, head_dim]
     const paddle::optional<paddle::Tensor>&
         smooth_weight,  // [num_kv_heads, head_dim]
+    const paddle::optional<paddle::Tensor>&
+        sinks,  // [num_heads]
     const paddle::Tensor& seq_lens_q,
     const paddle::Tensor& seq_lens_kv,
     const paddle::Tensor& seq_lens_encoder,
@@ -195,7 +65,8 @@ void CascadeAppendAttentionKernel(
     const bool is_decoder,
     const bool enable_prefill,
     cudaStream_t& stream,
-    paddle::Tensor* out) {
+    paddle::Tensor* out,
+    const int sliding_window) {
     if (cache_quant_type_str == "none") {
         CascadeAppendAttentionC16Kernel<T, OutT>(meta_data,
                                                 qkv,
@@ -208,6 +79,7 @@ void CascadeAppendAttentionKernel(
                                                 cache_v_zp,
                                                 shift_bias,
                                                 smooth_weight,
+                                                sinks,
                                                 seq_lens_q,
                                                 seq_lens_kv,
                                                 seq_lens_encoder,
@@ -230,9 +102,10 @@ void CascadeAppendAttentionKernel(
                                                 is_decoder,
                                                 enable_prefill,
                                                 stream,
-                                                out);
+                                                out,
+                                                sliding_window);
     } else if (cache_quant_type_str == "cache_int8") {
-        CascadeAppendAttentionC8Kernel<T, OutT>(meta_data,
+        CascadeAppendAttentionC8Kernel<T, OutT, false>(meta_data,
                                                 qkv,
                                                 cache_k,
                                                 cache_v,
@@ -243,6 +116,7 @@ void CascadeAppendAttentionKernel(
                                                 cache_v_zp,
                                                 shift_bias,
                                                 smooth_weight,
+                                                sinks,
                                                 seq_lens_q,
                                                 seq_lens_kv,
                                                 seq_lens_encoder,
@@ -264,9 +138,11 @@ void CascadeAppendAttentionKernel(
                                                 causal,
                                                 is_decoder,
                                                 enable_prefill,
+                                                cache_quant_type_str,
                                                 stream,
-                                                out);
-    } else if (cache_quant_type_str == "cache_fp8") {
+                                                out,
+                                                sliding_window);
+    } else if (cache_quant_type_str == "cache_fp8" or cache_quant_type_str == "block_wise_fp8") {
         CascadeAppendAttentionC8Kernel<T, OutT, true>(meta_data,
                                                 qkv,
                                                 cache_k,
@@ -278,6 +154,7 @@ void CascadeAppendAttentionKernel(
                                                 cache_v_zp,
                                                 shift_bias,
                                                 smooth_weight,
+                                                sinks,
                                                 seq_lens_q,
                                                 seq_lens_kv,
                                                 seq_lens_encoder,
@@ -299,8 +176,10 @@ void CascadeAppendAttentionKernel(
                                                 causal,
                                                 is_decoder,
                                                 enable_prefill,
+                                                cache_quant_type_str,
                                                 stream,
-                                                out);
+                                                out,
+                                                sliding_window);
     } else if (cache_quant_type_str == "cache_int4_zp") {
         CascadeAppendAttentionC4Kernel<T, OutT>(meta_data,
                                                 qkv,
@@ -313,6 +192,7 @@ void CascadeAppendAttentionKernel(
                                                 cache_v_zp,
                                                 shift_bias,
                                                 smooth_weight,
+                                                sinks,
                                                 seq_lens_q,
                                                 seq_lens_kv,
                                                 seq_lens_encoder,
@@ -335,7 +215,8 @@ void CascadeAppendAttentionKernel(
                                                 is_decoder,
                                                 enable_prefill,
                                                 stream,
-                                                out);
+                                                out,
+                                                sliding_window);
     } else {
         PD_THROW(
             "cache_quant_type_str should be one of [none, cache_int8, "

custom_ops/gpu_ops/append_attn/autogen_template_instantiation.py

@@ -0,0 +1,243 @@
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Universal template instantiation generator - fully based on configuration file template instantiation generation."""
+
+import argparse
+import json
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Any, Dict, List, Optional, Tuple
+
+
+@dataclass
+class TemplateConfig:
+    """Template configuration class."""
+
+    name: str  # Function name
+    function_name: str  # Actual function name
+    impl_file: str  # Implementation file path
+    template_params: List[str]  # Template parameter list (in order)
+    dispatch_params: Dict[str, List[Any]]  # Dispatch parameters
+    data_types: Optional[List[Tuple[str, str, str]]] = None  # Data type combinations (input_type, output_type, suffix)
+    max_instances_per_file: int = 60  # Maximum instances per file
+    file_prefix: str = ""  # File prefix
+    function_signature: str = ""  # Function signature template
+
+
+class UniversalTemplateInstantiator:
+    """Universal template instantiator - fully based on configuration file."""
+
+    def __init__(self, config_file: str):
+        """Initialize the instantiator."""
+        self.config_file = config_file
+        self.configs = self._load_configs()
+
+    def _load_configs(self) -> Dict[str, TemplateConfig]:
+        """Load configuration file."""
+        with open(self.config_file, "r", encoding="utf-8") as f:
+            config_data = json.load(f)
+
+        configs = {}
+        for name, config_dict in config_data.items():
+            config = TemplateConfig(**config_dict)
+            self._validate_config(config)
+            configs[name] = config
+        return configs
+
+    def _validate_config(self, config: TemplateConfig):
+        """Validate configuration completeness."""
+        has_t = "T" in config.template_params
+        has_out_t = "OutT" in config.template_params
+
+        if (has_t or has_out_t) and not config.data_types:
+            raise ValueError(
+                f"Configuration '{config.name}' has T or OutT in template_params but no data_types configured"
+            )
+
+        special_params = {"T", "OutT", "NUM_WARP_Q"}
+        for param_name in config.template_params:
+            if param_name not in special_params and param_name not in config.dispatch_params:
+                raise ValueError(f"Template parameter '{param_name}' in '{config.name}' not found in dispatch_params")
+
+        if "NUM_WARP_Q" in config.template_params and "BLOCK_SHAPE_Q" not in config.dispatch_params:
+            raise ValueError(
+                f"Template parameter 'NUM_WARP_Q' in '{config.name}' requires 'BLOCK_SHAPE_Q' in dispatch_params"
+            )
+
+    def _calculate_num_warp_q(self, block_shape_q: int) -> int:
+        """Calculate number of warps."""
+        if block_shape_q <= 32:
+            return 1
+        else:
+            return 4
+
+    def _build_template_args(self, config: TemplateConfig, t_in: str, t_out: str, params: Dict[str, Any]) -> str:
+        """Build template arguments."""
+        template_args_parts = []
+
+        for param_name in config.template_params:
+            if param_name == "T":
+                if t_in:
+                    template_args_parts.append(t_in)
+                else:
+                    raise ValueError("Template parameter 'T' requires input type, but data_types is empty or invalid")
+            elif param_name == "OutT":
+                if t_out:
+                    template_args_parts.append(t_out)
+                else:
+                    raise ValueError(
+                        "Template parameter 'OutT' requires output type, but data_types is empty or invalid"
+                    )
+            elif param_name == "NUM_WARP_Q":
+                if "BLOCK_SHAPE_Q" in params:
+                    num_warp_q = self._calculate_num_warp_q(params["BLOCK_SHAPE_Q"])
+                    template_args_parts.append(str(num_warp_q))
+                else:
+                    raise ValueError("Template parameter 'NUM_WARP_Q' requires 'BLOCK_SHAPE_Q' in dispatch_params")
+            elif param_name in params:
+                template_args_parts.append(str(params[param_name]))
+            else:
+                raise ValueError(f"Template parameter '{param_name}' not found in dispatch_params")
+
+        return f"<{', '.join(template_args_parts)}>"
+
+    def _generate_function_signature(self, config: TemplateConfig, template_args: str) -> str:
+        """Generate function signature."""
+        if config.function_signature:
+            return config.function_signature.format(function_name=config.function_name, template_args=template_args)
+        else:
+            raise ValueError(f"Function signature not found for {config.name}")
+
+    def _generate_file_header(self, config: TemplateConfig) -> str:
+        """Generate file header."""
+        return f"""// Generated by autogen_template_instantiation.py - Do not edit.
+
+#pragma once
+
+#include "../../{config.impl_file}"
+"""
+
+    def _generate_template_instantiation(
+        self, config: TemplateConfig, t_in: str, t_out: str, params: Dict[str, Any]
+    ) -> str:
+        """Generate template instantiation."""
+        template_args = self._build_template_args(config, t_in, t_out, params)
+        return self._generate_function_signature(config, template_args)
+
+    def generate_combinations_for_type(self, config: TemplateConfig, t_in: str, t_out: str) -> List[Dict[str, Any]]:
+        """Generate parameter combinations for specific type."""
+        combinations = []
+
+        def _generate_recursive(
+            params_dict: Dict[str, List[Any]], current_params: Dict[str, Any], param_names: List[str]
+        ):
+            if not param_names:
+                combinations.append(current_params.copy())
+                return
+
+            param_name = param_names[0]
+            for value in params_dict[param_name]:
+                current_params[param_name] = value
+                _generate_recursive(params_dict, current_params, param_names[1:])
+
+        _generate_recursive(config.dispatch_params, {}, list(config.dispatch_params.keys()))
+        return combinations
+
+    def split_combinations(self, combinations: List[Dict[str, Any]], max_per_file: int) -> List[List[Dict[str, Any]]]:
+        """Split combinations into multiple files."""
+        chunks = []
+        for i in range(0, len(combinations), max_per_file):
+            chunk = combinations[i : i + max_per_file]
+            chunks.append(chunk)
+        return chunks
+
+    def generate_file_content(
+        self,
+        config: TemplateConfig,
+        t_in: str,
+        t_out: str,
+        t_out_name: str,
+        file_index: int,
+        combinations: List[Dict[str, Any]],
+    ) -> str:
+        """Generate file content."""
+        content = self._generate_file_header(config)
+
+        for params in combinations:
+            content += self._generate_template_instantiation(config, t_in, t_out, params)
+
+        return content
+
+    def generate_for_function_type(self, function_name: str, output_dir: str):
+        """Generate template instantiation files for specific function type."""
+        if function_name not in self.configs:
+            raise ValueError(f"Function type '{function_name}' not found in config")
+
+        config = self.configs[function_name]
+        output_path = Path(output_dir)
+        output_path.mkdir(exist_ok=True)
+
+        if not config.data_types:
+            data_types = [("", "", "")]
+        else:
+            data_types = config.data_types
+
+        for t_in, t_out, t_out_name in data_types:
+            combinations = self.generate_combinations_for_type(config, t_in, t_out)
+            if combinations:
+                chunks = self.split_combinations(combinations, config.max_instances_per_file)
+                for i, chunk in enumerate(chunks):
+                    filename = f"{config.file_prefix}{t_out_name}_part_{i:02d}.cu"
+                    filepath = output_path / filename
+                    content = self.generate_file_content(config, t_in, t_out, t_out_name, i, chunk)
+                    with open(filepath, "w", encoding="utf-8") as f:
+                        f.write(content)
+
+    def generate_all(self, output_dir: str):
+        """Generate all configured function types."""
+        for function_name in self.configs.keys():
+            print(f"Generating template instantiations for {function_name}...")
+            self.generate_for_function_type(function_name, output_dir)
+            print(f"Completed generating {function_name} template instantiations.")
+
+
+def main():
+    """Main function."""
+    parser = argparse.ArgumentParser(description="Universal template instantiation generator")
+    parser.add_argument(
+        "--config",
+        "-c",
+        type=str,
+        default="gpu_ops/append_attn/template_config.json",
+        help="Configuration file path (JSON format)",
+    )
+    parser.add_argument(
+        "--output",
+        "-o",
+        type=str,
+        default="gpu_ops/append_attn/template_instantiation/autogen",
+        help="Output directory",
+    )
+
+    args = parser.parse_args()
+
+    try:
+        instantiator = UniversalTemplateInstantiator(args.config)
+        instantiator.generate_all(args.output)
+    except Exception as e:
+        print(f"Error: {e}")
+
+
+if __name__ == "__main__":
+    main()

custom_ops/gpu_ops/append_attn/decode_attention_func.cuh

@@ -13,8 +13,8 @@
 // limitations under the License.
 #pragma once
 
-
-#include "multi_head_latent_attention_kernel.h"
+#include "helper.h"
+#include "utils.cuh"
 
 template <size_t vec_size, typename T>
 struct softmax_state_t {

custom_ops/gpu_ops/append_attn/decoder_write_cache_with_rope_kernel.cu

@@ -21,7 +21,6 @@ void append_decode_cache_rope_qk_norm(const QKV_TYPE* qkv,
                                       T* value_cache,
                                       T* qkv_out,
                                       const int* block_tables,
-                              const int* batch_id_per_token,
                                       const int* cu_seqlens_q,
                                       const int* seq_lens,
                                       const int* seq_lens_encoder,
@@ -59,7 +58,6 @@ void append_decode_cache_rope_qk_norm(const QKV_TYPE* qkv,
                                             value_cache,
                                             qkv_out,
                                             block_tables,
-                                            batch_id_per_token,
                                             cu_seqlens_q,
                                             seq_lens,
                                             seq_lens_encoder,
@@ -84,7 +82,6 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
                               T* value_cache,
                               T* qkv_out,
                               const int* block_tables,
-                              const int* batch_id_per_token,
                               const int* cu_seqlens_q,
                               const int* seq_lens,
                               const int* seq_lens_encoder,
@@ -97,6 +94,7 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
                               const int num_heads,
                               const int kv_num_heads,
                               const int dim_head,
+                              const int rotary_dim,
                               const int block_size,
                               const int bsz,
                               const cudaStream_t& stream,
@@ -120,7 +118,6 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -137,13 +134,36 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
               kv_num_heads,
               rope_3d);
     } else {
-      append_decode_cache_T_neox_rope_kernel<T, PackSize>
-          <<<grid_size, blocksize, 0, stream>>>(reinterpret_cast<const T*>(qkv),
+      if (rotary_dim < dim_head) {
+        append_decode_cache_T_neox_partial_rope_kernel<T, PackSize>
+            <<<grid_size, blocksize, 0, stream>>>(
+                reinterpret_cast<const T*>(qkv),
+                key_cache,
+                value_cache,
+                qkv_out,
+                block_tables,
+                cu_seqlens_q,
+                seq_lens,
+                seq_lens_encoder,
+                cos_emb,
+                sin_emb,
+                max_seq_len,
+                max_blocks_per_seq,
+                num_heads,
+                dim_head,
+                rotary_dim,
+                block_size,
+                elem_nums,
+                kv_num_heads,
+                rope_3d);
+      } else {
+        append_decode_cache_T_neox_rope_kernel<T, PackSize>
+            <<<grid_size, blocksize, 0, stream>>>(
+                reinterpret_cast<const T*>(qkv),
                 key_cache,
                 value_cache,
                 qkv_out,
                 block_tables,
-                                                batch_id_per_token,
                 cu_seqlens_q,
                 seq_lens,
                 seq_lens_encoder,
@@ -158,6 +178,7 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
                 kv_num_heads,
                 rope_3d);
       }
+    }
   } else {
     if (qkv_out_scales) {
       append_decode_cache_T_rope_kernel<T, PackSize>
@@ -167,7 +188,6 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -190,7 +210,6 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
                                                 value_cache,
                                                 qkv_out,
                                                 block_tables,
-                                                batch_id_per_token,
                                                 cu_seqlens_q,
                                                 seq_lens,
                                                 seq_lens_encoder,
@@ -208,13 +227,15 @@ void append_decode_cache_rope(const QKV_TYPE* qkv,
   }
 }
 
-template <typename T, typename QKV_TYPE, bool is_scale_channel_wise = false, bool IsFP8=false>
+template <typename T,
+          typename QKV_TYPE,
+          bool is_scale_channel_wise = false,
+          bool IsFP8 = false>
 void append_decode_cache_int8_rope(const QKV_TYPE* qkv,
                                    uint8_t* key_cache,
                                    uint8_t* value_cache,
                                    T* qkv_out,
                                    const int* block_tables,
-                                   const int* batch_id_per_token,
                                    const int* cu_seqlens_q,
                                    const int* seq_lens,
                                    const int* seq_lens_encoder,
@@ -247,7 +268,6 @@ void append_decode_cache_int8_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -273,7 +293,6 @@ void append_decode_cache_int8_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -292,14 +311,18 @@ void append_decode_cache_int8_rope(const QKV_TYPE* qkv,
     }
   } else {
     if (qkv_out_scales) {
-      append_decode_cache_int8_rope_kernel<T, 4, 0, 128, is_scale_channel_wise, IsFP8>
+      append_decode_cache_int8_rope_kernel<T,
+                                           4,
+                                           0,
+                                           128,
+                                           is_scale_channel_wise,
+                                           IsFP8>
           <<<grids, num_warps * 32, 0, stream>>>(
               reinterpret_cast<const int*>(qkv),
               key_cache,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -318,14 +341,18 @@ void append_decode_cache_int8_rope(const QKV_TYPE* qkv,
               kv_num_heads,
               rope_3d);
     } else {
-      append_decode_cache_int8_rope_kernel<T, 4, 0, 128, is_scale_channel_wise, IsFP8>
+      append_decode_cache_int8_rope_kernel<T,
+                                           4,
+                                           0,
+                                           128,
+                                           is_scale_channel_wise,
+                                           IsFP8>
           <<<grids, num_warps * 32, 0, stream>>>(
               reinterpret_cast<const T*>(qkv),
               key_cache,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -351,7 +378,6 @@ void append_decode_cache_int4_rope(const QKV_TYPE* qkv,
                                    uint8_t* value_cache,
                                    T* qkv_out,
                                    const int* block_tables,
-                                   const int* batch_id_per_token,
                                    const int* cu_seqlens_q,
                                    const int* seq_lens,
                                    const int* seq_lens_encoder,
@@ -386,7 +412,6 @@ void append_decode_cache_int4_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -414,7 +439,6 @@ void append_decode_cache_int4_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -442,7 +466,6 @@ void append_decode_cache_int4_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -470,7 +493,6 @@ void append_decode_cache_int4_rope(const QKV_TYPE* qkv,
               value_cache,
               qkv_out,
               block_tables,
-              batch_id_per_token,
               cu_seqlens_q,
               seq_lens,
               seq_lens_encoder,
@@ -497,7 +519,6 @@ void DecoderWriteCacheWithRoPEKernel(
     const paddle::Tensor& qkv,
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,
@@ -534,11 +555,24 @@ void DecoderWriteCacheWithRoPEKernel(
   const float* cos_emb =
       rotary_embs ? rotary_embs.get().data<float>() : nullptr;
   const float* sin_emb;
+  int rotary_dim = dim_head;
   if (rotary_embs) {
     sin_emb =
         use_neox_rotary_style
             ? rotary_embs.get().data<float>() + max_seq_len * dim_head
             : rotary_embs.get().data<float>() + max_seq_len * dim_head / 2;
+    rotary_dim =
+        rotary_embs.get().dims()[rotary_embs.get().dims().size() - 1] * 2;
+    if (rotary_dim < dim_head) {
+      if (!use_neox_rotary_style || qkv_out_scales || q_norm_weight ||
+          k_norm_weight || cache_quant_type_str != "none") {
+        PADDLE_THROW(phi::errors::Fatal(
+            "partial_rotary_factor < 1.0 only supports neox_rotary_style=True, "
+            "qkv_out_scales is None, q_norm_weight/k_norm_weight) is None, and "
+            "cache_quant_type_str is 'none'."));
+      }
+      sin_emb = rotary_embs.get().data<float>() + max_seq_len * rotary_dim / 2;
+    }
   }
 
   if (q_norm_weight && k_norm_weight) {
@@ -549,7 +583,6 @@ void DecoderWriteCacheWithRoPEKernel(
           reinterpret_cast<DataType_*>(value_cache_out->data<T>()),
           reinterpret_cast<DataType_*>(qkv_out->data<T>()),
           block_tables.data<int>(),
-          batch_id_per_token.data<int>(),
           cu_seqlens_q.data<int>(),
           seq_lens.data<int>(),
           seq_lens_encoder.data<int>(),
@@ -572,9 +605,86 @@ void DecoderWriteCacheWithRoPEKernel(
           q_norm_weight ? q_norm_weight.get().data<float>() : nullptr,
           k_norm_weight ? k_norm_weight.get().data<float>() : nullptr,
           rms_norm_eps);
+    } else if (cache_quant_type_str == "block_wise_fp8") {
+      constexpr int num_warps = 4;
+      const int all_warps = ((num_heads + 2 * kv_num_heads) + num_warps - 1) /
+                            num_warps * num_warps;
+      dim3 grids(bsz, all_warps / num_warps);
+      append_decode_cache_int8_rope_qk_norm_kernel<DataType_,
+                                                   4,
+                                                   0,
+                                                   128,
+                                                   false,
+                                                   true,
+                                                   true>
+          <<<grids, num_warps * 32, 0, stream>>>(
+              reinterpret_cast<const DataType_*>(qkv_ptr),
+              key_cache_out->data<uint8_t>(),
+              value_cache_out->data<uint8_t>(),
+              reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+              block_tables.data<int>(),
+              cu_seqlens_q.data<int>(),
+              seq_lens.data<int>(),
+              seq_lens_encoder.data<int>(),
+              cos_emb,
+              sin_emb,
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  cache_k_scale.get().data<T>())),
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  (cache_v_scale.get().data<T>()))),
+              q_norm_weight.get().data<float>(),
+              k_norm_weight.get().data<float>(),
+              max_seq_len,
+              max_blocks_per_seq,
+              num_heads,
+              block_size,
+              127.0f,
+              -127.0f,
+              kv_num_heads,
+              rope_3d,
+              rms_norm_eps);
+    } else if ((cache_quant_type_str == "cache_fp8")) {
+      constexpr int num_warps = 4;
+      const int all_warps = ((num_heads + 2 * kv_num_heads) + num_warps - 1) /
+                            num_warps * num_warps;
+      dim3 grids(bsz, all_warps / num_warps);
+      append_decode_cache_int8_rope_qk_norm_kernel<DataType_,
+                                                   4,
+                                                   0,
+                                                   128,
+                                                   false,
+                                                   true,
+                                                   false>
+          <<<grids, num_warps * 32, 0, stream>>>(
+              reinterpret_cast<const DataType_*>(qkv_ptr),
+              key_cache_out->data<uint8_t>(),
+              value_cache_out->data<uint8_t>(),
+              reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+              block_tables.data<int>(),
+              cu_seqlens_q.data<int>(),
+              seq_lens.data<int>(),
+              seq_lens_encoder.data<int>(),
+              cos_emb,
+              sin_emb,
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  cache_k_scale.get().data<T>())),
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  (cache_v_scale.get().data<T>()))),
+              q_norm_weight.get().data<float>(),
+              k_norm_weight.get().data<float>(),
+              max_seq_len,
+              max_blocks_per_seq,
+              num_heads,
+              block_size,
+              127.0f,
+              -127.0f,
+              kv_num_heads,
+              rope_3d,
+              rms_norm_eps);
     } else {
       PD_THROW(
-          "append_decode_cache_rope_qk_norm not support cachekv quant yet");
+          "append_decode_cache_rope_qk_norm just supports cache_quant_type "
+          "none/block_wise_fp8/cache_fp8");
     }
   } else {
     if (cache_quant_type_str == "none") {
@@ -584,7 +694,6 @@ void DecoderWriteCacheWithRoPEKernel(
           reinterpret_cast<DataType_*>(value_cache_out->data<T>()),
           reinterpret_cast<DataType_*>(qkv_out->data<T>()),
           block_tables.data<int>(),
-          batch_id_per_token.data<int>(),
           cu_seqlens_q.data<int>(),
           seq_lens.data<int>(),
           seq_lens_encoder.data<int>(),
@@ -599,6 +708,7 @@ void DecoderWriteCacheWithRoPEKernel(
           num_heads,
           kv_num_heads,
           dim_head,
+          rotary_dim,
           block_size,
           bsz,
           stream,
@@ -606,7 +716,8 @@ void DecoderWriteCacheWithRoPEKernel(
           rope_3d);
     } else if (cache_quant_type_str == "cache_int8") {
       bool is_scale_channel_wise = false;
-      if (cache_k_scale && cache_k_scale.get().dims()[0] == dim_head * kv_num_heads) {
+      if (cache_k_scale &&
+          cache_k_scale.get().dims()[0] == dim_head * kv_num_heads) {
         is_scale_channel_wise = true;
       }
       if (is_scale_channel_wise) {
@@ -616,7 +727,6 @@ void DecoderWriteCacheWithRoPEKernel(
             value_cache_out->data<uint8_t>(),
             reinterpret_cast<DataType_*>(qkv_out->data<T>()),
             block_tables.data<int>(),
-          batch_id_per_token.data<int>(),
             cu_seqlens_q.data<int>(),
             seq_lens.data<int>(),
             seq_lens_encoder.data<int>(),
@@ -649,7 +759,6 @@ void DecoderWriteCacheWithRoPEKernel(
             value_cache_out->data<uint8_t>(),
             reinterpret_cast<DataType_*>(qkv_out->data<T>()),
             block_tables.data<int>(),
-            batch_id_per_token.data<int>(),
             cu_seqlens_q.data<int>(),
             seq_lens.data<int>(),
             seq_lens_encoder.data<int>(),
@@ -683,7 +792,6 @@ void DecoderWriteCacheWithRoPEKernel(
           value_cache_out->data<uint8_t>(),
           reinterpret_cast<DataType_*>(qkv_out->data<T>()),
           block_tables.data<int>(),
-            batch_id_per_token.data<int>(),
           cu_seqlens_q.data<int>(),
           seq_lens.data<int>(),
           seq_lens_encoder.data<int>(),
@@ -709,6 +817,43 @@ void DecoderWriteCacheWithRoPEKernel(
           stream,
           use_neox_rotary_style,
           rope_3d);
+    } else if (cache_quant_type_str == "block_wise_fp8") {
+      constexpr int num_warps = 4;
+      const int all_warps = ((num_heads + 2 * kv_num_heads) + num_warps - 1) /
+                            num_warps * num_warps;
+      dim3 grids(bsz, all_warps / num_warps);
+      append_decode_cache_int8_rope_qk_norm_kernel<DataType_,
+                                                   4,
+                                                   0,
+                                                   128,
+                                                   false,
+                                                   true>
+          <<<grids, num_warps * 32, 0, stream>>>(
+              reinterpret_cast<const DataType_*>(qkv_ptr),
+              key_cache_out->data<uint8_t>(),
+              value_cache_out->data<uint8_t>(),
+              reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+              block_tables.data<int>(),
+              cu_seqlens_q.data<int>(),
+              seq_lens.data<int>(),
+              seq_lens_encoder.data<int>(),
+              cos_emb,
+              sin_emb,
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  cache_k_scale.get().data<T>())),
+              const_cast<DataType_*>(reinterpret_cast<const DataType_*>(
+                  (cache_v_scale.get().data<T>()))),
+              nullptr,
+              nullptr,
+              max_seq_len,
+              max_blocks_per_seq,
+              num_heads,
+              block_size,
+              127.0f,
+              -127.0f,
+              kv_num_heads,
+              rope_3d,
+              rms_norm_eps);
     } else if (cache_quant_type_str == "cache_int4_zp") {
       append_decode_cache_int4_rope(
           reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
@@ -716,7 +861,6 @@ void DecoderWriteCacheWithRoPEKernel(
           value_cache_out->data<uint8_t>(),
           reinterpret_cast<DataType_*>(const_cast<T*>(qkv_out->data<T>())),
           block_tables.data<int>(),
-          batch_id_per_token.data<int>(),
           cu_seqlens_q.data<int>(),
           seq_lens.data<int>(),
           seq_lens_encoder.data<int>(),
@@ -756,7 +900,6 @@ void DecoderWriteCacheWithRoPEKernel(
   }
 }
 
-
 template void DecoderWriteCacheWithRoPEKernel<paddle::bfloat16, int>(
     const AppendAttnMetaData& meta_data,
     const paddle::Tensor&
@@ -764,7 +907,6 @@ template void DecoderWriteCacheWithRoPEKernel<paddle::bfloat16, int>(
               // kv_num_heads, head_dim] if GQA)
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,
@@ -794,7 +936,6 @@ DecoderWriteCacheWithRoPEKernel<paddle::bfloat16, paddle::bfloat16>(
               // kv_num_heads, head_dim] if GQA)
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,
@@ -823,7 +964,6 @@ template void DecoderWriteCacheWithRoPEKernel<paddle::float16, int>(
               // kv_num_heads, head_dim] if GQA)
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,
@@ -852,7 +992,6 @@ template void DecoderWriteCacheWithRoPEKernel<paddle::float16, paddle::float16>(
               // kv_num_heads, head_dim] if GQA)
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,

custom_ops/gpu_ops/append_attn/decoder_write_cache_with_rope_kernel.h

@@ -23,7 +23,6 @@ void DecoderWriteCacheWithRoPEKernel(
               // kv_num_heads, head_dim] if GQA)
     const paddle::Tensor& seq_lens,
     const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
     const paddle::Tensor& cu_seqlens_q,
     const paddle::Tensor& block_tables,
     const paddle::optional<paddle::Tensor>& rotary_embs,

custom_ops/gpu_ops/append_attn/encoder_write_cache_with_rope_impl.cuh

@@ -449,8 +449,8 @@ __global__ void GQAVariableLengthRotaryQKNormKernel(
   const int half_lastdim = last_dim / 2;
   const int offset = (q_num_head + kv_num_head) * last_dim;
   const int all_head_num = elem_cnt / last_dim;
-  for (int gloabl_hi = global_warp_idx; gloabl_hi < all_head_num; gloabl_hi += all_warp_num) {
-    int64_t linear_index = gloabl_hi * last_dim + threadIdx.x * VecSize;
+  for (int global_hi = global_warp_idx; global_hi < all_head_num; global_hi += all_warp_num) {
+    int64_t linear_index = global_hi * last_dim + threadIdx.x * VecSize;
     const int token_idx = linear_index / offset;
     const int ori_bi = batch_id_per_token[token_idx];
     if (seq_lens[ori_bi] == 0) continue;
@@ -900,6 +900,74 @@ __global__ void GQANeoxVariableLengthRotaryKernel(
   }
 }
 
+template <typename T, int VecSize = 1>
+__global__ void GQANeoxVariableLengthPartialRotaryKernel(
+    const T *qkv,
+    const float *cos_emb,
+    const float *sin_emb,
+    const int *batch_id_per_token,
+    const int *cu_seqlens_q,
+    const int *seq_lens,
+    const int *seq_lens_decoder,
+    const float *qkv_out_scales,
+    const T *qkv_biases,
+    T *qkv_out,
+    const int64_t elem_cnt,
+    const int q_num_head,
+    const int kv_num_head,
+    const int seq_len,
+    const int head_dim,
+    const int rotary_dim,
+    const bool rope_3d) {
+  using LoadT = AlignedVector<T, VecSize>;
+  using LoadEmbT = AlignedVector<float, VecSize>;
+  LoadT left_vec;
+  LoadT right_vec;
+  LoadEmbT cos_emb_vec;
+  LoadEmbT sin_emb_vec;
+  int64_t global_thread_idx = blockDim.x * blockIdx.x + threadIdx.x;
+  const int rotary_dim_half = rotary_dim / 2;
+  const int offset = (q_num_head + kv_num_head) * rotary_dim_half;
+  for (int64_t linear_index = global_thread_idx * VecSize,
+               step = gridDim.x * blockDim.x * VecSize;
+       linear_index < elem_cnt;
+       linear_index += step) {
+    const int token_idx = linear_index / offset;
+    const int ori_bi = batch_id_per_token[token_idx];
+    if (seq_lens && seq_lens[ori_bi] == 0) continue;
+    const int bias = linear_index % offset;
+    const int hi = bias / rotary_dim_half;
+    const int h_bias = bias % rotary_dim_half;
+
+    const int ori_seq_id = (token_idx - cu_seqlens_q[ori_bi]) + seq_lens_decoder[ori_bi];
+
+    const int emb_idx = ori_seq_id * rotary_dim_half + h_bias;
+    int64_t new_emb_idx = rope_3d ? emb_idx + ori_bi * head_dim * seq_len * 2 : emb_idx;
+    const int base_idx_left =
+        token_idx * (q_num_head + 2 * kv_num_head) * head_dim + hi * head_dim +
+        h_bias;
+    const int base_idx_right = base_idx_left + rotary_dim_half;
+
+    Load<T, VecSize>(&qkv[base_idx_left], &left_vec);
+    Load<T, VecSize>(&qkv[base_idx_right], &right_vec);
+    Load<float, VecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+    Load<float, VecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
+#pragma unroll
+    for (int i = 0; i < VecSize; i++) {
+      const float input_left = static_cast<float>(left_vec[i]);
+      const float input_right = static_cast<float>(right_vec[i]);
+      const float cos_tmp = cos_emb_vec[i];
+      const float sin_tmp = sin_emb_vec[i];
+      left_vec[i] =
+          static_cast<T>(input_left * cos_tmp - input_right * sin_tmp);
+      right_vec[i] =
+          static_cast<T>(input_right * cos_tmp + input_left * sin_tmp);
+    }
+    Store<T, VecSize>(left_vec, &qkv_out[base_idx_left]);
+    Store<T, VecSize>(right_vec, &qkv_out[base_idx_right]);
+  }
+}
+
 template <typename T, int VecSize = 1>
 __global__ void cache_kernel(
     const T *__restrict__ qkv,    // [num_tokens, num_heads + 2 * kv_num_heads,
@@ -936,7 +1004,8 @@ __global__ void cache_kernel(
     const uint32_t qkv_bias = bias % hidden_size;
     const uint32_t hi = qkv_bias / head_size;
     const uint32_t h_bias = qkv_bias % head_size;
-    const uint32_t ori_bi = batch_id_per_token[token_idx];
+    const int32_t ori_bi = batch_id_per_token[token_idx];
+    if (ori_bi == -1) continue;  // skip batch_id_per_token[token_idx]=-1
     if (seq_lens[ori_bi] == 0) continue;
     const uint32_t ori_seq_id = (token_idx - cu_seqlens_q[ori_bi]) + seq_lens_decoder[ori_bi];
 
@@ -1232,6 +1301,411 @@ __global__ void append_write_cache_kv_c8_qkv(
   }
 }
 
+template <typename T,
+          uint32_t num_frags_y,
+          uint32_t num_frags_z,
+          uint32_t HEAD_DIM,
+          uint32_t BLOCK_SIZE,
+          uint32_t NUM_WARPS,
+          bool is_need_kv_quant,
+          bool IsFP8 = true>
+__global__ void append_write_cache_kv_c8_qkv_dynamic(
+    uint8_t *__restrict__ cache_k,
+    uint8_t *__restrict__ cache_v,
+    const T *__restrict__ qkv_input,
+    T *__restrict__ cache_k_scales, // [block_num, num_heads, block_size]
+    T *__restrict__ cache_v_scales, // [block_num, num_heads, block_size]
+    const int *__restrict__ batch_ids,
+    const int *__restrict__ tile_ids,
+    const int *__restrict__ seq_lens_this_time,
+    const int *__restrict__ seq_lens_decoder,
+    const int *__restrict__ batch_id_per_token,
+    const int *__restrict__ cu_seqlens_q,
+    const int *__restrict__ block_tables,
+    const int max_seq_len,
+    const int max_blocks_per_seq,
+    const int num_heads,
+    const int kv_num_heads) {
+  constexpr uint32_t num_vecs_per_head = HEAD_DIM / num_elems_per_128b<T>();
+  constexpr uint32_t pad_len = BLOCK_SIZE;
+  const uint32_t btid = blockIdx.x, kv_head_idx = blockIdx.z;
+  const T cache_k_scale = cache_k_scales[kv_head_idx];
+  const T cache_v_scale = cache_v_scales[kv_head_idx];
+  const uint32_t tid = threadIdx.x, wid = threadIdx.y;
+  const uint32_t batch_id = batch_ids[btid];
+  const uint32_t tile_id = tile_ids[btid];
+  const uint32_t seq_len_this_time = seq_lens_this_time[batch_id];
+  if (seq_len_this_time <= 0) {
+    return;
+  }
+  const int *block_table_now = nullptr;
+
+  block_table_now = block_tables + batch_id * max_blocks_per_seq;
+
+  const uint32_t num_rows_per_block =
+      NUM_WARPS * num_frags_z * 16;  // BLOCK_SIZE
+  const uint32_t start_len = seq_lens_decoder[batch_id];
+  const uint32_t bf_pad_len = start_len % pad_len;
+  const uint32_t start_len_pad = start_len - bf_pad_len;
+  const uint32_t end_len = start_len + seq_len_this_time;
+
+  const uint32_t tile_start = start_len_pad + tile_id * num_rows_per_block;
+  int block_id = __ldg(&block_table_now[tile_start / BLOCK_SIZE]);
+  uint32_t chunk_start = tile_start + wid * num_frags_z * 16 + tid / 8;
+
+  const uint32_t start_token_idx = cu_seqlens_q[batch_id];
+  const uint32_t kv_batch_stride = (num_heads + 2 * kv_num_heads) * HEAD_DIM;
+  const uint32_t kv_h_stride = HEAD_DIM;
+  __shared__ T k_smem_ori[num_rows_per_block * HEAD_DIM];
+  __shared__ T v_smem_ori[num_rows_per_block * HEAD_DIM];
+  __shared__ T v_scale_smem[BLOCK_SIZE];
+  if (tile_start >= start_len) {
+    constexpr int KV_VEC_SIZE = 16 / sizeof(uint8_t);  // 16
+    using LoadPadKVT = AlignedVector<uint8_t, KV_VEC_SIZE>;
+    // pad zero for this kv_head_idx for this block
+    LoadPadKVT pad_cache_vec;
+    *(reinterpret_cast<uint4*>(pad_cache_vec.val)) = make_uint4(0, 0, 0, 0);
+    // reset k
+    constexpr int num_vecs_per_head_k = HEAD_DIM / KV_VEC_SIZE;
+    constexpr int num_token_each_time_k = 32 / num_vecs_per_head_k;
+    uint32_t tgt_idx =
+        (block_id * kv_num_heads + kv_head_idx) * BLOCK_SIZE * HEAD_DIM +
+        tid % num_vecs_per_head_k * KV_VEC_SIZE;
+    for (int block_i = tid / num_vecs_per_head_k;
+          block_i < BLOCK_SIZE;
+          block_i += num_token_each_time_k) {
+      Store<uint8_t, KV_VEC_SIZE>(pad_cache_vec,
+                                  &cache_k[tgt_idx + block_i * HEAD_DIM]);
+    }
+
+    // reset v
+    const int num_vecs_per_head_v = BLOCK_SIZE / KV_VEC_SIZE;
+    const int num_token_each_time_v = 32 / num_vecs_per_head_v;
+    tgt_idx =
+        (block_id * kv_num_heads + kv_head_idx) * HEAD_DIM * BLOCK_SIZE +
+        tid % num_vecs_per_head_v * KV_VEC_SIZE;
+    for (int block_i = tid / num_vecs_per_head_v; block_i < HEAD_DIM;
+          block_i += num_token_each_time_v) {
+      Store<uint8_t, KV_VEC_SIZE>(
+          pad_cache_vec, &cache_v[tgt_idx + block_i * BLOCK_SIZE]);
+    }
+  }
+  smem_t k_smem(k_smem_ori);
+  smem_t v_smem(v_smem_ori);
+
+  uint32_t kv_smem_offset_w = smem_t::get_permuted_offset<num_vecs_per_head>(
+      wid * num_frags_z * 16 + tid / 8, tid % 8);  // 4 * 8 per warp
+
+  /*
+   0 | 1
+   2 | 3
+  */
+  uint32_t k_smem_offset_r = smem_t::get_permuted_offset<num_vecs_per_head>(
+      wid * num_frags_z * 16 + 8 * (tid / 16) + tid % 8, (tid % 16) / 8);
+
+  constexpr uint32_t num_frags_v = num_frags_y / NUM_WARPS;
+  /*
+   0 | 2
+   1 | 3
+  */
+  uint32_t v_smem_offset_r = smem_t::get_permuted_offset<num_vecs_per_head>(
+      tid % 16, wid * num_frags_v * 2 + tid / 16);
+
+  // load kv gmem to smem
+  const uint32_t real_start_token_idx = start_token_idx - bf_pad_len +
+                                        tile_id * num_rows_per_block +
+                                        wid * num_frags_z * 16 + tid / 8;
+  uint32_t k_read_idx = real_start_token_idx * kv_batch_stride +
+                        (num_heads + kv_head_idx) * kv_h_stride +
+                        tid % 8 * num_elems_per_128b<T>();
+  uint32_t v_read_idx = real_start_token_idx * kv_batch_stride +
+                        (num_heads + kv_num_heads + kv_head_idx) * kv_h_stride +
+                        tid % 8 * num_elems_per_128b<T>();
+#pragma unroll
+  for (uint32_t fz = 0; fz < num_frags_z; ++fz) {
+#pragma unroll
+    for (uint32_t j = 0; j < 4; ++j) {
+#pragma unroll
+      for (uint32_t fy = 0; fy < num_frags_y / 4;
+           ++fy) {  // (num_frags_y * 16) / (8 *  num_elems_per_128b<T>())
+        if (chunk_start >= start_len && chunk_start < end_len) {
+          k_smem.load_128b_async<SharedMemFillMode::kNoFill>(
+              kv_smem_offset_w, qkv_input + k_read_idx, chunk_start < end_len);
+          v_smem.load_128b_async<SharedMemFillMode::kNoFill>(
+              kv_smem_offset_w, qkv_input + v_read_idx, chunk_start < end_len);
+        }
+        kv_smem_offset_w =
+            k_smem.advance_offset_by_column<8>(kv_smem_offset_w, fy);
+        k_read_idx += 8 * num_elems_per_128b<T>();
+        v_read_idx += 8 * num_elems_per_128b<T>();
+      }
+      kv_smem_offset_w =
+          k_smem.advance_offset_by_row<4, num_vecs_per_head>(kv_smem_offset_w) -
+          2 * num_frags_y;
+      chunk_start += 4;
+      k_read_idx +=
+          4 * kv_batch_stride - 2 * num_frags_y * num_elems_per_128b<T>();
+      v_read_idx +=
+          4 * kv_batch_stride - 2 * num_frags_y * num_elems_per_128b<T>();
+    }
+  }
+  commit_group();
+  wait_group<0>();
+  __syncthreads();
+
+  // reduce scale
+  // 16 rows per warp
+  uint32_t kv_reduce_frag[4];
+  T *kv_reduce_frag_T = reinterpret_cast<T*>(kv_reduce_frag);
+
+   T k_local_max_value[num_frags_z * 2];
+   T v_local_max_value[num_frags_z * 2];
+#pragma unroll
+  for (int i = 0; i < num_frags_z * 2; i++) {
+    k_local_max_value[i] = -INFINITY;
+  }
+#pragma unroll
+  for (int i = 0; i < num_frags_z * 2; i++) {
+    v_local_max_value[i] = -INFINITY;
+  }
+  const int num_kv_heads = gridDim.z;
+  const int scale_offset = block_id * num_kv_heads * BLOCK_SIZE + kv_head_idx * BLOCK_SIZE;
+  T *cache_k_scale_now = cache_k_scales + scale_offset;
+  T *cache_v_scale_now = cache_v_scales + scale_offset;
+  // k scale
+#pragma unroll
+  for (uint32_t fz = 0; fz < num_frags_z; ++fz) {
+#pragma unroll
+    for (uint32_t fy = 0; fy < num_frags_y; ++fy) {
+      // reduce per thread, 4 threads each row
+      k_smem.ldmatrix_m8n8x4(k_smem_offset_r, kv_reduce_frag);
+#pragma unroll
+      for (int i = 0; i < 4; i++) {
+        k_local_max_value[fz * 2] = __hmax(__habs(kv_reduce_frag_T[i]), k_local_max_value[fz * 2]);
+      }
+#pragma unroll
+      for (int i = 0; i < 4; i++) {
+        k_local_max_value[fz * 2 + 1] = __hmax(__habs(kv_reduce_frag_T[i + 4]), k_local_max_value[fz * 2 + 1]);
+      }
+      k_smem_offset_r = k_smem.advance_offset_by_column<2>(k_smem_offset_r, fy);
+    }
+    // reduce per row
+    for (int i = 0; i < 2; i++) {
+      T local_max_value = __habs(k_local_max_value[fz * 2 + i]);
+      local_max_value = __hmax(local_max_value, __shfl_xor_sync(0xffffffff, local_max_value, 2));
+      local_max_value = __hmax(local_max_value, __shfl_xor_sync(0xffffffff, local_max_value, 1));
+      // used for quant
+      k_local_max_value[fz * 2 + i] = __hdiv(448, local_max_value);
+    }
+    // store
+    if (tid % 4 == 0) {
+      const int offset_now = wid * num_frags_z * 16 + tid / 4;
+      // used for dequant
+      if (tile_start + offset_now >= start_len) {
+        if (tile_start + offset_now < end_len) {
+          cache_k_scale_now[offset_now] = __hdiv(1, k_local_max_value[fz * 2]);
+        } else {
+          cache_k_scale_now[offset_now] = 0;
+        }
+      }
+      if (tile_start + offset_now + 8 >= start_len) {
+        if (tile_start + offset_now + 8 < end_len) {
+          cache_k_scale_now[offset_now + 8] = __hdiv(1, k_local_max_value[fz * 2 + 1]);
+        } else {
+          cache_k_scale_now[offset_now + 8] = 0;
+        }
+      }
+    }
+    __syncthreads();
+    k_smem_offset_r -= 2 * num_frags_y; // num_frags_z = 1
+  }
+  // v scale
+  #pragma unroll
+  for (uint32_t fz = 0; fz < num_frags_z; ++fz) {
+#pragma unroll
+    for (uint32_t fy = 0; fy < num_frags_y; ++fy) {
+      // reduce per thread, 4 threads each row
+      v_smem.ldmatrix_m8n8x4(k_smem_offset_r, kv_reduce_frag);
+#pragma unroll
+      for (int i = 0; i < 4; i++) {
+        v_local_max_value[fz * 2] = __hmax(__habs(kv_reduce_frag_T[i]), v_local_max_value[fz * 2]);
+      }
+#pragma unroll
+      for (int i = 0; i < 4; i++) {
+        v_local_max_value[fz * 2 + 1] = __hmax(__habs(kv_reduce_frag_T[i + 4]), v_local_max_value[fz * 2 + 1]);
+      }
+      k_smem_offset_r = v_smem.advance_offset_by_column<2>(k_smem_offset_r, fy);
+    }
+    // reduce per row
+    for (int i = 0; i < 2; i++) {
+      T local_max_value = __habs(v_local_max_value[fz * 2 + i]);
+      local_max_value = __hmax(local_max_value, __shfl_xor_sync(0xffffffff, local_max_value, 2));
+      local_max_value = __hmax(local_max_value, __shfl_xor_sync(0xffffffff, local_max_value, 1));
+      v_local_max_value[fz * 2 + i] = __hdiv(448, local_max_value);
+    }
+    // store
+    if (tid % 4 == 0) {
+      const int offset_now = wid * num_frags_z * 16 + tid / 4;
+      // used for dequant
+      if (tile_start + offset_now >= start_len) {
+        if (tile_start + offset_now < end_len) {
+          cache_v_scale_now[offset_now] = __hdiv(1, v_local_max_value[fz * 2]);
+          v_scale_smem[offset_now] = v_local_max_value[fz * 2];
+        } else {
+          cache_v_scale_now[offset_now] = 0;
+          v_scale_smem[offset_now] = 0;
+        }
+      }
+      if (tile_start + offset_now + 8 >= start_len) {
+        if (tile_start + offset_now + 8 < end_len) {
+          cache_v_scale_now[offset_now + 8] = __hdiv(1, v_local_max_value[fz * 2 + 1]);
+          v_scale_smem[offset_now + 8] = v_local_max_value[fz * 2 + 1];
+        } else {
+          cache_v_scale_now[offset_now + 8] = 0;
+          v_scale_smem[offset_now + 8] = 0;
+        }
+      }
+    }
+    __syncthreads();
+    k_smem_offset_r -= 2 * num_frags_y; // num_frags_z = 1
+  }
+  __syncthreads();
+
+  // mask, quant, store
+  using LoadKVT = AlignedVector<uint8_t, 4>;
+  LoadKVT cache_vec1;
+  LoadKVT cache_vec2;
+
+  uint32_t chunk_start_k = tile_start + wid * num_frags_z * 16 + tid / 4;
+  uint32_t kv_frag[4];
+  const uint32_t write_n_stride = kv_num_heads * BLOCK_SIZE * HEAD_DIM;
+  const uint32_t write_h_stride = BLOCK_SIZE * HEAD_DIM;
+  const uint32_t write_b_stride = HEAD_DIM;
+  const uint32_t write_d_stride = BLOCK_SIZE;
+  uint32_t k_write_idx = block_id * write_n_stride +
+                         kv_head_idx * write_h_stride +
+                         (wid * num_frags_z * 16 + tid / 4) * write_b_stride +
+                         tid % 4 * 4;  // 4 * int8 = 8 * int4 = 32bit
+#pragma unroll
+  for (uint32_t fz = 0; fz < num_frags_z; ++fz) {
+    uint32_t k_write_idx_now_z = k_write_idx + fz * 16 * write_b_stride;
+#pragma unroll
+    for (uint32_t fy = 0; fy < num_frags_y; ++fy) {
+      uint32_t k_write_idx_now = k_write_idx_now_z +
+                                 fy % 2 * 8 * write_b_stride +
+                                 fy / 2 * 32;  // + fy % 2 * 16;
+      // load
+      k_smem.ldmatrix_m8n8x4(k_smem_offset_r, kv_frag);
+      // quant
+      T *k_frag_T = reinterpret_cast<T *>(kv_frag);
+      if (bf_pad_len != 0) {
+        Load<uint8_t, 4>(cache_k + k_write_idx_now, &cache_vec1);
+        Load<uint8_t, 4>(cache_k + k_write_idx_now + 16, &cache_vec2);
+      }
+#pragma unroll
+      for (uint32_t v_id = 0; v_id < 8; ++v_id) {
+        uint8_t uint_quant_value;
+        if (chunk_start_k + (v_id / 4) * 8 >= start_len &&
+            chunk_start_k + (v_id / 4) * 8 < end_len) {
+          uint_quant_value = QuantToC8<T, is_need_kv_quant, IsFP8>(k_local_max_value[fz * 2 + v_id / 4], k_frag_T[v_id], 127.0f, -127.0f);
+        } else {
+          uint_quant_value = 0;
+        }
+        if (bf_pad_len != 0) {
+          if (v_id < 4) {
+            cache_vec1[v_id] |= uint_quant_value;
+          } else {
+            cache_vec2[v_id % 4] |= uint_quant_value;
+          }
+        } else {
+          if (v_id < 4) {
+            cache_vec1[v_id] = uint_quant_value;
+          } else {
+            cache_vec2[v_id - 4] = uint_quant_value;
+          }
+        }
+      }
+      // store
+      Store<uint8_t, 4>(cache_vec1, cache_k + k_write_idx_now);
+      Store<uint8_t, 4>(cache_vec2, cache_k + k_write_idx_now + 16);
+      k_smem_offset_r = k_smem.advance_offset_by_column<2>(k_smem_offset_r, fy);
+    }
+    k_smem_offset_r =
+        k_smem.advance_offset_by_row<16, num_vecs_per_head>(k_smem_offset_r) -
+        2 * num_frags_y;
+    chunk_start_k += 16;
+  }
+
+  uint32_t chunk_start_v = tile_start + tid % 4 * 2;
+  uint32_t v_write_idx = block_id * write_n_stride +
+                         kv_head_idx * write_h_stride +
+                         (wid * num_frags_v * 16 + tid / 4) * write_d_stride +
+                         tid % 4 * 4;  // 4 * int8 = 8 * int4 = 32bit
+  const uint32_t num_frags_z_v = num_frags_z * NUM_WARPS;
+  T v_scales[num_frags_z_v * 4];
+  for (int v_i = 0; v_i < num_frags_z_v; v_i++) {
+    const int offset = v_i * 16;
+    const int t_offset = tid % 4 * 2;
+    v_scales[v_i * 4] = v_scale_smem[offset + t_offset];
+    v_scales[v_i * 4 + 1] = v_scale_smem[offset + t_offset + 1];
+    v_scales[v_i * 4 + 2] = v_scale_smem[offset + t_offset + 8];
+    v_scales[v_i * 4 + 3] = v_scale_smem[offset + t_offset + 9];
+  }
+
+#pragma unroll
+  for (uint32_t fy = 0; fy < num_frags_v; ++fy) {
+    uint32_t v_write_idx_now_v = v_write_idx + fy * 16 * write_d_stride;
+#pragma unroll
+    for (uint32_t fz = 0; fz < num_frags_z_v; ++fz) {
+      uint32_t v_write_idx_now = v_write_idx_now_v +
+                                 fz % 2 * 8 * write_d_stride +
+                                 fz / 2 * 32;  // + fz % 2 * 16;
+      // load
+      v_smem.ldmatrix_m8n8x4_trans(v_smem_offset_r, kv_frag);
+      // quant
+      T *v_frag_T = reinterpret_cast<T *>(kv_frag);
+      if (bf_pad_len != 0) {
+        Load<uint8_t, 4>(cache_v + v_write_idx_now, &cache_vec1);
+        Load<uint8_t, 4>(cache_v + v_write_idx_now + 16, &cache_vec2);
+      }
+#pragma unroll
+      for (uint32_t v_id = 0; v_id < 8; ++v_id) {
+        uint8_t uint_quant_value;
+        if (chunk_start_v + v_id % 2 + (v_id % 4) / 2 * 8 >= start_len &&
+            chunk_start_v + v_id % 2 + (v_id % 4) / 2 * 8 < end_len) {
+          uint_quant_value = QuantToC8<T, is_need_kv_quant, IsFP8>(v_scales[fz * 4 + v_id % 4], v_frag_T[v_id], 127.0f, -127.0f);
+          // store now
+        } else {
+          uint_quant_value = 0;
+        }
+        if (bf_pad_len != 0) {
+          if (v_id < 4) {
+            cache_vec1[v_id] |= uint_quant_value;
+          } else {
+            cache_vec2[v_id % 4] |= uint_quant_value;
+          }
+        } else {
+          if (v_id < 4) {
+            cache_vec1[v_id] = uint_quant_value;
+          } else {
+            cache_vec2[v_id % 4] = uint_quant_value;
+          }
+        }
+      }
+      // store
+      Store<uint8_t, 4>(cache_vec1, cache_v + v_write_idx_now);
+      Store<uint8_t, 4>(cache_vec2, cache_v + v_write_idx_now + 16);
+      chunk_start_v += 16;
+      v_smem_offset_r =
+          k_smem.advance_offset_by_row<16, num_vecs_per_head>(v_smem_offset_r);
+    }
+    v_smem_offset_r = k_smem.advance_offset_by_column<2>(
+                          v_smem_offset_r, wid * num_frags_v + fy) -
+                      16 * num_frags_z_v * num_vecs_per_head;
+    chunk_start_v -= 16 * num_frags_z_v;
+  }
+}
+
 // Write Cache KV in Append
 template <typename T,
           uint32_t num_frags_y,
@@ -1705,7 +2179,9 @@ void gqa_rotary_qk_norm_variable(
       qkv_out_scales
           ? token_num * (num_heads + 2 * kv_num_heads) * dim_head
           : token_num * (num_heads + kv_num_heads) * dim_head;  // for all q k v
-  assert(dim_head == 128 && "dim_head must be 128");
+  if (dim_head != 128) {
+    PADDLE_THROW("gqa rotary with qk norm only support head_dim=128, but got %d.", dim_head);
+  }
   constexpr int HEAD_DIM = 128;
   constexpr int PackSize = HEAD_DIM / kWarpSize;
   const int pack_num = elem_nums / PackSize;
@@ -1755,6 +2231,7 @@ void gqa_rotary_qk_variable(
     const int seq_len,
     const int input_output_len,
     const int dim_head,
+    const int rotary_dim,
     const cudaStream_t &stream,
     bool use_neox_style = false,
     bool rope_3d = false) {
@@ -1835,6 +2312,37 @@ void gqa_rotary_qk_variable(
               dim_head,
               rope_3d);
     } else {
+      if (rotary_dim < dim_head){
+        PD_CHECK((rotary_dim / 2) % PackSize == 0);
+        elem_nums =
+            qkv_out_scales
+                ? token_num * (num_heads + 2 * kv_num_heads) * rotary_dim
+                : token_num * (num_heads + kv_num_heads) * rotary_dim;  // for all q k v
+        if (use_neox_style) {
+          elem_nums /= 2;
+        }
+        const int pack_num_new = elem_nums / PackSize;
+        GetNumBlocks<128>(pack_num_new, &grid_size);
+        GQANeoxVariableLengthPartialRotaryKernel<T, PackSize>
+            <<<grid_size, blocksize, 0, stream>>>(
+                reinterpret_cast<const T *>(qkv_input),
+                cos_emb,
+                rotary_emb + input_output_len * rotary_dim / 2,
+                batch_id_per_token,
+                cu_seqlens_q,
+                seq_lens,
+                seq_lens_decoder,
+                qkv_out_scales,
+                qkv_bias,
+                qkv_out,
+                elem_nums,
+                num_heads,
+                kv_num_heads,
+                seq_len,
+                dim_head,
+                rotary_dim,
+                rope_3d);
+      }else{
         GQANeoxVariableLengthRotaryKernel<T, PackSize>
           <<<grid_size, blocksize, 0, stream>>>(
               reinterpret_cast<const T *>(qkv_input),
@@ -1855,6 +2363,7 @@ void gqa_rotary_qk_variable(
               rope_3d);
       }
     }
+  }
 }
 
 template <typename T, typename QKV_TYPE>
@@ -2006,10 +2515,11 @@ void CascadeAppendWriteCacheKVC8QKV(
     int num_blocks_x_cpu,
     int max_seq_len,
     bool is_scale_channel_wise,
-    const bool is_fp8,
+    const std::string& cache_quant_type,
     cudaStream_t &stream,
     paddle::Tensor *cache_k_out,
     paddle::Tensor *cache_v_out) {
+  using NV_TYPE = typename cascade_attn_type_traits<T>::type;
   auto max_blocks_per_seq = meta_data.max_blocks_per_seq;
   auto num_tokens = meta_data.token_nums;
   auto num_heads = meta_data.q_num_heads;
@@ -2027,6 +2537,7 @@ void CascadeAppendWriteCacheKVC8QKV(
   dim3 blocks(32, num_warps);
 
   const uint32_t smem_size = (BLOCK_SIZE * HEAD_DIM) * sizeof(T) * 2;
+  if (cache_quant_type != "block_wise_fp8") {
     auto kernel_fn = append_write_cache_kv_c8_qkv<T,
                                                   num_frags_y,
                                                   num_frags_z,
@@ -2034,7 +2545,7 @@ void CascadeAppendWriteCacheKVC8QKV(
                                                   BLOCK_SIZE,
                                                   num_warps,
                                                   true, false>;
-  if (is_fp8) {
+    if (cache_quant_type == "cache_fp8") {
       kernel_fn = append_write_cache_kv_c8_qkv<T,
                                               num_frags_y,
                                               num_frags_z,
@@ -2070,6 +2581,33 @@ void CascadeAppendWriteCacheKVC8QKV(
                                             max_blocks_per_seq,
                                             num_heads,
                                             kv_num_heads);
+  } else {
+    auto kernel_fn = append_write_cache_kv_c8_qkv_dynamic<NV_TYPE,
+                                                          num_frags_y,
+                                                          num_frags_z,
+                                                          HEAD_DIM,
+                                                          BLOCK_SIZE,
+                                                          num_warps,
+                                                          true, true>;
+    cudaFuncSetAttribute(
+        kernel_fn, cudaFuncAttributeMaxDynamicSharedMemorySize, smem_size);
+    kernel_fn<<<grids, blocks, 0, stream>>>(cache_k_out->data<uint8_t>(),
+                                            cache_v_out->data<uint8_t>(),
+                                            reinterpret_cast<const NV_TYPE*>(qkv.data<T>()),
+                                            const_cast<NV_TYPE*>(reinterpret_cast<const NV_TYPE*>(cache_k_scale.data<T>())),
+                                            const_cast<NV_TYPE*>(reinterpret_cast<const NV_TYPE*>(cache_v_scale.data<T>())),
+                                            batch_ids.data<int>(),
+                                            tile_ids_per_batch.data<int>(),
+                                            seq_lens_this_time.data<int>(),
+                                            seq_lens_decoder.data<int>(),
+                                            batch_id_per_token.data<int>(),
+                                            cu_seqlens_q.data<int>(),
+                                            block_table.data<int>(),
+                                            max_seq_len,
+                                            max_blocks_per_seq,
+                                            num_heads,
+                                            kv_num_heads);
+  }
 }
 
 template <typename T, uint32_t HEAD_DIM, uint32_t BLOCK_SIZE>

custom_ops/gpu_ops/append_attn/encoder_write_cache_with_rope_kernel.h

@@ -55,9 +55,19 @@ void EncoderWriteCacheWithRopeKernel(
   auto kv_num_heads = meta_data.kv_num_heads;
   auto head_dim = meta_data.head_dims;
   bool is_scale_channel_wise = false;
+  int rotary_dim = head_dim;
   if (cache_k_scale && cache_k_scale.get().dims()[0] == head_dim * kv_num_heads) {
     is_scale_channel_wise = true;
   }
+  if (rotary_embs){
+    rotary_dim = rotary_embs.get().dims()[rotary_embs.get().dims().size()-1] * 2;
+    if(rotary_dim < head_dim){
+      if (!use_neox_style || q_norm_weight || k_norm_weight || num_heads == kv_num_heads || is_scale_channel_wise){
+        PADDLE_THROW(phi::errors::Fatal(
+          "partial_rotary_factor < 1.0 only supports use_neox_rotary_style=True, q_norm_weight/k_norm_weight) is None, GQA and is_scale_channel_wise=false."));
+      }
+    }
+  }
 
   if (q_norm_weight && k_norm_weight) {
     if (num_heads != kv_num_heads && !is_scale_channel_wise && !use_neox_style) {
@@ -125,6 +135,7 @@ void EncoderWriteCacheWithRopeKernel(
           max_seq_len,
           rope_3d ? rotary_embs.get().dims()[3] : rotary_embs.get().dims()[2],
           head_dim,
+          rotary_dim,
           stream,
           use_neox_style,
           rope_3d);
@@ -167,7 +178,7 @@ void EncoderWriteCacheWithRopeKernel(
                                     stream,
                                     key_cache_out,
                                     value_cache_out);
-  } else if (cache_quant_type_str == "cache_int8" or cache_quant_type_str == "cache_fp8") {
+  } else if (cache_quant_type_str == "cache_int8" or cache_quant_type_str == "cache_fp8" or cache_quant_type_str == "block_wise_fp8") {
     DISPATCH_HEAD_DIM(
         head_dim, HEAD_DIM, {DISPATCH_BLOCK_SIZE(block_size, BLOCK_SIZE, {
           CascadeAppendWriteCacheKVC8QKV<T, HEAD_DIM, BLOCK_SIZE>(
@@ -187,7 +198,7 @@ void EncoderWriteCacheWithRopeKernel(
               num_blocks,
               max_seq_len,
               is_scale_channel_wise,
-              cache_quant_type_str == "cache_fp8",
+              cache_quant_type_str,
               stream,
               key_cache_out,
               value_cache_out);

custom_ops/gpu_ops/append_attn/get_block_shape_and_split_kv_block.cu

@@ -11,14 +11,17 @@
 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 // See the License for the specific language governing permissions and
 // limitations under the License.
-
+#include "cute/tensor.hpp"
 #include "helper.h"
 #include "paddle/extension.h"
+#ifndef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
 #include "paddle/phi/core/memory/memcpy.h"
+#endif
+#include "utils.cuh"
 
 template <int THREADBLOCK_SIZE>
 __global__ void
-GetMaxLenKernel(const int *seq_lens, const int *seq_lens_this_time,
+GetMaxLenKernel(const int *seq_lens_decoder, const int *seq_lens_this_time,
                 const int *seq_lens_encoder,
                 const int *seq_lens_this_time_merged,
                 const int *seq_lens_encoder_merged, const int *seq_mapping,
@@ -36,41 +39,27 @@ GetMaxLenKernel(const int *seq_lens, const int *seq_lens_this_time,
   int max_just_dec_merged_len_this_time_this_thread = 0;
   int max_system_len_this_thread = 0;
   int max_dec_len_without_system_this_thread = 0;
+  int max_len_kv_this_thread = 0;
   for (int i = tid; i < batch_size; i += blockDim.x) {
     const int seq_len_this_time = seq_lens_this_time[i];
+    const int seq_len_decoder = seq_lens_decoder[i];
     max_len_this_time_this_thread =
         max(seq_len_this_time, max_len_this_time_this_thread);
     max_len_encoder_this_thread =
         max(seq_lens_encoder[i], max_len_encoder_this_thread);
-    max_len_decoder_this_thread = max(seq_lens[i], max_len_decoder_this_thread);
+    max_len_decoder_this_thread = max(seq_len_decoder, max_len_decoder_this_thread);
     if (seq_len_this_time <= 0)
       continue;
-    const int max_just_dec_len_now = seq_lens_encoder[i] > 0 ? 0 : seq_lens[i];
+    const int max_just_dec_len_now = seq_lens_encoder[i] > 0 ? 0 : seq_len_decoder;
     max_len_this_thread =
-        max(seq_lens[i] + seq_len_this_time, max_len_this_thread);
+        max(seq_len_decoder + seq_len_this_time, max_len_this_thread);
     max_just_dec_len_this_thread =
         max(max_just_dec_len_this_thread, max_just_dec_len_now);
-    if (system_lens) {
-      const int real_bid = seq_mapping[i];
-      const int system_len_now = system_lens[real_bid];
-      max_system_len_this_thread =
-          max(max_system_len_this_thread, system_len_now);
-      max_dec_len_without_system_this_thread =
-          max(max_dec_len_without_system_this_thread,
-              max_just_dec_len_now - system_len_now);
-    }
-  }
-  if (system_lens) {
-    for (int i = tid; i < batch_size; i += blockDim.x) {
-      const int ori_seq_len_this_time = seq_lens_this_time_merged[i];
-      if (ori_seq_len_this_time <= 0)
+
+    if (seq_len_decoder == 0)
       continue;
-      const int max_just_dec_merged_len_this_time_now =
-          seq_lens_encoder_merged[i] > 0 ? 0 : ori_seq_len_this_time;
-      max_just_dec_merged_len_this_time_this_thread =
-          max(max_just_dec_merged_len_this_time_this_thread,
-              max_just_dec_merged_len_this_time_now);
-    }
+    max_len_kv_this_thread =
+        max(seq_len_this_time + seq_len_decoder, max_len_kv_this_thread);
   }
   int total_max_len_this_time =
       BlockReduce(temp_storage)
@@ -93,6 +82,8 @@ GetMaxLenKernel(const int *seq_lens, const int *seq_lens_this_time,
   int total_dec_len_without_system =
       BlockReduce(temp_storage)
           .Reduce(max_dec_len_without_system_this_thread, MaxOp<int>());
+  int total_max_len_kv =
+      BlockReduce(temp_storage).Reduce(max_len_kv_this_thread, MaxOp<int>());
   if (tid == 0) {
     max_lens[0] = total_max_len_this_time;
     max_lens[1] = total_max_len_encoder;
@@ -102,6 +93,7 @@ GetMaxLenKernel(const int *seq_lens, const int *seq_lens_this_time,
     max_lens[5] = total_just_dec_merged;
     max_lens[6] = total_system_len;
     max_lens[7] = total_dec_len_without_system;
+    max_lens[8] = total_max_len_kv;
   }
 }
 
@@ -116,29 +108,146 @@ void GetMaxLen(const paddle::Tensor &seq_lens_tensor,
       max_len_tensor.data<int>(), batch_size);
 }
 
-__global__ void split_q_block(const int *__restrict__ seq_lens_q,
+template <uint32_t config_size>
+__global__ void search_chunk_size_for_mla(
+    const int *__restrict__ seq_lens_q,
     const int *__restrict__ seq_lens_encoder,
+    const int *__restrict__ seq_lens_decoder,
+    int *__restrict__ num_blocks_x,
+    int *__restrict__ res_chunk_size,
+    const int bsz,
+    const int set_chunk_size,
+    const int block_size,
+    const int sm_cout) {
+  const uint32_t conf_id = threadIdx.x;
+  int gridx = 0;
+  if (set_chunk_size > 0 && conf_id == 0) {
+    for (uint32_t bid = 0; bid < bsz; bid++) {
+      int seq_len = seq_lens_q[bid];
+      int seq_len_encoder = seq_lens_encoder[bid];
+      int seq_len_decoder = seq_lens_decoder[bid] + seq_len;
+      if (seq_len == 0 || seq_len_encoder > 0) continue;
+
+      int loop_times;
+      loop_times = cute::ceil_div(seq_len_decoder, set_chunk_size);
+      gridx += loop_times;
+    }
+    *num_blocks_x = gridx;
+    *res_chunk_size = set_chunk_size;
+  } else if (conf_id < config_size) {
+    __shared__ int gridx_shared[config_size];
+    // chunk_size is a multiple of 64
+    const int chunk_size = block_size << conf_id;
+    for (uint32_t bid = 0; bid < bsz; bid++) {
+      int seq_len = seq_lens_q[bid];
+      int seq_len_encoder = seq_lens_encoder[bid];
+      int seq_len_decoder = seq_lens_decoder[bid] + seq_len;
+      if (seq_len == 0 || seq_len_encoder > 0) continue;
+
+      int loop_times;
+      loop_times = cute::ceil_div(seq_len_decoder, chunk_size);
+      gridx += loop_times;
+    }
+    gridx_shared[conf_id] = gridx;
+    __syncthreads();
+    if (threadIdx.x == 0) {
+      uint32_t res_id = 0;
+      uint32_t max_last_wave_block = 0;
+      for (uint32_t i = 1; i < config_size; i++) {
+          uint32_t last_wave_block = gridx_shared[i] % sm_cout;
+          if (last_wave_block >= max_last_wave_block) {
+              res_id = i;
+              max_last_wave_block = last_wave_block;
+          }
+      }
+      *num_blocks_x = gridx_shared[res_id];
+      *res_chunk_size = block_size << res_id;
+    }
+  }
+}
+
+__global__ void split_block_for_mla(const int *__restrict__ seq_lens_q,
+                                    const int *__restrict__ seq_lens_encoder,
+                                    const int *__restrict__ seq_lens_decoder,
                                     int *__restrict__ batch_ids,
                                     int *__restrict__ tile_ids_per_batch,
-                              int *__restrict__ num_blocks_x, const int bsz,
-                              const int num_rows_per_block,
-                              const int group_size) {
+                                    const int bsz,
+                                    const int chunk_size) {
   if (threadIdx.x == 0) {
-    int gridx = 0;
     int index = 0;
     for (uint32_t bid = 0; bid < bsz; bid++) {
       int seq_len = seq_lens_q[bid];
-      if (seq_lens_encoder && seq_lens_encoder[bid] > 0) {
-        seq_len = 0;
+      int seq_len_encoder = seq_lens_encoder[bid];
+      int seq_len_decoder = seq_lens_decoder[bid] + seq_len;
+
+      if (seq_len == 0) continue;
+
+      int loop_times;
+      loop_times = cute::ceil_div(seq_len_decoder, chunk_size);
+      if (seq_len_encoder > 0) {
+          loop_times = 0;
       }
-      const int loop_times = div_up(seq_len * group_size, num_rows_per_block);
       for (uint32_t tile_id = 0; tile_id < loop_times; tile_id++) {
           batch_ids[index] = bid;
           tile_ids_per_batch[index++] = tile_id;
       }
-      gridx += loop_times;
     }
-    *num_blocks_x = gridx;
+  }
+}
+
+__global__ void split_q_block(const int *__restrict__ seq_lens_q,
+                              const int *__restrict__ seq_lens_encoder,
+                              int *__restrict__ batch_ids,
+                              int *__restrict__ tile_ids_per_batch,
+                              int *__restrict__ num_blocks_x,
+                              const int bsz,
+                              const int num_rows_per_block,
+                              const int group_size) {
+  // one block one warp
+  const int lane_id = threadIdx.x % warpSize;
+  int prev_offset = 0;
+
+  // loop on warp tile：[base, base+32)
+  for (int base = 0; base < bsz; base += warpSize) {
+    const int bid = base + lane_id;
+
+    // calculate loop_times for bid
+    int loop_times = 0;
+    if (bid < bsz) {
+      int seq_len = seq_lens_q[bid];
+      if (seq_lens_encoder && seq_lens_encoder[bid] > 0) {
+        seq_len = 0;
+      }
+      loop_times = div_up(seq_len * group_size, num_rows_per_block);
+    }
+
+    // prefix sum for each lane, get the start offset in this tile
+    // inclusive scan
+    int x = loop_times;
+    for (int offset = 1; offset < warpSize; offset <<= 1) {
+      int y = __shfl_up_sync(0xffffffff, x, offset);
+      if (lane_id >= offset) x += y;
+    }
+    // exclusive prefix sum
+    int bid_offset = x - loop_times;
+    int tile_sum = __shfl_sync(0xffffffff, x, warpSize - 1);
+
+    // write batch_ids and tile_ids_per_batch
+    if (bid < bsz && loop_times > 0) {
+      int write_base = prev_offset + bid_offset;
+      for (int t = 0; t < loop_times; ++t) {
+        int pos = write_base + t;
+        batch_ids[pos] = bid;
+        tile_ids_per_batch[pos] = t;
+      }
+    }
+
+    // for next warp tile
+    prev_offset += tile_sum;
+  }
+
+  if (threadIdx.x == 0) {
+    *num_blocks_x = prev_offset;
   }
 }
 
@@ -168,37 +277,22 @@ __global__ void split_kv_block(const int *__restrict__ seq_lens_decoder,
   }
 }
 
-template <int THREADBLOCK_SIZE>
-__global__ void
-get_max_len_kv_ernel(int *max_seq_lens_out, const int *seq_lens_this_time,
-                     const int *seq_lens_decoder, const int batch_size) {
-  const int tid = threadIdx.x;
-
-  typedef cub::BlockReduce<int, THREADBLOCK_SIZE> BlockReduce;
-  __shared__ typename BlockReduce::TempStorage temp_storage;
-
-  int max_len_this_thread = 0;
-  for (int i = tid; i < batch_size; i += blockDim.x) {
-    if (seq_lens_decoder[i] == 0)
-      continue;
-    max_len_this_thread =
-        max(seq_lens_this_time[i] + seq_lens_decoder[i], max_len_this_thread);
-  }
-  int total =
-      BlockReduce(temp_storage).Reduce(max_len_this_thread, MaxOp<int>());
-  if (tid == 0) {
-    *max_seq_lens_out = total;
-  }
-}
-
-std::vector<paddle::Tensor> GetBlockShapeAndSplitKVBlock(
+void GetBlockShapeAndSplitKVBlock(
     const paddle::Tensor &seq_lens_encoder,
     const paddle::Tensor &seq_lens_decoder,
     const paddle::Tensor &seq_lens_this_time,
     paddle::Tensor &decoder_batch_ids,          // Inplace
     paddle::Tensor &decoder_tile_ids_per_batch, // Inplace
-    paddle::Tensor &decoder_num_blocks_x_cpu,   // Inplace, Pinned Memory
-    paddle::Tensor &max_len_tensor_cpu,         // Inplace, Pinned Memory
+    paddle::Tensor &decoder_num_blocks_cpu,     // Inplace, Pinned Memory
+    paddle::Tensor &decoder_num_blocks_device,  // Inplace
+    paddle::Tensor &decoder_chunk_size_device,  // Inplace
+    paddle::Tensor &max_len_tensor_cpu,         // Inplace, CPU
+    paddle::Tensor &encoder_batch_ids,          // Inplace
+    paddle::Tensor &encoder_tile_ids_per_batch, // Inplace
+    paddle::Tensor &encoder_num_blocks_x_cpu,   // Inplace, CPU
+    paddle::Tensor &kv_batch_ids,               // Inplace
+    paddle::Tensor &kv_tile_ids_per_batch,      // Inplace
+    paddle::Tensor &kv_num_blocks_x_cpu,        // Inplace, CPU
     const int encoder_block_shape_q,
     const int decoder_block_shape_q,
     const int group_size,
@@ -222,32 +316,126 @@ std::vector<paddle::Tensor> GetBlockShapeAndSplitKVBlock(
   int max_just_dec_merged_len_this_time = max_len_cpu_ptr[5];
   int max_system_len = max_len_cpu_ptr[6];
   int max_just_dec_len_without_system = max_len_cpu_ptr[7];
+  int max_kv_len_this_time = max_len_cpu_ptr[8];
 
-  paddle::Tensor encoder_batch_ids;
-  paddle::Tensor encoder_tile_ids_per_batch;
-  paddle::Tensor encoder_num_blocks_x_cpu;  /*cpu*/
-  paddle::Tensor kv_batch_ids;
-  paddle::Tensor kv_tile_ids_per_batch;
-  paddle::Tensor kv_num_blocks_x_cpu;       /*cpu*/
-  paddle::Tensor max_len_kv_cpu;            /*cpu*/
+  // decoder
+  if (max_dec_len_this_time > 0) {
 
-  auto max_len_kv =
-      GetEmptyTensor({1}, paddle::DataType::INT32, seq_lens_decoder.place());
-  get_max_len_kv_ernel<128><<<1, 128, 0, stream>>>(
-      max_len_kv.data<int>(), seq_lens_this_time.data<int>(),
-      seq_lens_decoder.data<int>(), bsz);
+    const bool mla_backend = checkAttentionBackend();
+    if (mla_backend && group_size <= 64) {
+      const int set_chunk_size = get_mla_dec_chunk_size(bsz);
 
-  max_len_kv_cpu = max_len_kv.copy_to(paddle::CPUPlace(), false);
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+          decoder_chunk_size_device.data<int>(), 64, sizeof(int32_t), stream));
 
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+          decoder_num_blocks_device.data<int>(), 0, sizeof(int32_t), stream));
+
+      int device;
+      cudaGetDevice(&device);
+      int sm_cout;
+      cudaDeviceGetAttribute(&sm_cout, cudaDevAttrMultiProcessorCount, device);
+      constexpr int config_size =
+          12;  // search space for chunk size:[64, 128, 256, ... 131072]
+
+      search_chunk_size_for_mla<config_size>
+          <<<1, 32, 0, stream>>>(seq_lens_this_time.data<int>(),
+                                 seq_lens_encoder.data<int>(),
+                                 seq_lens_decoder.data<int>(),
+                                 decoder_num_blocks_device.data<int>(),
+                                 decoder_chunk_size_device.data<int>(),
+                                 bsz,
+                                 set_chunk_size,
+                                 block_size,
+                                 sm_cout);
+
+      decoder_num_blocks_cpu.copy_(
+          decoder_num_blocks_device, decoder_num_blocks_cpu.place(), false);
+      auto decoder_chunk_size_cpu =
+          decoder_chunk_size_device.copy_to(paddle::CPUPlace(), false);
+      const int chunk_size = decoder_chunk_size_cpu.data<int>()[0];
+
+      //  NOTE: (changwenbin) When using auto_chunk,
+      // decode_max_tile_size must take into account the maximum case, where * 1024 can cover 128K.
+      // const uint32_t decoder_batch_shape = seq_lens_decoder.dims()[0] * 1024;
+
+      const uint32_t decoder_max_tile_size_per_bs_q =
+          div_up((decoder_step_token_num * group_size), decoder_block_shape_q);
+      const uint32_t decoder_batch_shape =
+          bsz * 1024 * decoder_max_tile_size_per_bs_q;
+
+      PADDLE_ENFORCE_GPU_SUCCESS(
+          cudaMemsetAsync(decoder_batch_ids.data<int>(),
+                          0,
+                          decoder_batch_shape * sizeof(int32_t),
+                          stream));
+      PADDLE_ENFORCE_GPU_SUCCESS(
+          cudaMemsetAsync(decoder_tile_ids_per_batch.data<int>(),
+                          0,
+                          decoder_batch_shape * sizeof(int32_t),
+                          stream));
+
+
+      split_block_for_mla<<<1, 32, 0, stream>>>(
+          seq_lens_this_time.data<int>(),
+          seq_lens_encoder.data<int>(),
+          seq_lens_decoder.data<int>(),
+          decoder_batch_ids.data<int>(),
+          decoder_tile_ids_per_batch.data<int>(),
+          bsz,
+          chunk_size);
+
+    } else {
+      // Note:(changwenbin)In order to adapt to cudagraph, the maximum value
+      // should be taken here
+      const uint32_t decoder_max_tile_size_per_bs_q =
+          div_up((decoder_step_token_num * group_size), decoder_block_shape_q);
+      const uint32_t decoder_batch_shape =
+          bsz * 1024 * decoder_max_tile_size_per_bs_q;
+
+      PADDLE_ENFORCE_GPU_SUCCESS(
+          cudaMemsetAsync(decoder_batch_ids.data<int>(),
+                          0,
+                          decoder_batch_shape * sizeof(int32_t),
+                          stream));
+      PADDLE_ENFORCE_GPU_SUCCESS(
+          cudaMemsetAsync(decoder_tile_ids_per_batch.data<int>(),
+                          0,
+                          decoder_batch_shape * sizeof(int32_t),
+                          stream));
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+          decoder_num_blocks_device.data<int>(), 0, sizeof(int32_t), stream));
+
+      split_q_block<<<1, 32, 0, stream>>>(
+          seq_lens_this_time.data<int>(),
+          seq_lens_encoder.data<int>(),
+          decoder_batch_ids.data<int>(),
+          decoder_tile_ids_per_batch.data<int>(),
+          decoder_num_blocks_device.data<int>(),
+          bsz,
+          decoder_block_shape_q,
+          group_size);
+
+      decoder_num_blocks_cpu.copy_(
+          decoder_num_blocks_device, decoder_num_blocks_cpu.place(), false);
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+        decoder_chunk_size_device.data<int>(), 64, sizeof(int32_t), stream));
+    }
+  } else {
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+          decoder_chunk_size_device.data<int>(), 64, sizeof(int32_t), stream));
+      PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(
+          decoder_num_blocks_device.data<int>(), 0, sizeof(int32_t), stream));
+      decoder_num_blocks_cpu.copy_(
+          decoder_num_blocks_device, decoder_num_blocks_cpu.place(), false);
+  }
+
+  // encoder
   if (max_enc_len_this_time > 0) {
-    const uint32_t max_tile_size_per_bs_kv =
-        div_up(max_enc_dec_len_this_time, block_size);
-    kv_batch_ids =
-        GetEmptyTensor({bsz * max_tile_size_per_bs_kv}, paddle::DataType::INT32,
-                       seq_lens_encoder.place());
-    kv_tile_ids_per_batch =
-        GetEmptyTensor({bsz * max_tile_size_per_bs_kv}, paddle::DataType::INT32,
-                       seq_lens_encoder.place());
+    const uint32_t max_tile_size_per_bs_kv = div_up(max_enc_dec_len_this_time, block_size);
+    const uint32_t kv_batch_shape = bsz * max_tile_size_per_bs_kv;
+    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(kv_batch_ids.data<int>(), 0, kv_batch_shape * sizeof(int32_t), stream));
+    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(kv_tile_ids_per_batch.data<int>(), 0, kv_batch_shape * sizeof(int32_t), stream));
     auto kv_num_blocks_x =
         GetEmptyTensor({1}, paddle::DataType::INT32, seq_lens_encoder.place());
 
@@ -258,16 +446,12 @@ std::vector<paddle::Tensor> GetBlockShapeAndSplitKVBlock(
         kv_tile_ids_per_batch.data<int>(), kv_num_blocks_x.data<int>(), bsz,
         block_size, block_size);
 
-    kv_num_blocks_x_cpu = kv_num_blocks_x.copy_to(paddle::CPUPlace(), false);
-
-    const uint32_t encoder_max_tile_size_per_bs_q =
-        div_up((max_enc_dec_len_this_time * group_size), encoder_block_shape_q);
-    encoder_batch_ids =
-        GetEmptyTensor({bsz * encoder_max_tile_size_per_bs_q},
-                       paddle::DataType::INT32, seq_lens_encoder.place());
-    encoder_tile_ids_per_batch =
-        GetEmptyTensor({bsz * encoder_max_tile_size_per_bs_q},
-                       paddle::DataType::INT32, seq_lens_encoder.place());
+    kv_num_blocks_x_cpu.copy_(kv_num_blocks_x, kv_num_blocks_x_cpu.place(), false);
+    // Clear buffer
+    const uint32_t encoder_max_tile_size_per_bs_q = div_up((max_enc_dec_len_this_time * group_size), encoder_block_shape_q);
+    const uint32_t encoder_batch_shape = bsz * encoder_max_tile_size_per_bs_q;
+    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(encoder_batch_ids.data<int>(), 0, encoder_batch_shape * sizeof(int32_t), stream));
+    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(encoder_tile_ids_per_batch.data<int>(), 0, encoder_batch_shape * sizeof(int32_t), stream));
     auto encoder_num_blocks_x =
         GetEmptyTensor({1}, paddle::DataType::INT32, seq_lens_encoder.place());
     split_q_block<<<1, 32, 0, stream>>>(seq_lens_encoder.data<int>(), nullptr,
@@ -275,54 +459,35 @@ std::vector<paddle::Tensor> GetBlockShapeAndSplitKVBlock(
                                         encoder_tile_ids_per_batch.data<int>(),
                                         encoder_num_blocks_x.data<int>(), bsz,
                                         encoder_block_shape_q, group_size);
-    encoder_num_blocks_x_cpu =
-        encoder_num_blocks_x.copy_to(paddle::CPUPlace(), false);
-  } else {
-    encoder_batch_ids =
-        GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
-    encoder_tile_ids_per_batch =
-        GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
-    encoder_num_blocks_x_cpu =
-        GetEmptyTensor({0}, paddle::DataType::INT32, paddle::CPUPlace());
-    kv_batch_ids =
-        GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
-    kv_tile_ids_per_batch =
-        GetEmptyTensor({0}, paddle::DataType::INT32, seq_lens_encoder.place());
-    kv_num_blocks_x_cpu =
-        GetEmptyTensor({0}, paddle::DataType::INT32, paddle::CPUPlace());
+    encoder_num_blocks_x_cpu.copy_(encoder_num_blocks_x, encoder_num_blocks_x_cpu.place(), false);
   }
 
-  if (max_just_dec_len_this_time > 0) {
-    // Clear buffer
-    const uint32_t decoder_max_tile_size_per_bs_q = div_up((decoder_step_token_num * group_size), decoder_block_shape_q);
-    const uint32_t decoder_batch_shape = bsz * decoder_max_tile_size_per_bs_q;
-    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(decoder_batch_ids.data<int>(), 0, decoder_batch_shape * sizeof(int32_t), stream));
-    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(decoder_tile_ids_per_batch.data<int>(), 0, decoder_batch_shape * sizeof(int32_t), stream));
-    PADDLE_ENFORCE_GPU_SUCCESS(cudaMemsetAsync(decoder_num_blocks_x_cpu.data<int>(), 0, sizeof(int32_t), stream));
+}
 
-    auto decoder_num_blocks_x =
-        GetEmptyTensor({1}, paddle::DataType::INT32, seq_lens_encoder.place());
-    split_q_block<<<1, 32, 0, stream>>>(
-        seq_lens_this_time.data<int>(),
-        seq_lens_encoder.data<int>(),
-        decoder_batch_ids.data<int>(),
-        decoder_tile_ids_per_batch.data<int>(),
-        decoder_num_blocks_x.data<int>(),
-        bsz,
-        decoder_block_shape_q,
-        group_size);
-    decoder_num_blocks_x_cpu.copy_(decoder_num_blocks_x, decoder_num_blocks_x_cpu.place(), false);
-  }
+std::vector<std::vector<int64_t>> GetBlockShapeAndSplitKVBlockInferShape(
+    const std::vector<int64_t> &seq_lens_encoder,
+    const std::vector<int64_t> &seq_lens_decoder,
+    const std::vector<int64_t> &seq_lens_this_time,
+    const int encoder_block_shape_q,
+    const int decoder_block_shape_q,
+    const int group_size,
+    const int block_size,
+    const int decoder_step_token_num
+) {
+  return {};
+}
 
-  return {
-    encoder_batch_ids,
-    encoder_tile_ids_per_batch,
-    encoder_num_blocks_x_cpu, /*cpu*/
-    kv_batch_ids,
-    kv_tile_ids_per_batch,
-    kv_num_blocks_x_cpu,      /*cpu*/
-    max_len_kv_cpu,           /*cpu*/
-  };
+std::vector<paddle::DataType> GetBlockShapeAndSplitKVBlockInferDtype(
+    const paddle::DataType &seq_lens_encoder,
+    const paddle::DataType &seq_lens_decoder,
+    const paddle::DataType &seq_lens_this_time,
+    const int encoder_block_shape_q,
+    const int decoder_block_shape_q,
+    const int group_size,
+    const int block_size,
+    const int decoder_step_token_num
+) {
+  return {};
 }
 
 PD_BUILD_STATIC_OP(get_block_shape_and_split_kv_block)
@@ -332,17 +497,19 @@ PD_BUILD_STATIC_OP(get_block_shape_and_split_kv_block)
       "seq_lens_this_time",
       "decoder_batch_ids",
       "decoder_tile_ids_per_batch",
-      "decoder_num_blocks_x_cpu",
-      "max_len_tensor_cpu"
+      "decoder_num_blocks_cpu",
+      "decoder_num_blocks_device",
+      "decoder_chunk_size_device",
+      "max_len_tensor_cpu",
+      "encoder_batch_ids",
+      "encoder_tile_ids_per_batch",
+      "encoder_num_blocks_x_cpu",
+      "kv_batch_ids",
+      "kv_tile_ids_per_batch",
+      "kv_num_blocks_x_cpu",
     })
     .Outputs({
-      paddle::Optional("encoder_batch_ids"),
-      paddle::Optional("encoder_tile_ids_per_batch"),
-      paddle::Optional("encoder_num_blocks_x_cpu"),
-      paddle::Optional("kv_batch_ids"),
-      paddle::Optional("kv_tile_ids_per_batch"),
-      paddle::Optional("kv_num_blocks_x_cpu"),
-      "max_len_kv_cpu"
+
     })
     .Attrs({
       "encoder_block_shape_q: int",
@@ -351,4 +518,6 @@ PD_BUILD_STATIC_OP(get_block_shape_and_split_kv_block)
       "block_size: int",
       "decoder_step_token_num: int"
     })
-    .SetKernelFn(PD_KERNEL(GetBlockShapeAndSplitKVBlock));
+    .SetKernelFn(PD_KERNEL(GetBlockShapeAndSplitKVBlock))
+    .SetInferShapeFn(PD_INFER_SHAPE(GetBlockShapeAndSplitKVBlockInferShape))
+    .SetInferDtypeFn(PD_INFER_DTYPE(GetBlockShapeAndSplitKVBlockInferDtype));

custom_ops/gpu_ops/append_attn/gqa_rope_write_cache.cu

@@ -217,7 +217,7 @@ __global__ void append_cache_kv_c16(
 
   // load k_smem 64 rows 128 cols
   for (int fz = 0; fz < 4; fz++) { // 4 rows pre warp once, 16 rows all 4 warps once, need 4 iter
-    for (int fy = 0; fy < 2; fy++) { // 8 * 128b = 64 * bf16 noce, need 2 iter
+    for (int fy = 0; fy < 2; fy++) { // 8 * 128b = 64 * bf16 once, need 2 iter
       k_smem.load_128b_async<SharedMemFillMode::kNoFill>(
             k_smem_offset_w, cur_cache_k + k_read_idx, end_idx > 0);
       k_smem_offset_w =
@@ -235,7 +235,7 @@ __global__ void append_cache_kv_c16(
   // deal k_smem 64 rows 128 cols
   for (int fz = 0; fz < 1; fz++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 1 iter
     uint32_t row_idx = wid * 16 + tid / 4;
-    for (int fy = 0; fy < 8; fy++) { // 2 * 128b = 16 * bf16 noce, need 8 iter
+    for (int fy = 0; fy < 8; fy++) { // 2 * 128b = 16 * bf16 once, need 8 iter
       uint32_t col_idx = fy * 16 + tid % 4 * 2;
       k_smem.ldmatrix_m8n8x4(k_smem_offset_r, kv_frag);
       // layout
@@ -278,7 +278,7 @@ __global__ void append_cache_kv_c16(
 
   // load v_smem 64 rows 128 cols
   for (int fz = 0; fz < 4; fz++) { // // 4 rows pre warp once, 16 rows all 4 warps once, need 4 iter
-    for (int fy = 0; fy < 2; fy++) { // 8 * 128b = 64 * bf16 noce, need 2 iter
+    for (int fy = 0; fy < 2; fy++) { // 8 * 128b = 64 * bf16 once, need 2 iter
       v_smem.load_128b_async<SharedMemFillMode::kNoFill>(
             v_smem_offset_w, cur_cache_v + v_read_idx, end_idx > 0);
       v_smem_offset_w =
@@ -296,7 +296,7 @@ __global__ void append_cache_kv_c16(
   // deal v_smem 64 rows 128 cols
   for (int fz = 0; fz < 1; fz++) { //  16 rows pre warp once, 64 rows all 4 warps once, need 1 iter
     uint32_t row_idx = wid * 16 + tid / 4;
-    for (int fy = 0; fy < 8; fy++) { // 2 * 128b = 16 * bf16 noce, need 8 iter
+    for (int fy = 0; fy < 8; fy++) { // 2 * 128b = 16 * bf16 once, need 8 iter
       uint32_t col_idx = fy * 16 + tid % 4 * 2;
       v_smem.ldmatrix_m8n8x4(v_smem_offset_r, kv_frag);
       // layout
@@ -400,7 +400,7 @@ __global__ void append_cache_kv_c8(
 
   // load v_smem 64 rows, 128 cols
   for (int fz = 0; fz < 4; fz++) { // 4 rows pre warp once, 16 rows all 4 warps once, need 4 iter
-    for (int fy = 0; fy < 1; fy++) { // 8 * 128b = 128 * uint8 noce, need 1 iter
+    for (int fy = 0; fy < 1; fy++) { // 8 * 128b = 128 * uint8 once, need 1 iter
       k_smem.load_128b_async<SharedMemFillMode::kNoFill>(
             k_smem_offset_w, cur_cache_k + k_read_idx, end_idx > 0);
       k_smem_offset_w =
@@ -418,7 +418,7 @@ __global__ void append_cache_kv_c8(
   // deal k_smem 64 rows, 128 cols
   for (int fz = 0; fz < 1; fz++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 1 iter
     uint32_t row_idx = wid * 16 + tid / 4;
-    for (int fy = 0; fy < 4; fy++) { // 2 * 128b = 32 * uint8 noce, need 4 iter
+    for (int fy = 0; fy < 4; fy++) { // 2 * 128b = 32 * uint8 once, need 4 iter
       uint32_t col_idx = fy * 32 + tid % 4 * 2;
       k_smem.ldmatrix_m8n8x4(k_smem_offset_r, k_frag);
       // layout
@@ -466,7 +466,7 @@ __global__ void append_cache_kv_c8(
                           tid % 4 * num_elems_per_128b<CacheT>();
   // load v_smem 128 rows 64 cols
   for (int fy = 0; fy < 4; fy++) { // 8 rows pre warp once, 32 rows all 4 warps once, need 4 iter
-    for (int fz = 0; fz < 1; fz++) { // 4 * 128b = 64 * uint8 noce, need 1 iter
+    for (int fz = 0; fz < 1; fz++) { // 4 * 128b = 64 * uint8 once, need 1 iter
       v_smem.load_128b_async<SharedMemFillMode::kNoFill>(
               v_smem_offset_w, cur_cache_v + v_read_idx, end_idx > 0);
       v_smem_offset_w =
@@ -485,7 +485,7 @@ __global__ void append_cache_kv_c8(
   // deal v_smem 128 rows 64 cols
   for (int fy = 0; fy < 2; fy++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 2 iter
     uint32_t dim_idx = fy * NUM_WARPS * 16 + wid * 16 + tid / 4;
-    for (int fz = 0; fz < 2; fz++) { // 2 * 128b = 32 * uint8 noce, need 2 iter
+    for (int fz = 0; fz < 2; fz++) { // 2 * 128b = 32 * uint8 once, need 2 iter
       uint32_t kv_idx = fz * 32 + tid % 4 * 2;
       v_smem.ldmatrix_m8n8x4(v_smem_offset_r, v_frag);
       // layout
@@ -614,7 +614,7 @@ __global__ void append_cache_kv_c4(
 
   // load k_smem 64 rows 128 cols
   for (int fz = 0; fz < 2; fz++) { // 4 rows pre warp once, 16 rows all 4 warps once, need 4 iter
-    for (int fy = 0; fy < 1; fy++) { // 4 * 128b = 128 * int4 noce, need 1 iter
+    for (int fy = 0; fy < 1; fy++) { // 4 * 128b = 128 * int4 once, need 1 iter
       k_smem.load_128b_async<SharedMemFillMode::kNoFill>(
             k_smem_offset_w, cur_cache_k + k_read_idx, end_idx > 0);
       k_smem_offset_w =
@@ -632,7 +632,7 @@ __global__ void append_cache_kv_c4(
   // deal k_smem 64 rows 128 cols
   for (int fz = 0; fz < 1; fz++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 1 iter
     uint32_t row_idx = wid * 16 + tid / 4;
-    for (int fy = 0; fy < 2; fy++) { // 2 * 128b = 64 * int4 noce, need 2 iter
+    for (int fy = 0; fy < 2; fy++) { // 2 * 128b = 64 * int4 once, need 2 iter
       uint32_t col_idx = fy * 64 + tid % 4 * 2;
       k_smem.ldmatrix_m8n8x4(k_smem_offset_r, k_frag);
 
@@ -685,7 +685,7 @@ __global__ void append_cache_kv_c4(
                           tid % 2 * num_elems_per_128b<CacheT>();
   // load v_smem 128 rows 64 rows
   for (int fy = 0; fy < 2; fy++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 2 iter
-    for (int fz = 0; fz < 1; fz++) { // 2 * 128b = 64 * int4 noce, need 1 iter
+    for (int fz = 0; fz < 1; fz++) { // 2 * 128b = 64 * int4 once, need 1 iter
       v_smem.load_128b_async<SharedMemFillMode::kNoFill>(
               v_smem_offset_w, cur_cache_v + v_read_idx, end_idx > 0);
       v_smem_offset_w =
@@ -704,7 +704,7 @@ __global__ void append_cache_kv_c4(
   // deal v_smem 128 rows 64 cols
   for (int fy = 0; fy < 2; fy++) { // 16 rows pre warp once, 64 rows all 4 warps once, need 2 iter
     uint32_t dim_idx = fy * NUM_WARPS * 16 + wid * 16 + tid / 4;
-    for (int fz = 0; fz < 1; fz++) { // 2 * 128b = 64 * int4 noce, need 1 iter
+    for (int fz = 0; fz < 1; fz++) { // 2 * 128b = 64 * int4 once, need 1 iter
       uint32_t kv_idx = fz * 64 + tid % 4 * 2;
       v_smem.ldmatrix_m8n8x4(v_smem_offset_r, v_frag);
       // layout
@@ -1000,7 +1000,7 @@ std::vector<paddle::Tensor> GQARopeWriteCacheKernel(
       stream,
       const_cast<paddle::Tensor*>(&key_cache),
       const_cast<paddle::Tensor*>(&value_cache));
-  } else if (cache_quant_type == "cache_int8" || cache_quant_type == "cache_fp8") {
+  } else if (cache_quant_type == "cache_int8" || cache_quant_type == "cache_fp8" || cache_quant_type == "block_wise_fp8") {
     CascadeAppendWriteCacheKVC8QKV<data_t, 128, 64>(
         meta_data,
         *const_cast<paddle::Tensor*>(&key_cache),
@@ -1018,7 +1018,7 @@ std::vector<paddle::Tensor> GQARopeWriteCacheKernel(
         kv_num_blocks_data,
         max_seq_len,
         false, // is_scale_channel_wise
-        cache_quant_type == "cache_fp8", // is_fp8
+        cache_quant_type,
         stream,
         const_cast<paddle::Tensor*>(&key_cache),
         const_cast<paddle::Tensor*>(&value_cache));

custom_ops/gpu_ops/append_attn/mem_util.cuh

@@ -15,6 +15,7 @@
 
 #include <cuda_runtime.h>
 #include <stdint.h>
+#include <cooperative_groups/memcpy_async.h>
 
 enum class SharedMemFillMode { kFillZero, kNoFill };
 
@@ -42,18 +43,35 @@ __device__ __forceinline__ void ldmatrix_m8n8x4_trans_impl(uint32_t* R,
 }
 
 __device__ __forceinline__ void commit_group() {
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  {}
+#else
   asm volatile("cp.async.commit_group;\n" ::);
+#endif
 }
 
 template <size_t n>
 __device__ __forceinline__ void wait_group() {
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  cooperative_groups::wait(cooperative_groups::this_thread_block());
+#else
   asm volatile("cp.async.wait_group %0;\n" ::"n"(n));
+#endif
 }
 
 template <PrefetchMode prefetch_mode, typename T>
 __device__ __forceinline__ void load_128b(T* smem_ptr, const T* gmem_ptr) {
   uint32_t smem_int_ptr =
       static_cast<uint32_t>(__cvta_generic_to_shared(smem_ptr));
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  if constexpr (prefetch_mode == PrefetchMode::kPrefetch) {
+    memset(__cvta_shared_to_generic(smem_int_ptr), 0, 16);
+    memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 16);
+  } else {
+    memset(__cvta_shared_to_generic(smem_int_ptr), 0, 16);
+    memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 16);
+  }
+#else
   if constexpr (prefetch_mode == PrefetchMode::kPrefetch) {
     asm volatile(
         "cp.async.cg.shared.global.L2::128B [%0], [%1], %2, %3;\n" ::"r"(
@@ -68,6 +86,7 @@ __device__ __forceinline__ void load_128b(T* smem_ptr, const T* gmem_ptr) {
         "n"(16),
         "r"(16));
   }
+#endif
 }
 
 template <PrefetchMode prefetch_mode, SharedMemFillMode fill_mode, typename T>
@@ -76,6 +95,28 @@ __device__ __forceinline__ void pred_load_128b(T* smem_ptr,
                                                bool predicate) {
   uint32_t smem_int_ptr =
       static_cast<uint32_t>(__cvta_generic_to_shared(smem_ptr));
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
+    int src_in_bytes = predicate ? 16 : 0;
+    if constexpr (prefetch_mode == PrefetchMode::kPrefetch) {
+      memset(__cvta_shared_to_generic(smem_int_ptr), 0, 16);
+      memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, src_in_bytes);
+    } else {
+      memset(__cvta_shared_to_generic(smem_int_ptr), 0, 16);
+      memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, src_in_bytes);
+    }
+  } else {
+    if constexpr (prefetch_mode == PrefetchMode::kPrefetch) {
+      if (predicate) {
+        memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 16);
+      }
+    } else {
+      if (predicate) {
+        memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 16);
+      }
+    }
+  }
+#else
   if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
     int src_in_bytes = predicate ? 16 : 0;
     if constexpr (prefetch_mode == PrefetchMode::kPrefetch) {
@@ -115,6 +156,7 @@ __device__ __forceinline__ void pred_load_128b(T* smem_ptr,
           "n"(16));
     }
   }
+#endif
 }
 
 template <PrefetchMode prefetch_mode, SharedMemFillMode fill_mode, typename T>
@@ -123,6 +165,17 @@ __device__ __forceinline__ void pred_load_64b(T* smem_ptr,
                                               bool predicate) {
   uint32_t smem_int_ptr =
       static_cast<uint32_t>(__cvta_generic_to_shared(smem_ptr));
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
+    int src_in_bytes = predicate ? 8 : 0;
+    memset(__cvta_shared_to_generic(smem_int_ptr), 0, 8);
+    memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, src_in_bytes);
+  } else {
+    if (predicate) {
+      memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 8);
+    }
+  }
+#else
   if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
     int src_in_bytes = predicate ? 8 : 0;
     asm volatile(
@@ -141,6 +194,7 @@ __device__ __forceinline__ void pred_load_64b(T* smem_ptr,
         "l"(gmem_ptr),
         "n"(8));
   }
+#endif
 }
 
 template <PrefetchMode prefetch_mode, SharedMemFillMode fill_mode, typename T>
@@ -149,6 +203,17 @@ __device__ __forceinline__ void pred_load_32b(T* smem_ptr,
                                               bool predicate) {
   uint32_t smem_int_ptr =
       static_cast<uint32_t>(__cvta_generic_to_shared(smem_ptr));
+#ifdef PADDLE_WITH_CUSTOM_DEVICE_METAX_GPU
+  if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
+    int src_in_bytes = predicate ? 4 : 0;
+    memset(__cvta_shared_to_generic(smem_int_ptr), 0, 4);
+    memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, src_in_bytes);
+  } else {
+    if (predicate) {
+      memcpy(__cvta_shared_to_generic(smem_int_ptr), (void *)gmem_ptr, 4);
+    }
+  }
+#else
   if constexpr (fill_mode == SharedMemFillMode::kFillZero) {
     int src_in_bytes = predicate ? 4 : 0;
     asm volatile(
@@ -167,6 +232,7 @@ __device__ __forceinline__ void pred_load_32b(T* smem_ptr,
         "l"(gmem_ptr),
         "n"(4));
   }
+#endif
 }
 
 template <size_t num_bits, PrefetchMode prefetch_mode, typename T>

custom_ops/gpu_ops/append_attn/multi_head_latent_attention_kernel.h

@@ -1,4 +1,4 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
+// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -12,8 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 #pragma once
+
 #include "helper.h"
 #include "utils.cuh"
+#include "multiquery_decoder_attention_impl.cuh"
 
 template <typename T>
 void DecodeMLAAttentionKernel(
@@ -31,6 +33,71 @@ void DecodeMLAAttentionKernel(
   const paddle::Tensor &block_table,
   int max_seq_len,
   int max_dec_len,
+  float softmax_scale,
+  float in_scale,
+  bool causal,
+  cudaStream_t &stream,
+  paddle::Tensor *out) {
+  const auto token_num = meta_data.token_nums;
+  const auto block_size = meta_data.block_size;
+  const auto bsz = meta_data.batch_size;
+  const auto num_heads = meta_data.q_num_heads;
+  const auto group_size = meta_data.q_num_heads / meta_data.kv_num_heads;
+  const auto head_dim_qk = meta_data.head_dims;
+  const auto head_dim_v = meta_data.head_dims_v;
+  const float rope_scale = 0.0;
+  const float rope_theta = 0.0;
+  const uint32_t deal_each_time = get_cascade_attention_deal_each_time();
+  const uint32_t num_stage = get_cascade_attention_num_stages();
+  const uint32_t num_threads = get_cascade_attention_num_threads();
+
+  DISPATCH_CAUSAL(causal, CAUSAL,
+    {DISPATCH_MLA_GROUP_SIZE(group_size, GROUP_SIZE,
+      {DISPATCH_MLA_HEAD_DIM(head_dim_qk, HEAD_DIM_QK,
+        {DISPATCH_MLA_HEAD_DIM(head_dim_v, HEAD_DIM_V,
+          {DISPATCH_BLOCK_SIZE(block_size, BLOCK_SIZE,
+              {DISPATCH_DEAL_EACH_TIME(deal_each_time, DEAL_EACH_TIME,
+                  {MultiQueryDecoderAttention<T, GROUP_SIZE, HEAD_DIM_QK, HEAD_DIM_V, BLOCK_SIZE, CAUSAL, 2, 16, DEAL_EACH_TIME>(
+                  meta_data, stream, q, cache_k, cache_v, attn_mask, shift_bias, smooth_weight, seq_lens_q, seq_lens_kv, batch_id_per_token, cu_seqlens_q,
+                  block_table, max_seq_len, max_dec_len, rope_scale, rope_theta, softmax_scale, in_scale, out);})})})})})});
+}
+
+template void DecodeMLAAttentionKernel<paddle::bfloat16>(
+  const AppendAttnMetaData& meta_data,
+  const paddle::Tensor &q, // [token_num, num_heads, head_dim]
+  const paddle::Tensor &cache_k,
+  const paddle::Tensor &cache_v,
+  const paddle::optional<paddle::Tensor>& attn_mask,
+  const paddle::optional<paddle::Tensor>& shift_bias,
+  const paddle::optional<paddle::Tensor>& smooth_weight,
+  const paddle::Tensor &seq_lens_q, // q_seq_len is 1
+  const paddle::Tensor &seq_lens_kv,
+  const paddle::Tensor &batch_id_per_token,
+  const paddle::Tensor &cu_seqlens_q,
+  const paddle::Tensor &block_table,
+  int max_seq_len,
+  int max_dec_len,
+  float softmax_scale,
+  float in_scale,
+  bool causal,
+  cudaStream_t &stream,
+  paddle::Tensor *out);
+
+template void DecodeMLAAttentionKernel<paddle::float16>(
+  const AppendAttnMetaData& meta_data,
+  const paddle::Tensor &q, // [token_num, num_heads, head_dim]
+  const paddle::Tensor &cache_k,
+  const paddle::Tensor &cache_v,
+  const paddle::optional<paddle::Tensor>& attn_mask,
+  const paddle::optional<paddle::Tensor>& shift_bias,
+  const paddle::optional<paddle::Tensor>& smooth_weight,
+  const paddle::Tensor &seq_lens_q, // q_seq_len is 1
+  const paddle::Tensor &seq_lens_kv,
+  const paddle::Tensor &batch_id_per_token,
+  const paddle::Tensor &cu_seqlens_q,
+  const paddle::Tensor &block_table,
+  int max_seq_len,
+  int max_dec_len,
  float softmax_scale,
   float in_scale,
   bool causal,

custom_ops/gpu_ops/append_attn/multiquery_attention_c16_kernel.h

@@ -0,0 +1,56 @@
+// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "append_attention_func.cuh"
+
+template <typename T,
+          uint32_t GROUP_SIZE,
+          uint32_t HEAD_DIM,
+          uint32_t BLOCK_SIZE,
+          bool CAUSAL,
+          uint32_t BLOCK_SHAPE_Q,
+          uint32_t NUM_WARP_Q,
+          typename OutT,
+          bool ENABLE_PREFILL = true>
+void MultiQueryAppendAttention(
+    const AppendAttnMetaData &meta_data,
+    const paddle::Tensor &qkv,
+    const paddle::Tensor &cache_k,
+    const paddle::Tensor &cache_v,
+    const paddle::optional<paddle::Tensor> &attn_mask,
+    const paddle::optional<paddle::Tensor> &shift_bias,
+    const paddle::optional<paddle::Tensor> &smooth_weight,
+    const paddle::optional<paddle::Tensor> &sinks,
+    const paddle::Tensor &seq_lens_q,
+    const paddle::Tensor &seq_lens_kv,
+    const paddle::Tensor &seq_lens_encoder,
+    const paddle::Tensor &batch_id_per_token,
+    const paddle::Tensor &cu_seqlens_q,
+    const paddle::Tensor &block_table,
+    const paddle::Tensor &batch_ids,
+    const paddle::Tensor &tile_ids_per_batch,
+    const int num_blocks_x_cpu,
+    const int max_seq_len,
+    const int max_dec_len,
+    const float quant_max_bound,
+    const float quant_min_bound,
+    const float in_scale,
+    const int max_partition_size,
+    const int encoder_max_partition_size,
+    const int speculate_max_draft_token_num,
+    const bool is_decoder,
+    cudaStream_t &stream,
+    paddle::Tensor *out,
+    const int sliding_window);

custom_ops/gpu_ops/append_attn/multiquery_attention_c4_kernel.h

@@ -0,0 +1,60 @@
+// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "append_attention_func.cuh"
+
+template <typename T,
+          uint32_t GROUP_SIZE,
+          uint32_t HEAD_DIM,
+          uint32_t BLOCK_SIZE,
+          bool CAUSAL,
+          uint32_t BLOCK_SHAPE_Q,
+          uint32_t NUM_WARP_Q,
+          typename OutT = T,
+          bool ENABLE_PREFILL = true>
+void MultiQueryAppendC4Attention(
+    const AppendAttnMetaData &meta_data,
+    const paddle::Tensor &qkv,
+    const paddle::Tensor &cache_k,
+    const paddle::Tensor &cache_v,
+    const paddle::optional<paddle::Tensor> &attn_mask,
+    const paddle::Tensor &cache_k_scale,
+    const paddle::Tensor &cache_v_scale,
+    const paddle::optional<paddle::Tensor> &cache_k_zp,
+    const paddle::optional<paddle::Tensor> &cache_v_zp,
+    const paddle::optional<paddle::Tensor> &shift_bias,
+    const paddle::optional<paddle::Tensor> &smooth_weight,
+    const paddle::optional<paddle::Tensor> &sinks,
+    const paddle::Tensor &seq_lens_q,
+    const paddle::Tensor &seq_lens_kv,
+    const paddle::Tensor &seq_lens_encoder,
+    const paddle::Tensor &batch_id_per_token,
+    const paddle::Tensor &cu_seqlens_q,
+    const paddle::Tensor &block_table,
+    const paddle::Tensor &batch_ids,
+    const paddle::Tensor &tile_ids_per_batch,
+    const int num_blocks_x_cpu,
+    const int max_seq_len,
+    const int max_dec_len,
+    const float quant_max_bound,
+    const float quant_min_bound,
+    const float in_scale,
+    const int max_partition_size,
+    const int encoder_max_partition_size,
+    const int speculate_max_draft_token_num,
+    const bool is_decoder,
+    cudaStream_t &stream,
+    paddle::Tensor *out,
+    const int sliding_window);

custom_ops/gpu_ops/append_attn/multiquery_attention_c8_kernel.h

@@ -0,0 +1,60 @@
+// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "append_attention_func.cuh"
+
+template <typename T,
+          uint32_t GROUP_SIZE,
+          uint32_t HEAD_DIM,
+          uint32_t BLOCK_SIZE,
+          bool CAUSAL,
+          uint32_t BLOCK_SHAPE_Q,
+          uint32_t NUM_WARP_Q,
+          typename OutT = T,
+          bool ENABLE_PREFILL = true,
+          bool IsFP8 = false,
+          bool IsDynamicC8 = false>
+void MultiQueryAppendC8Attention(
+    const AppendAttnMetaData &meta_data,
+    const paddle::Tensor &qkv,
+    const paddle::Tensor &cache_k,
+    const paddle::Tensor &cache_v,
+    const paddle::optional<paddle::Tensor> &attn_mask,
+    const paddle::Tensor &cache_k_scale,
+    const paddle::Tensor &cache_v_scale,
+    const paddle::optional<paddle::Tensor> &shift_bias,
+    const paddle::optional<paddle::Tensor> &smooth_weight,
+    const paddle::optional<paddle::Tensor> &sinks,
+    const paddle::Tensor &seq_lens_q,
+    const paddle::Tensor &seq_lens_kv,
+    const paddle::Tensor &seq_lens_encoder,
+    const paddle::Tensor &batch_id_per_token,
+    const paddle::Tensor &cu_seqlens_q,
+    const paddle::Tensor &block_table,
+    const paddle::Tensor &batch_ids,
+    const paddle::Tensor &tile_ids_per_batch,
+    const int num_blocks_x_cpu,
+    const int max_seq_len,
+    const int max_dec_len,
+    const float quant_max_bound,
+    const float quant_min_bound,
+    const float in_scale,
+    const int max_partition_size,
+    const int encoder_max_partition_size,
+    const int speculate_max_draft_token_num,
+    const bool is_decoder,
+    cudaStream_t &stream,
+    paddle::Tensor *out,
+    const int sliding_window);

custom_ops/gpu_ops/append_attn/multiquery_decoder_attention_impl.cuh

@@ -11,8 +11,10 @@
 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 // See the License for the specific language governing permissions and
 // limitations under the License.
+#pragma once
 
 #include "decode_attention_func.cuh"
+#include "multiquery_decoder_attention_kernel.h"
 
 #define CHECK(call)                                                             \
 do                                                                              \
@@ -471,90 +473,3 @@ void MultiQueryDecoderAttention(
   // CHECK(cudaGetLastError());
   // CHECK(cudaDeviceSynchronize());
 }
-
-template <typename T>
-void DecodeMLAAttentionKernel(
-  const AppendAttnMetaData& meta_data,
-  const paddle::Tensor &q, // [token_num, num_heads, head_dim]
-  const paddle::Tensor &cache_k,
-  const paddle::Tensor &cache_v,
-  const paddle::optional<paddle::Tensor>& attn_mask,
-  const paddle::optional<paddle::Tensor>& shift_bias,
-  const paddle::optional<paddle::Tensor>& smooth_weight,
-  const paddle::Tensor &seq_lens_q, // q_seq_len is 1
-  const paddle::Tensor &seq_lens_kv,
-  const paddle::Tensor &batch_id_per_token,
-  const paddle::Tensor &cu_seqlens_q,
-  const paddle::Tensor &block_table,
-  int max_seq_len,
-  int max_dec_len,
-  float softmax_scale,
-  float in_scale,
-  bool causal,
-  cudaStream_t &stream,
-  paddle::Tensor *out) {
-  const auto token_num = meta_data.token_nums;
-  const auto block_size = meta_data.block_size;
-  const auto bsz = meta_data.batch_size;
-  const auto num_heads = meta_data.q_num_heads;
-  const auto group_size = meta_data.q_num_heads / meta_data.kv_num_heads;
-  const auto head_dim_qk = meta_data.head_dims;
-  const auto head_dim_v = meta_data.head_dims_v;
-  const float rope_scale = 0.0;
-  const float rope_theta = 0.0;
-  const uint32_t deal_each_time = get_cascade_attention_deal_each_time();
-  const uint32_t num_stage = get_cascade_attention_num_stages();
-  const uint32_t num_threads = get_cascade_attention_num_threads();
-
-  DISPATCH_CAUSAL(causal, CAUSAL,
-    {DISPATCH_MLA_GROUP_SIZE(group_size, GROUP_SIZE,
-      {DISPATCH_MLA_HEAD_DIM(head_dim_qk, HEAD_DIM_QK,
-        {DISPATCH_MLA_HEAD_DIM(head_dim_v, HEAD_DIM_V,
-          {DISPATCH_BLOCK_SIZE(block_size, BLOCK_SIZE,
-              {DISPATCH_DEAL_EACH_TIME(deal_each_time, DEAL_EACH_TIME,
-                  {MultiQueryDecoderAttention<T, GROUP_SIZE, HEAD_DIM_QK, HEAD_DIM_V, BLOCK_SIZE, CAUSAL, 2, 16, DEAL_EACH_TIME>(
-                  meta_data, stream, q, cache_k, cache_v, attn_mask, shift_bias, smooth_weight, seq_lens_q, seq_lens_kv, batch_id_per_token, cu_seqlens_q,
-                  block_table, max_seq_len, max_dec_len, rope_scale, rope_theta, softmax_scale, in_scale, out);})})})})})});
-}
-
-template void DecodeMLAAttentionKernel<paddle::bfloat16>(
-  const AppendAttnMetaData& meta_data,
-  const paddle::Tensor &q, // [token_num, num_heads, head_dim]
-  const paddle::Tensor &cache_k,
-  const paddle::Tensor &cache_v,
-  const paddle::optional<paddle::Tensor>& attn_mask,
-  const paddle::optional<paddle::Tensor>& shift_bias,
-  const paddle::optional<paddle::Tensor>& smooth_weight,
-  const paddle::Tensor &seq_lens_q, // q_seq_len is 1
-  const paddle::Tensor &seq_lens_kv,
-  const paddle::Tensor &batch_id_per_token,
-  const paddle::Tensor &cu_seqlens_q,
-  const paddle::Tensor &block_table,
-  int max_seq_len,
-  int max_dec_len,
-  float softmax_scale,
-  float in_scale,
-  bool causal,
-  cudaStream_t &stream,
-  paddle::Tensor *out);
-
-template void DecodeMLAAttentionKernel<paddle::float16>(
-  const AppendAttnMetaData& meta_data,
-  const paddle::Tensor &q, // [token_num, num_heads, head_dim]
-  const paddle::Tensor &cache_k,
-  const paddle::Tensor &cache_v,
-  const paddle::optional<paddle::Tensor>& attn_mask,
-  const paddle::optional<paddle::Tensor>& shift_bias,
-  const paddle::optional<paddle::Tensor>& smooth_weight,
-  const paddle::Tensor &seq_lens_q, // q_seq_len is 1
-  const paddle::Tensor &seq_lens_kv,
-  const paddle::Tensor &batch_id_per_token,
-  const paddle::Tensor &cu_seqlens_q,
-  const paddle::Tensor &block_table,
-  int max_seq_len,
-  int max_dec_len,
- float softmax_scale,
-  float in_scale,
-  bool causal,
-  cudaStream_t &stream,
-  paddle::Tensor *out);

custom_ops/gpu_ops/append_attn/multiquery_decoder_attention_kernel.h

@@ -0,0 +1,39 @@
+// Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "decode_attention_func.cuh"
+
+template <typename T, uint32_t GROUP_SIZE, uint32_t HEAD_DIM_QK, uint32_t HEAD_DIM_V, uint32_t BLOCK_SIZE, bool CAUSAL, uint32_t NUM_STAGE, uint32_t cache_bytes, uint32_t DEAL_EACH_TIME>
+void MultiQueryDecoderAttention(
+  const AppendAttnMetaData& meta_data,
+  cudaStream_t &stream,
+  const paddle::Tensor &q,
+  const paddle::Tensor &cache_k, // [max_block_num, num_kv_heads, block_size, head_dim]
+  const paddle::Tensor &cache_v, // [num_kv_heads, head_dim]
+  const paddle::optional<paddle::Tensor>& attn_mask,
+  const paddle::optional<paddle::Tensor>& shift_bias,
+  const paddle::optional<paddle::Tensor>& smooth_weight,
+  const paddle::Tensor &seq_lens_q,
+  const paddle::Tensor &seq_lens_kv,
+  const paddle::Tensor &batch_id_per_token,
+  const paddle::Tensor &cu_seqlens_q,
+  const paddle::Tensor &block_table,
+  const int max_seq_len,
+  const int max_dec_len,
+  const float rope_scale,
+  const float rope_theta,
+  const float softmax_scale,
+  const float in_scale,
+  paddle::Tensor *out);

custom_ops/gpu_ops/append_attn/speculate_write_cache_with_rope_impl.cuh

@@ -18,6 +18,167 @@
 #include "mma_tensor_op.cuh"
 #include "utils.cuh"
 
+template <typename T, int VecSize = 1, typename InT = T>
+__global__ void append_speculate_cache_T_rope_qk_norm_kernel(
+    const InT* __restrict__ qkv,  // [token_num, num_heads + 2 * gqa_group_size,
+                                  // head_size]
+    T* __restrict__ key_cache,    // [num_blocks, gqa_group_size, block_size,
+                                  // head_size // 2]
+    T* __restrict__ value_cache,  // [num_blocks, gqa_group_size, block_size,
+                                  // head_size // 2]
+    T* __restrict__ q_out,
+    const int* __restrict__ block_tables,        // [bsz, max_blocks_per_seq]
+    const int* __restrict__ batch_id_per_token,  // [num_tokens]
+    const int* __restrict__ cu_seqlens_q,
+    const int* __restrict__ seq_lens_decoder,  // [bsz]
+    const float* __restrict__ cos_emb,
+    const float* __restrict__ sin_emb,
+    const float*
+        qkv_out_scales,   // [(num_heads + 2 * gqa_group_size) * head_size]
+    const T* qkv_biases,  // [num_head + 2 * gqa_group_size, dim_head]
+    const int max_seq_len,
+    const int max_blocks_per_seq,
+    const int num_heads,
+    const int output_inner_dim,
+    const int head_size,
+    const int block_size,
+    const int elem_cnt,
+    const int gqa_group_size,
+    const float* q_norm_weight,
+    const float* k_norm_weight,
+    const float rms_norm_eps,
+    const bool rope_3d) {
+  using LoadT = AlignedVector<T, VecSize>;
+  using LoadFloat = AlignedVector<float, VecSize>;
+  using LoadInT = AlignedVector<InT, VecSize>;
+  constexpr int HalfVecSize = VecSize / 2;
+  using LoadEmbT = AlignedVector<float, HalfVecSize>;
+  LoadInT src_vec;
+  LoadFloat scale_vec;
+  LoadT bias_vec;
+  LoadEmbT cos_emb_vec;
+  LoadEmbT sin_emb_vec;
+  LoadFloat tmp_vec;
+  LoadFloat q_norm_vec;
+  LoadFloat k_norm_vec;
+
+  int64_t global_warp_idx = blockDim.y * blockIdx.x + threadIdx.y;
+  int64_t all_warp_num = gridDim.x * blockDim.y;
+  int64_t all_head_dim = elem_cnt / head_size;
+
+  const int64_t hidden_size = (num_heads + 2 * gqa_group_size) * head_size;
+  const int half_head_size = head_size / 2;
+  for (int global_hi = global_warp_idx; global_hi < all_head_dim;
+       global_hi += all_warp_num) {
+    int64_t linear_index = global_hi * head_size + threadIdx.x * VecSize;
+    const int token_id = linear_index / hidden_size;
+
+    const int ori_bi = batch_id_per_token[token_id];
+    if (ori_bi == -1) continue;  // NOTE(gongshaotian): For CUDAGraph padding
+    if (seq_lens_decoder[ori_bi] == 0) continue;
+    const int bias = linear_index % hidden_size;
+    const int hi = bias / head_size;  // q + k + v
+    const int h_bias = bias % head_size;
+    const int start_token_idx = cu_seqlens_q[ori_bi];
+    const int write_seq_id =
+        seq_lens_decoder[ori_bi] + token_id - start_token_idx;
+    if (write_seq_id == 0) continue;
+
+    const int* block_table_now = block_tables + ori_bi * max_blocks_per_seq;
+    const int block_idx = block_table_now[write_seq_id / block_size];
+    if (block_idx < 0) {
+      return;  // NOTE(gongshaotian): For CUDAGraph padding
+    }
+    const int block_offset = write_seq_id % block_size;
+
+    const int write_q_idx =
+        token_id * output_inner_dim * head_size + hi * head_size + h_bias;
+
+    const int bias_idx = hi * head_size + h_bias;
+    Load<InT, VecSize>(&qkv[linear_index], &src_vec);
+    if (qkv_biases) {
+      Load<T, VecSize>(&qkv_biases[bias_idx], &bias_vec);
+    }
+    if (qkv_out_scales) {
+      Load<float, VecSize>(&qkv_out_scales[bias_idx], &scale_vec);
+    }
+    if (hi < num_heads + gqa_group_size) {
+      // q k rope
+      const int64_t emb_idx = write_seq_id * half_head_size + h_bias / 2;
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + ori_bi * max_seq_len * head_size : emb_idx;
+      Load<float, HalfVecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, HalfVecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
+    }
+    float thread_m2 = 0.0f;
+    float warp_m2 = 0.0f;
+#pragma unroll
+    for (int i = 0; i < HalfVecSize; i++) {
+      // add_bias + rope
+      float input_left = static_cast<float>(src_vec[2 * i]);
+      float input_right = static_cast<float>(src_vec[2 * i + 1]);
+      if (qkv_out_scales) {
+        input_left *= scale_vec[2 * i];
+        input_right *= scale_vec[2 * i + 1];
+      }
+      if (qkv_biases) {
+        input_left = input_left + static_cast<float>(bias_vec[2 * i]);
+        input_right = input_right + static_cast<float>(bias_vec[2 * i + 1]);
+      }
+      if (hi < num_heads + gqa_group_size) {
+        const float cos_tmp = cos_emb_vec[i];
+        const float sin_tmp = sin_emb_vec[i];
+        float tmp1 = input_left * cos_tmp - input_right * sin_tmp;
+        float tmp2 = input_right * cos_tmp + input_left * sin_tmp;
+        thread_m2 += tmp1 * tmp1 + tmp2 * tmp2;
+        tmp_vec[2 * i] = tmp1;
+        tmp_vec[2 * i + 1] = tmp2;
+      } else {
+        bias_vec[2 * i] = static_cast<T>(input_left);
+        bias_vec[2 * i + 1] = static_cast<T>(input_right);
+      }
+    }
+    if (hi < (num_heads + gqa_group_size)) {
+      WelfordWarpAllReduce<float, 32>(thread_m2, &warp_m2);
+      float row_variance = max(warp_m2 / head_size, 0.0f);
+      float row_inv_var = Rsqrt(row_variance + rms_norm_eps);
+      if (hi < num_heads) {
+        Load<float, VecSize>(&q_norm_weight[threadIdx.x * VecSize],
+                             &q_norm_vec);
+#pragma unroll
+        for (int i = 0; i < VecSize; i++) {
+          bias_vec[i] =
+              static_cast<T>(tmp_vec[i] * row_inv_var * q_norm_vec[i]);
+        }
+      } else {
+        Load<float, VecSize>(&k_norm_weight[threadIdx.x * VecSize],
+                             &k_norm_vec);
+#pragma unroll
+        for (int i = 0; i < VecSize; i++) {
+          bias_vec[i] =
+              static_cast<T>(tmp_vec[i] * row_inv_var * k_norm_vec[i]);
+        }
+      }
+    }
+    if (hi < num_heads) {
+      // write q
+      Store<T, VecSize>(bias_vec, &q_out[write_q_idx]);
+    } else {
+      //  write k/v
+      const int kv_head_idx = (hi - num_heads) % gqa_group_size;
+      const int tgt_idx = (block_idx * gqa_group_size * block_size * head_size +
+                           kv_head_idx * block_size * head_size +
+                           block_offset * head_size + h_bias);
+      // write
+      if (hi < num_heads + gqa_group_size) {
+        Store<T, VecSize>(bias_vec, &key_cache[tgt_idx]);
+      } else {
+        Store<T, VecSize>(bias_vec, &value_cache[tgt_idx]);
+      }
+    }
+  }
+}
+
 template <int VecSize = 4, int HeadDim = 128>
 __global__ void append_clear_cache_int8_block(
     uint8_t* __restrict__ key_cache,    // [num_blocks, gqa_group_size,
@@ -43,6 +204,7 @@ __global__ void append_clear_cache_int8_block(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -91,7 +253,6 @@ __global__ void append_clear_cache_int8_block(
   }
 }
 
-
 template <int VecSize = 4, int HeadDim = 128>
 __global__ void append_clear_cache_int4_block(
     uint8_t* __restrict__ key_cache,    // [num_blocks, gqa_group_size,
@@ -117,6 +278,7 @@ __global__ void append_clear_cache_int4_block(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -193,7 +355,8 @@ __global__ void append_speculate_cache_rope_kernel(
     const int head_size,
     const int block_size,
     const int elem_cnt,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   using LoadT = AlignedVector<T, VecSize>;
   using LoadFloat = AlignedVector<float, VecSize>;
   using LoadInT = AlignedVector<InT, VecSize>;
@@ -215,6 +378,8 @@ __global__ void append_speculate_cache_rope_kernel(
        linear_index += step) {
     const int token_id = linear_index / hidden_size;
     const int ori_bi = batch_id_per_token[token_id];
+    if (ori_bi == -1) continue;  // NOTE(gongshaotian): For CUDAGraph padding
+
     if (seq_lens_decoder[ori_bi] == 0) continue;
     const int bias = linear_index % hidden_size;
     const int hi = bias / head_size;  // q + k + v
@@ -227,15 +392,7 @@ __global__ void append_speculate_cache_rope_kernel(
     const int* block_table_now = block_tables + ori_bi * max_blocks_per_seq;
     const int block_idx = block_table_now[write_seq_id / block_size];
     if (block_idx < 0) {
-      printf(
-          "Fatal Error!!!, block idx %d when write_seq_id is %d\n some key var "
-          "%d %d %d %d\n",
-          block_idx,
-          write_seq_id,
-          ori_bi,
-          seq_lens_decoder[ori_bi],
-          token_id,
-          cu_seqlens_q[ori_bi]);
+      return;  // NOTE(gongshaotian): For CUDAGraph padding
     }
     const int block_offset = write_seq_id % block_size;
 
@@ -253,8 +410,10 @@ __global__ void append_speculate_cache_rope_kernel(
     if (hi < num_heads + gqa_group_size) {
       // q k rope
       const int64_t emb_idx = write_seq_id * half_head_size + h_bias / 2;
-      Load<float, HalfVecSize>(&cos_emb[emb_idx], &cos_emb_vec);
-      Load<float, HalfVecSize>(&sin_emb[emb_idx], &sin_emb_vec);
+      int64_t new_emb_idx =
+          rope_3d ? emb_idx + ori_bi * max_seq_len * head_size : emb_idx;
+      Load<float, HalfVecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, HalfVecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
     }
 #pragma unroll
     for (int i = 0; i < HalfVecSize; i++) {
@@ -326,7 +485,8 @@ __global__ void append_speculate_cache_neox_rope_kernel(
     const int head_size,
     const int block_size,
     const int elem_cnt,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   using LoadT = AlignedVector<T, VecSize>;
   using LoadFloat = AlignedVector<float, VecSize>;
   using LoadInT = AlignedVector<InT, VecSize>;
@@ -348,6 +508,7 @@ __global__ void append_speculate_cache_neox_rope_kernel(
        linear_index += step) {
     const int token_id = linear_index / half_hidden_size;
     const int ori_bi = batch_id_per_token[token_id];
+    if (ori_bi == -1) continue;  // NOTE(gongshaotian): For CUDAGraph padding
     if (seq_lens_decoder[ori_bi] == 0) continue;
     const int bias = linear_index % half_hidden_size;
     const int hi = bias / half_head_size;  // q + k + v
@@ -360,15 +521,7 @@ __global__ void append_speculate_cache_neox_rope_kernel(
     const int* block_table_now = block_tables + ori_bi * max_blocks_per_seq;
     const int block_idx = block_table_now[write_seq_id / block_size];
     if (block_idx < 0) {
-      printf(
-          "Fatal Error!!!, block idx %d when write_seq_id is %d\n some key var "
-          "%d %d %d %d\n",
-          block_idx,
-          write_seq_id,
-          ori_bi,
-          seq_lens_decoder[ori_bi],
-          token_id,
-          cu_seqlens_q[ori_bi]);
+      return;  // NOTE(gongshaotian): For CUDAGraph padding
     }
     const int block_offset = write_seq_id % block_size;
 
@@ -390,8 +543,10 @@ __global__ void append_speculate_cache_neox_rope_kernel(
     if (hi < num_heads + gqa_group_size) {
       // q k rope
       const int64_t emb_idx = write_seq_id * head_size + h_bias;
-      Load<float, VecSize>(&cos_emb[emb_idx], &cos_emb_vec);
-      Load<float, VecSize>(&sin_emb[emb_idx], &sin_emb_vec);
+      int64_t new_emb_idx =
+          rope_3d ? emb_idx + ori_bi * max_seq_len * head_size * 2 : emb_idx;
+      Load<float, VecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, VecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
     }
 #pragma unroll
     for (int i = 0; i < VecSize; i++) {
@@ -443,6 +598,277 @@ __global__ void append_speculate_cache_neox_rope_kernel(
   }
 }
 
+template <typename T,
+          int VecSize = 4,
+          int RoundType = 0,
+          int HeadDim = 128,
+          bool IsFP8 = false>
+__global__ void append_speculate_cache_fp8_rope_qk_norm_dynamic_kernel(
+    const T* __restrict__ quant_qkv,    // [num_head, num_heads + 2 *
+                                        // gqa_group_size, head_size]
+    uint8_t* __restrict__ key_cache,    // [num_blocks, gqa_group_size,
+                                        // block_size, head_size // 2]
+    uint8_t* __restrict__ value_cache,  // [num_blocks, gqa_group_size,
+                                        // block_size, head_size // 2]
+    T* __restrict__ qkv_out,
+    const int* __restrict__ block_tables,        // [bsz, max_blocks_per_seq]
+    const int* __restrict__ batch_id_per_token,  // [num_tokens]
+    const int* __restrict__ cu_seqlens_q,
+    const int* __restrict__ seq_lens,          // [bsz]
+    const int* __restrict__ seq_lens_encoder,  // [bsz]
+    const float* __restrict__ cos_emb,
+    const float* __restrict__ sin_emb,
+    T* __restrict__ cache_k_scale,
+    T* __restrict__ cache_v_scale,
+    const float* q_norm_weight,
+    const float* k_norm_weight,
+    const int max_seq_len,
+    const int max_blocks_per_seq,
+    const int num_heads,
+    const int block_size,
+    const float max_bound,
+    const float min_bound,
+    const int gqa_group_size,
+    const bool rope_3d,
+    const float rms_norm_eps) {
+  static_assert(HeadDim == 128, "just support HeadDim be 128 now!");
+  static_assert(VecSize == 4, "just support VecSize be 4 now, 32 * 4!");
+  constexpr int NUM_WARPS = 4;
+  const int tid = threadIdx.x;
+  const int wid = tid / 32;
+  const int lane_id = tid % 32;
+  const int token_id = blockIdx.x;
+
+  const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
+
+  const int start_token_idx = cu_seqlens_q[bid];
+  const int head_idx = blockIdx.y * NUM_WARPS + wid;
+  int q_head_idx, k_head_idx, v_idx;
+  const int64_t hidden_size = (num_heads + 2 * gqa_group_size) * HeadDim;
+  constexpr int half_head_size = HeadDim / 2;
+  if (seq_lens_encoder[bid] > 0) return;
+  const int write_seq_id = seq_lens[bid] + token_id - start_token_idx;
+  if (write_seq_id == 0) return;
+  const int* block_table_now = block_tables + bid * max_blocks_per_seq;
+  const int block_idx = __ldg(&block_table_now[write_seq_id / block_size]);
+  const int block_offset = write_seq_id % block_size;
+
+  int cache_offset;
+  if (head_idx < num_heads) {
+    cache_offset = 0;
+  } else if (head_idx < num_heads + 2 * gqa_group_size) {
+    cache_offset = block_idx * gqa_group_size * block_size +
+                   (head_idx - num_heads) % gqa_group_size * block_size +
+                   block_offset;
+  }
+  T* cache_k_scale_now = cache_k_scale + cache_offset;
+  T* cache_v_scale_now = cache_v_scale + cache_offset;
+
+  float thread_m2 = 0.0f;
+  float warp_m2 = 0.0f;
+
+  if (head_idx < num_heads) {
+    // q
+    using LoadT = AlignedVector<T, VecSize>;
+    using LoadBiasT = AlignedVector<T, VecSize>;
+    using LoadOutScaleT = AlignedVector<float, VecSize>;
+    constexpr int HalfVecSize = VecSize / 2;
+    using LoadEmbT = AlignedVector<float, HalfVecSize>;
+
+    LoadT src_vec;
+    LoadBiasT bias_vec;
+    LoadOutScaleT out_scale_vec;
+    LoadEmbT cos_emb_vec;
+    LoadEmbT sin_emb_vec;
+    const T* qkv_now = quant_qkv + token_id * hidden_size;
+    T* qkv_out_now = qkv_out + token_id * hidden_size;
+#pragma unroll
+    for (uint32_t head_bias = lane_id * VecSize; head_bias < HeadDim;
+         head_bias += 32 * VecSize) {
+      const int bias_idx = head_idx * HeadDim + head_bias;
+      Load<T, VecSize>(&qkv_now[bias_idx], &src_vec);
+
+      // q rope
+      const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, HalfVecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, HalfVecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
+#pragma unroll
+      for (int i = 0; i < HalfVecSize; i++) {
+        // dequant + add_bias + rope
+        float input_left = static_cast<float>(src_vec[2 * i]);
+        float input_right = static_cast<float>(src_vec[2 * i + 1]);
+        const float cos_tmp = cos_emb_vec[i];
+        const float sin_tmp = sin_emb_vec[i];
+        float tmp1 = input_left * cos_tmp - input_right * sin_tmp;
+        float tmp2 = input_right * cos_tmp + input_left * sin_tmp;
+        thread_m2 += tmp1 * tmp1 + tmp2 * tmp2;
+        bias_vec[2 * i] = static_cast<T>(tmp1);
+        bias_vec[2 * i + 1] = static_cast<T>(tmp2);
+      }
+      // qk norm
+      if (q_norm_weight) {
+        WelfordWarpAllReduce<float, 32>(thread_m2, &warp_m2);
+        float row_variance = max(warp_m2 / HeadDim, 0.0f);
+        float row_inv_var = Rsqrt(row_variance + rms_norm_eps);
+        LoadOutScaleT q_norm_vec;
+        Load<float, VecSize>(&q_norm_weight[lane_id * VecSize], &q_norm_vec);
+#pragma unroll
+        for (int i = 0; i < VecSize; i++) {
+          bias_vec[i] = static_cast<T>(static_cast<float>(bias_vec[i]) *
+                                       row_inv_var * q_norm_vec[i]);
+        }
+      }
+      Store<T, VecSize>(bias_vec, &qkv_out_now[bias_idx]);
+    }
+  } else if (head_idx < num_heads + 2 * gqa_group_size) {
+    // k
+    constexpr int KV_VEC_SIZE = 16 / sizeof(uint8_t);  // 16
+    using LoadPadKVT = AlignedVector<uint8_t, KV_VEC_SIZE>;
+    const uint32_t kv_head_idx = (head_idx - num_heads) % gqa_group_size;
+
+    constexpr int K_VEC_SIZE = 4;
+    constexpr int HALF_K_VEC_SIZE = 2;
+    using LoadKVResT = AlignedVector<uint8_t, K_VEC_SIZE>;
+    using LoadKVT = AlignedVector<uint8_t, HALF_K_VEC_SIZE>;
+    using LoadT = AlignedVector<T, HALF_K_VEC_SIZE>;
+    using LoadBiasT = AlignedVector<T, HALF_K_VEC_SIZE>;
+    using LoadOutScaleT = AlignedVector<float, HALF_K_VEC_SIZE>;
+    using LoadEmbT = AlignedVector<float, 1>;
+    LoadKVResT cache_vec;
+    LoadT src_vec1, src_vec2;
+    LoadBiasT bias_vec1, bias_vec2;
+    LoadOutScaleT out_scale_vec1, out_scale_vec2;
+    LoadEmbT cos_emb_vec1, cos_emb_vec2;
+    LoadEmbT sin_emb_vec1, sin_emb_vec2;
+
+    const T* qkv_now = quant_qkv + token_id * hidden_size;
+    const int head_bias = lane_id / 4 * 16 + lane_id % 4 * 2;
+    const int bias_idx = head_idx * HeadDim + head_bias;
+    Load<T, HALF_K_VEC_SIZE>(&qkv_now[bias_idx], &src_vec1);
+    Load<T, HALF_K_VEC_SIZE>(&qkv_now[bias_idx + 8], &src_vec2);
+    T scale = T(1.0f);
+    const int k_head_idx = head_idx - num_heads;
+    const int v_head_idx = head_idx - num_heads - gqa_group_size;
+    if (head_idx < num_heads + gqa_group_size) {
+      const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, 1>(&cos_emb[new_emb_idx], &cos_emb_vec1);
+      Load<float, 1>(&cos_emb[new_emb_idx + 4], &cos_emb_vec2);
+      Load<float, 1>(&sin_emb[new_emb_idx], &sin_emb_vec1);
+      Load<float, 1>(&sin_emb[new_emb_idx + 4], &sin_emb_vec2);
+    }
+
+    float input_left = static_cast<float>(src_vec1[0]);
+    float input_right = static_cast<float>(src_vec1[1]);
+    if (head_idx < num_heads + gqa_group_size) {
+      float cos_tmp = cos_emb_vec1[0];
+      float sin_tmp = sin_emb_vec1[0];
+      float tmp1 = input_left * cos_tmp - input_right * sin_tmp;
+      float tmp2 = input_right * cos_tmp + input_left * sin_tmp;
+      thread_m2 += tmp1 * tmp1 + tmp2 * tmp2;
+      bias_vec1[0] = static_cast<T>(tmp1);
+      bias_vec1[1] = static_cast<T>(tmp2);
+    } else {
+      bias_vec1[0] = static_cast<T>(input_left);
+      bias_vec1[1] = static_cast<T>(input_right);
+    }
+
+    input_left = static_cast<float>(src_vec2[0]);
+    input_right = static_cast<float>(src_vec2[1]);
+    if (head_idx < num_heads + gqa_group_size) {
+      float cos_tmp = cos_emb_vec2[0];
+      float sin_tmp = sin_emb_vec2[0];
+      float tmp1 = input_left * cos_tmp - input_right * sin_tmp;
+      float tmp2 = input_right * cos_tmp + input_left * sin_tmp;
+      thread_m2 += tmp1 * tmp1 + tmp2 * tmp2;
+      bias_vec2[0] = static_cast<T>(tmp1);
+      bias_vec2[1] = static_cast<T>(tmp2);
+    } else {
+      bias_vec2[0] = static_cast<T>(input_left);
+      bias_vec2[1] = static_cast<T>(input_right);
+    }
+    if (k_norm_weight) {
+      if (head_idx < num_heads + gqa_group_size) {
+        LoadOutScaleT k_norm_vec1, k_norm_vec2;
+        Load<float, HALF_K_VEC_SIZE>(&k_norm_weight[head_bias], &k_norm_vec1);
+        Load<float, HALF_K_VEC_SIZE>(&k_norm_weight[head_bias + 8],
+                                     &k_norm_vec2);
+        // qk norm
+        WelfordWarpAllReduce<float, 32>(thread_m2, &warp_m2);
+        float row_variance = max(warp_m2 / HeadDim, 0.0f);
+        float row_inv_var = Rsqrt(row_variance + rms_norm_eps);
+
+        for (int i = 0; i < HALF_K_VEC_SIZE; i++) {
+          bias_vec1[i] = static_cast<T>(static_cast<float>(bias_vec1[i]) *
+                                        row_inv_var * k_norm_vec1[i]);
+          bias_vec2[i] = static_cast<T>(static_cast<float>(bias_vec2[i]) *
+                                        row_inv_var * k_norm_vec2[i]);
+        }
+      }
+    }
+    // reduce max, 1 head per warp
+    T local_max = -INFINITY;
+#pragma unroll
+    for (int i = 0; i < HALF_K_VEC_SIZE; i++) {
+      local_max = __hmax(local_max, __habs(bias_vec1[i]));
+      local_max = __hmax(local_max, __habs(bias_vec2[i]));
+    }
+#pragma unroll
+    for (int m_offset = 16; m_offset > 0; m_offset /= 2) {
+      local_max =
+          __hmax(local_max, __shfl_xor_sync(0xffffffff, local_max, m_offset));
+    }
+
+    scale = __hdiv(448, local_max);
+
+    if (lane_id == 0) {
+      if (head_idx < num_heads + gqa_group_size) {
+        cache_k_scale_now[0] = __hdiv(1, scale);
+      } else {
+        cache_v_scale_now[0] = __hdiv(1, scale);
+      }
+    }
+
+#pragma unroll
+    for (uint32_t i = 0; i < HALF_K_VEC_SIZE; i++) {
+      cache_vec[i] = QuantToC8<T, true, IsFP8, RoundType>(
+          scale, bias_vec1[i], max_bound, min_bound);
+      cache_vec[i + HALF_K_VEC_SIZE] = QuantToC8<T, true, IsFP8, RoundType>(
+          scale, bias_vec2[i], max_bound, min_bound);
+    }
+    if (head_idx < num_heads + gqa_group_size) {
+      const int start_block_16 =
+          block_offset / 16 * 16 + block_offset % 8 + lane_id / 4 % 2 * 8;
+      const uint32_t tgt_cache_idx =
+          block_idx * gqa_group_size * block_size * HeadDim +
+          kv_head_idx * block_size * HeadDim + start_block_16 * HeadDim +
+          lane_id / 4 / 2 * 32 + (block_offset % 16) / 8 * 16 + lane_id % 4 * 4;
+      Store<uint8_t, K_VEC_SIZE>(cache_vec, &key_cache[tgt_cache_idx]);
+    } else {
+      const uint32_t base_tgt_cache_idx =
+          block_idx * gqa_group_size * HeadDim * block_size +
+          kv_head_idx * HeadDim * block_size +
+          (lane_id / 4 * 16 + lane_id % 4 * 2) * block_size +
+          block_offset / 16 % 2 * 8 * block_size + block_offset / 16 / 2 * 32;
+      const uint32_t tgt_cache_idx1 = base_tgt_cache_idx +
+                                      block_offset % 8 / 2 * 4     // per 4
+                                      + block_offset % 16 / 8 * 2  // per 2
+                                      + block_offset % 2;          // per 1
+      const uint32_t tgt_cache_idx2 = tgt_cache_idx1 + block_size;
+      const uint32_t tgt_cache_idx3 = tgt_cache_idx1 + 16;
+      const uint32_t tgt_cache_idx4 = tgt_cache_idx3 + block_size;
+      value_cache[tgt_cache_idx1] = cache_vec[0];
+      value_cache[tgt_cache_idx2] = cache_vec[1];
+      value_cache[tgt_cache_idx3] = cache_vec[2];
+      value_cache[tgt_cache_idx4] = cache_vec[3];
+    }
+  }
+}
+
 template <typename T,
           int VecSize = 4,
           int RoundType = 0,
@@ -476,7 +902,8 @@ __global__ void append_speculate_cache_int8_rope_kernel(
     const int block_size,
     const float max_bound,
     const float min_bound,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   static_assert(HeadDim == 128, "just support HeadDim be 128 now!");
   static_assert(VecSize == 4, "just support VecSize be 4 now, 32 * 4!");
   constexpr int NUM_WARPS = 4;
@@ -486,6 +913,7 @@ __global__ void append_speculate_cache_int8_rope_kernel(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -522,8 +950,10 @@ __global__ void append_speculate_cache_int8_rope_kernel(
 
       // q rope
       const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
-      Load<float, HalfVecSize>(&cos_emb[emb_idx], &cos_emb_vec);
-      Load<float, HalfVecSize>(&sin_emb[emb_idx], &sin_emb_vec);
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, HalfVecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, HalfVecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
       if (qkv_out_scales) {
         Load<float, VecSize>(&qkv_out_scales[bias_idx], &out_scale_vec);
       }
@@ -583,10 +1013,12 @@ __global__ void append_speculate_cache_int8_rope_kernel(
     T scale;
     if (head_idx < num_heads + gqa_group_size) {
       const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
-      Load<float, 1>(&cos_emb[emb_idx], &cos_emb_vec1);
-      Load<float, 1>(&cos_emb[emb_idx + 4], &cos_emb_vec2);
-      Load<float, 1>(&sin_emb[emb_idx], &sin_emb_vec1);
-      Load<float, 1>(&sin_emb[emb_idx + 4], &sin_emb_vec2);
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, 1>(&cos_emb[new_emb_idx], &cos_emb_vec1);
+      Load<float, 1>(&cos_emb[new_emb_idx + 4], &cos_emb_vec2);
+      Load<float, 1>(&sin_emb[new_emb_idx], &sin_emb_vec1);
+      Load<float, 1>(&sin_emb[new_emb_idx + 4], &sin_emb_vec2);
       scale = __ldg(&cache_k_scales[kv_head_idx]);
     } else {
       scale = __ldg(&cache_v_scales[kv_head_idx]);
@@ -644,8 +1076,10 @@ __global__ void append_speculate_cache_int8_rope_kernel(
     }
 #pragma unroll
     for (uint32_t i = 0; i < HALF_K_VEC_SIZE; i++) {
-      cache_vec[i] = QuantToC8<T,true, IsFP8, RoundType>(scale, bias_vec1[i], max_bound, min_bound);
-      cache_vec[i + HALF_K_VEC_SIZE] = QuantToC8<T,true, IsFP8, RoundType>(scale, bias_vec2[i], max_bound, min_bound);
+      cache_vec[i] = QuantToC8<T, true, IsFP8, RoundType>(
+          scale, bias_vec1[i], max_bound, min_bound);
+      cache_vec[i + HALF_K_VEC_SIZE] = QuantToC8<T, true, IsFP8, RoundType>(
+          scale, bias_vec2[i], max_bound, min_bound);
     }
     if (head_idx < num_heads + gqa_group_size) {
       const int start_block_16 =
@@ -708,7 +1142,8 @@ __global__ void append_speculate_cache_int8_neox_rope_kernel(
     const int block_size,
     const float max_bound,
     const float min_bound,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   static_assert(HeadDim == 128, "just support HeadDim be 128 now!");
   static_assert(VecSize == 4, "just support VecSize be 4 now, 32 * 4!");
   constexpr int NUM_WARPS = 4;
@@ -718,6 +1153,7 @@ __global__ void append_speculate_cache_int8_neox_rope_kernel(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -757,8 +1193,10 @@ __global__ void append_speculate_cache_int8_neox_rope_kernel(
 
       // q rope
       const uint32_t emb_idx = write_seq_id * HeadDim + head_bias;
-      Load<float, VecSize>(&cos_emb[emb_idx], &cos_emb_vec);
-      Load<float, VecSize>(&sin_emb[emb_idx], &sin_emb_vec);
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim * 2 : emb_idx;
+      Load<float, VecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, VecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
       if (qkv_out_scales) {
         Load<float, VecSize>(&qkv_out_scales[bias_idx_left],
                              &left_out_scale_vec);
@@ -853,10 +1291,12 @@ __global__ void append_speculate_cache_int8_neox_rope_kernel(
 
         T scale;
         const uint32_t emb_idx = write_seq_id * HeadDim + head_bias;
-        Load<float, HALF_K_VEC_SIZE>(&cos_emb[emb_idx], &cos_emb_vec1);
-        Load<float, HALF_K_VEC_SIZE>(&cos_emb[emb_idx + 8], &cos_emb_vec2);
-        Load<float, HALF_K_VEC_SIZE>(&sin_emb[emb_idx], &sin_emb_vec1);
-        Load<float, HALF_K_VEC_SIZE>(&sin_emb[emb_idx + 8], &sin_emb_vec2);
+        uint32_t new_emb_idx =
+            rope_3d ? emb_idx + bid * max_seq_len * HeadDim * 2 : emb_idx;
+        Load<float, HALF_K_VEC_SIZE>(&cos_emb[new_emb_idx], &cos_emb_vec1);
+        Load<float, HALF_K_VEC_SIZE>(&cos_emb[new_emb_idx + 8], &cos_emb_vec2);
+        Load<float, HALF_K_VEC_SIZE>(&sin_emb[new_emb_idx], &sin_emb_vec1);
+        Load<float, HALF_K_VEC_SIZE>(&sin_emb[new_emb_idx + 8], &sin_emb_vec2);
         scale = __ldg(&cache_k_scales[kv_head_idx]);
 #pragma unroll
         for (int i = 0; i < HALF_K_VEC_SIZE; i++) {
@@ -1088,7 +1528,8 @@ __global__ void append_speculate_cache_int4_rope_kernel(
     const int block_size,
     const float max_bound,
     const float min_bound,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   static_assert(HeadDim == 128, "just support HeadDim be 128 now!");
   static_assert(VecSize == 4, "just support VecSize be 4 now, 32 * 4!");
   constexpr int NUM_WARPS = 4;
@@ -1099,6 +1540,7 @@ __global__ void append_speculate_cache_int4_rope_kernel(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -1145,8 +1587,10 @@ __global__ void append_speculate_cache_int4_rope_kernel(
       // Load<float, VecSize>(&qkv_out_scales[bias_idx], &out_scale_vec);
       // q rope
       const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
-      Load<float, HalfVecSize>(&cos_emb[emb_idx], &cos_emb_vec);
-      Load<float, HalfVecSize>(&sin_emb[emb_idx], &sin_emb_vec);
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, HalfVecSize>(&cos_emb[new_emb_idx], &cos_emb_vec);
+      Load<float, HalfVecSize>(&sin_emb[new_emb_idx], &sin_emb_vec);
 #pragma unroll
       for (int i = 0; i < HalfVecSize; i++) {
         // dequant + add_bias + rope
@@ -1235,10 +1679,12 @@ __global__ void append_speculate_cache_int4_rope_kernel(
     //                              &out_scale_vec2);
     if (head_idx < num_heads + gqa_group_size) {
       const uint32_t emb_idx = write_seq_id * half_head_size + head_bias / 2;
-      Load<float, 1>(&cos_emb[emb_idx], &cos_emb_vec1);
-      Load<float, 1>(&cos_emb[emb_idx + 4], &cos_emb_vec2);
-      Load<float, 1>(&sin_emb[emb_idx], &sin_emb_vec1);
-      Load<float, 1>(&sin_emb[emb_idx + 4], &sin_emb_vec2);
+      uint32_t new_emb_idx =
+          rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+      Load<float, 1>(&cos_emb[new_emb_idx], &cos_emb_vec1);
+      Load<float, 1>(&cos_emb[new_emb_idx + 4], &cos_emb_vec2);
+      Load<float, 1>(&sin_emb[new_emb_idx], &sin_emb_vec1);
+      Load<float, 1>(&sin_emb[new_emb_idx + 4], &sin_emb_vec2);
       Load<T, HALF_K_VEC_SIZE>(&cache_k_scales[cache_idx], &scale_vec1);
       Load<T, HALF_K_VEC_SIZE>(&cache_k_scales[cache_idx + 8], &scale_vec2);
       Load<T, HALF_K_VEC_SIZE>(&cache_k_zero_points[cache_idx], &zp_vec1);
@@ -1341,7 +1787,6 @@ __global__ void append_speculate_cache_int4_rope_kernel(
       }
       Store<uint8_t, K_VEC_SIZE>(cache_vec, &key_cache[tgt_cache_idx]);
     } else {
-
       const uint32_t base_tgt_cache_idx =
           block_idx * gqa_group_size * HeadDim * half_block_size +
           kv_head_idx * HeadDim * half_block_size +
@@ -1431,7 +1876,8 @@ __global__ void append_speculate_cache_int4_neox_rope_kernel(
     const int block_size,
     const float max_bound,
     const float min_bound,
-    const int gqa_group_size) {
+    const int gqa_group_size,
+    const bool rope_3d) {
   static_assert(HeadDim == 128, "just support HeadDim be 128 now!");
   static_assert(VecSize == 4, "just support VecSize be 4 now, 32 * 4!");
   constexpr int NUM_WARPS = 4;
@@ -1442,6 +1888,7 @@ __global__ void append_speculate_cache_int4_neox_rope_kernel(
   const int token_id = blockIdx.x;
 
   const int bid = batch_id_per_token[token_id];
+  if (bid == -1) return;  // NOTE(gongshaotian): For CUDAGraph padding
 
   const int start_token_idx = cu_seqlens_q[bid];
   const int head_idx = blockIdx.y * NUM_WARPS + wid;
@@ -1581,10 +2028,12 @@ __global__ void append_speculate_cache_int4_neox_rope_kernel(
                                      &right_out_scale_vec2);
 
         const uint32_t emb_idx = write_seq_id * HeadDim + head_bias;
-        Load<float, HALF_K_VEC_SIZE>(&cos_emb[emb_idx], &cos_emb_vec1);
-        Load<float, HALF_K_VEC_SIZE>(&cos_emb[emb_idx + 8], &cos_emb_vec2);
-        Load<float, HALF_K_VEC_SIZE>(&sin_emb[emb_idx], &sin_emb_vec1);
-        Load<float, HALF_K_VEC_SIZE>(&sin_emb[emb_idx + 8], &sin_emb_vec2);
+        uint32_t new_emb_idx =
+            rope_3d ? emb_idx + bid * max_seq_len * HeadDim : emb_idx;
+        Load<float, HALF_K_VEC_SIZE>(&cos_emb[new_emb_idx], &cos_emb_vec1);
+        Load<float, HALF_K_VEC_SIZE>(&cos_emb[new_emb_idx + 8], &cos_emb_vec2);
+        Load<float, HALF_K_VEC_SIZE>(&sin_emb[new_emb_idx], &sin_emb_vec1);
+        Load<float, HALF_K_VEC_SIZE>(&sin_emb[new_emb_idx + 8], &sin_emb_vec2);
         Load<T, HALF_K_VEC_SIZE>(&cache_k_scales[left_cache_idx],
                                  &left_scale_vec1);
         Load<T, HALF_K_VEC_SIZE>(&cache_k_scales[left_cache_idx + 8],
@@ -1618,7 +2067,6 @@ __global__ void append_speculate_cache_int4_neox_rope_kernel(
           right_bias_vec1[i] =
               static_cast<T>(input_right * cos_tmp + input_left * sin_tmp);
 
-
           input_left = static_cast<float>(left_src_vec2[i]);
           input_right = static_cast<float>(right_src_vec2[i]);
           cos_tmp = cos_emb_vec2[i];

custom_ops/gpu_ops/append_attn/speculate_write_cache_with_rope_kernel.cu

@@ -15,6 +15,78 @@
 #include "speculate_write_cache_with_rope_kernel.h"
 #include "utils.cuh"
 
+template <typename T, typename QKV_TYPE>
+void append_speculate_cache_rope_qk_norm(const QKV_TYPE* qkv,
+                              T* key_cache,
+                              T* value_cache,
+                              T* qkv_out,
+                              const int* block_tables,
+                              const int* batch_id_per_token,
+                              const int* cu_seqlens_q,
+                              const int* seq_lens,
+                              const int* seq_lens_encoder,
+                              const float* cos_emb,
+                              const float* sin_emb,
+                              const float* qkv_out_scales,
+                              const T* qkv_biases,
+                              const int max_seq_len,
+                              const int max_blocks_per_seq,
+                              const int num_heads,
+                              const int kv_num_heads,
+                              const int dim_head,
+                              const int block_size,
+                              const int bsz,
+                              const int token_num,
+                              const cudaStream_t& stream,
+                              const bool use_neox_style,
+                              const float* q_norm_weight,
+                              const float* k_norm_weight,
+                              const float rms_norm_eps,
+                              const bool rope_3d) {
+  int output_inner_dim = num_heads + 2 * kv_num_heads;
+  const uint32_t elem_nums =
+      use_neox_style ? token_num * (num_heads + 2 * kv_num_heads) * dim_head / 2
+                     : token_num * (num_heads + 2 * kv_num_heads) * dim_head;
+  constexpr int HEAD_DIM = 128;
+
+  constexpr int PackSize = HEAD_DIM / kWarpSize;
+  const int pack_num = elem_nums / PackSize;
+  const int blocksize = 128;
+  int grid_size = 1;
+  GetNumBlocks<128>(pack_num, &grid_size);
+  if (use_neox_style) {
+      PD_THROW(
+          "append_speculate_cache_rope_qk_norm not support neox rope yet");
+  } else {
+    dim3 block_dim(kWarpSize, blocksize / kWarpSize, 1);
+    append_speculate_cache_T_rope_qk_norm_kernel<T, PackSize>
+      <<<grid_size, block_dim, 0, stream>>>(qkv,
+                                            key_cache,
+                                            value_cache,
+                                            qkv_out,
+                                            block_tables,
+                                            batch_id_per_token,
+                                            cu_seqlens_q,
+                                            seq_lens,
+                                            cos_emb,
+                                            sin_emb,
+                                            qkv_out_scales,
+                                            qkv_biases,
+                                            max_seq_len,
+                                            max_blocks_per_seq,
+                                            num_heads,
+                                            output_inner_dim,
+                                            dim_head,
+                                            block_size,
+                                            elem_nums,
+                                            kv_num_heads,
+                                            q_norm_weight,
+                                            k_norm_weight,
+                                            rms_norm_eps,
+                                            rope_3d);
+  }
+}
+
 // rope + write
 template <typename T, typename QKV_TYPE>
 void append_speculate_cache_rope(const QKV_TYPE* qkv,
@@ -39,7 +111,8 @@ void append_speculate_cache_rope(const QKV_TYPE* qkv,
                                  const int bsz,
                                  const int token_num,
                                  const cudaStream_t& stream,
-                                 const bool use_neox_style) {
+                                 const bool use_neox_style,
+                                 const bool rope_3d) {
   int output_inner_dim = num_heads + 2 * kv_num_heads;
 
   const uint32_t elem_nums =
@@ -73,7 +146,8 @@ void append_speculate_cache_rope(const QKV_TYPE* qkv,
             dim_head,
             block_size,
             elem_nums,
-            kv_num_heads);
+            kv_num_heads,
+            rope_3d);
   } else {
     append_speculate_cache_rope_kernel<T, PackSize>
         <<<grid_size, threads_per_block, 0, stream>>>(
@@ -96,10 +170,83 @@ void append_speculate_cache_rope(const QKV_TYPE* qkv,
             dim_head,
             block_size,
             elem_nums,
-            kv_num_heads);
+            kv_num_heads,
+            rope_3d);
   }
 }
 
+template <typename T>
+void append_speculate_cache_fp8_dynamic_rope(const T* qkv,
+                                             uint8_t* key_cache,
+                                             uint8_t* value_cache,
+                                             T* qkv_out,
+                                             const int* block_tables,
+                                             const int* batch_id_per_token,
+                                             const int* cu_seqlens_q,
+                                             const int* seq_lens,
+                                             const int* seq_lens_encoder,
+                                             const float* cos_emb,
+                                             const float* sin_emb,
+                                             T* cache_k_scale,
+                                             T* cache_v_scale,
+                                             const float* q_norm_weight,
+                                             const float* k_norm_weight,
+                                             const int max_seq_len,
+                                             const int max_blocks_per_seq,
+                                             const int num_heads,
+                                             const int kv_num_heads,
+                                             const int dim_head,
+                                             const int block_size,
+                                             const int bsz,
+                                             const int token_num,
+                                             const cudaStream_t& stream,
+                                             const bool rope_3d,
+                                             const float rms_norm_eps) {
+  constexpr int num_warps = 4;
+  const int all_warps =
+      ((num_heads + 2 * kv_num_heads) + num_warps - 1) / num_warps * num_warps;
+  dim3 grids(token_num, all_warps / num_warps);
+
+  append_clear_cache_int8_block<4>
+      <<<grids, num_warps * 32, 0, stream>>>(key_cache,
+                                             value_cache,
+                                             seq_lens,
+                                             block_tables,
+                                             batch_id_per_token,
+                                             cu_seqlens_q,
+                                             seq_lens_encoder,
+                                             max_seq_len,
+                                             max_blocks_per_seq,
+                                             num_heads,
+                                             block_size,
+                                             kv_num_heads);
+  append_speculate_cache_fp8_rope_qk_norm_dynamic_kernel<T, 4, 0, 128, true>
+      <<<grids, num_warps * 32, 0, stream>>>(qkv,
+                                             key_cache,
+                                             value_cache,
+                                             qkv_out,
+                                             block_tables,
+                                             batch_id_per_token,
+                                             cu_seqlens_q,
+                                             seq_lens,
+                                             seq_lens_encoder,
+                                             cos_emb,
+                                             sin_emb,
+                                             cache_k_scale,
+                                             cache_v_scale,
+                                             q_norm_weight,
+                                             k_norm_weight,
+                                             max_seq_len,
+                                             max_blocks_per_seq,
+                                             num_heads,
+                                             block_size,
+                                             127.0f,
+                                             -127.0f,
+                                             kv_num_heads,
+                                             rope_3d,
+                                             rms_norm_eps);
+}
+
 template <typename T, typename QKV_TYPE, bool IsFP8=false>
 void append_speculate_cache_int8_rope(const QKV_TYPE* qkv,
                                       uint8_t* key_cache,
@@ -125,7 +272,8 @@ void append_speculate_cache_int8_rope(const QKV_TYPE* qkv,
                                       const int bsz,
                                       const int token_num,
                                       const cudaStream_t& stream,
-                                      const bool use_neox_style) {
+                                      const bool use_neox_style,
+                                      const bool rope_3d) {
   constexpr int num_warps = 4;
   const int all_warps =
       ((num_heads + 2 * kv_num_heads) + num_warps - 1) / num_warps * num_warps;
@@ -167,7 +315,8 @@ void append_speculate_cache_int8_rope(const QKV_TYPE* qkv,
                                                block_size,
                                                127.0f,
                                                -127.0f,
-                                               kv_num_heads);
+                                               kv_num_heads,
+                                               rope_3d);
   } else {
     append_speculate_cache_int8_rope_kernel<T, 4, 0, 128, QKV_TYPE, IsFP8>
         <<<grids, num_warps * 32, 0, stream>>>(qkv,
@@ -191,7 +340,8 @@ void append_speculate_cache_int8_rope(const QKV_TYPE* qkv,
                                                block_size,
                                                127.0f,
                                                -127.0f,
-                                               kv_num_heads);
+                                               kv_num_heads,
+                                               rope_3d);
   }
 }
 
@@ -222,7 +372,8 @@ void append_speculate_cache_int4_rope(const QKV_TYPE* qkv,
                                       const int bsz,
                                       const int token_num,
                                       const cudaStream_t& stream,
-                                      const bool use_neox_style) {
+                                      const bool use_neox_style,
+                                      const bool rope_3d) {
   constexpr int num_warps = 4;
   const int all_warps =
       ((num_heads + 2 * kv_num_heads) + num_warps - 1) / num_warps * num_warps;
@@ -266,7 +417,8 @@ void append_speculate_cache_int4_rope(const QKV_TYPE* qkv,
                                                block_size,
                                                7.0f,
                                                -8.0f,
-                                               kv_num_heads);
+                                               kv_num_heads,
+                                               rope_3d);
   } else {
     append_speculate_cache_int4_rope_kernel<T, 4>
         <<<grids, num_warps * 32, 0, stream>>>(qkv,
@@ -292,7 +444,8 @@ void append_speculate_cache_int4_rope(const QKV_TYPE* qkv,
                                                block_size,
                                                7.0f,
                                                -8.0f,
-                                               kv_num_heads);
+                                               kv_num_heads,
+                                               rope_3d);
   }
 }
 template <typename T, typename QKV_TYPE>
@@ -313,11 +466,15 @@ void SpeculateWriteCacheWithRoPEKernel(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out) {
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps) {
   typedef cascade_attn_type_traits<T> traits_;
   typedef cascade_attn_type_traits<QKV_TYPE> qkt_nv_type_;
   typedef typename traits_::type DataType_;
@@ -342,6 +499,73 @@ void SpeculateWriteCacheWithRoPEKernel(
             ? rotary_embs.get().data<float>() + max_seq_len * dim_head
             : rotary_embs.get().data<float>() + max_seq_len * dim_head / 2;
   }
+  if (q_norm_weight && k_norm_weight) {
+    if (cache_quant_type_str == "none") {
+      append_speculate_cache_rope_qk_norm(
+          reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
+          reinterpret_cast<DataType_*>(key_cache_out->data<T>()),
+          reinterpret_cast<DataType_*>(value_cache_out->data<T>()),
+          reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+          block_tables.data<int>(),
+          batch_id_per_token.data<int>(),
+          cu_seqlens_q.data<int>(),
+          seq_lens.data<int>(),
+          seq_lens_encoder.data<int>(),
+          cos_emb,
+          sin_emb,
+          qkv_out_scales ? qkv_out_scales.get().data<float>() : nullptr,
+          qkv_biases ? reinterpret_cast<DataType_*>(
+                          const_cast<T*>(qkv_biases.get().data<T>()))
+                    : nullptr,
+          max_seq_len,
+          max_blocks_per_seq,
+          num_heads,
+          kv_num_heads,
+          dim_head,
+          block_size,
+          bsz,
+          token_nums,
+          stream,
+          use_neox_rotary_style,
+          reinterpret_cast<const float*>(q_norm_weight.get().data<float>()),
+          reinterpret_cast<const float*>(k_norm_weight.get().data<float>()),
+          rms_norm_eps,
+          rope_3d);
+    } else if (cache_quant_type_str == "block_wise_fp8") {
+        append_speculate_cache_fp8_dynamic_rope(
+            reinterpret_cast<const DataType_*>(qkv_ptr),
+            key_cache_out->data<uint8_t>(),
+            value_cache_out->data<uint8_t>(),
+            reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+            block_tables.data<int>(),
+            batch_id_per_token.data<int>(),
+            cu_seqlens_q.data<int>(),
+            seq_lens.data<int>(),
+            seq_lens_encoder.data<int>(),
+            cos_emb,
+            sin_emb,
+            const_cast<DataType_*>(reinterpret_cast<const DataType_*>(cache_k_scale.get().data<T>())),
+            const_cast<DataType_*>(reinterpret_cast<const DataType_*>(cache_v_scale.get().data<T>())),
+            q_norm_weight.get().data<float>(),
+            k_norm_weight.get().data<float>(),
+            max_seq_len,
+            max_blocks_per_seq,
+            num_heads,
+            kv_num_heads,
+            dim_head,
+            block_size,
+            bsz,
+            token_nums,
+            stream,
+            rope_3d,
+            rms_norm_eps
+        );
+    } else {
+      PD_THROW(
+          "append_decode_cache_rope_qk_norm not support cachekv quant yet");
+    }
+
+  } else {
     if (cache_quant_type_str == "none") {
         append_speculate_cache_rope(
             reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
@@ -368,7 +592,8 @@ void SpeculateWriteCacheWithRoPEKernel(
             bsz,
             token_nums,
             stream,
-        use_neox_rotary_style);
+            use_neox_rotary_style,
+            rope_3d);
     } else if (cache_quant_type_str == "cache_int8") {
         append_speculate_cache_int8_rope(
             reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
@@ -401,7 +626,8 @@ void SpeculateWriteCacheWithRoPEKernel(
             bsz,
             token_nums,
             stream,
-        use_neox_rotary_style);
+            use_neox_rotary_style,
+            rope_3d);
     } else if (cache_quant_type_str == "cache_fp8") {
         append_speculate_cache_int8_rope<DataType_, QKV_TYPE, true>(
             reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
@@ -434,7 +660,37 @@ void SpeculateWriteCacheWithRoPEKernel(
             bsz,
             token_nums,
             stream,
-        use_neox_rotary_style);
+            use_neox_rotary_style,
+            rope_3d);
+    } else if (cache_quant_type_str == "block_wise_fp8") {
+        append_speculate_cache_fp8_dynamic_rope(
+            reinterpret_cast<const DataType_*>(qkv_ptr),
+            key_cache_out->data<uint8_t>(),
+            value_cache_out->data<uint8_t>(),
+            reinterpret_cast<DataType_*>(qkv_out->data<T>()),
+            block_tables.data<int>(),
+            batch_id_per_token.data<int>(),
+            cu_seqlens_q.data<int>(),
+            seq_lens.data<int>(),
+            seq_lens_encoder.data<int>(),
+            cos_emb,
+            sin_emb,
+            const_cast<DataType_*>(reinterpret_cast<const DataType_*>(cache_k_scale.get().data<T>())),
+            const_cast<DataType_*>(reinterpret_cast<const DataType_*>(cache_v_scale.get().data<T>())),
+            nullptr, // q_norm_weight
+            nullptr, // k_norm_weight
+            max_seq_len,
+            max_blocks_per_seq,
+            num_heads,
+            kv_num_heads,
+            dim_head,
+            block_size,
+            bsz,
+            token_nums,
+            stream,
+            rope_3d,
+            rms_norm_eps
+        );
     } else if (cache_quant_type_str == "cache_int4_zp") {
         append_speculate_cache_int4_rope(
             reinterpret_cast<const QKV_TYPE*>(qkv_ptr),
@@ -473,12 +729,14 @@ void SpeculateWriteCacheWithRoPEKernel(
             bsz,
             token_nums,
             stream,
-        use_neox_rotary_style);
+            use_neox_rotary_style,
+            rope_3d);
     } else {
         PD_THROW(
             "cache_quant_type_str should be one of [none, cache_int8, "
             "cache_int4_zp]");
     }
+  }
 }
 
 template void SpeculateWriteCacheWithRoPEKernel<paddle::bfloat16, int>(
@@ -500,11 +758,15 @@ template void SpeculateWriteCacheWithRoPEKernel<paddle::bfloat16, int>(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out);
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps);
 
 template void
 SpeculateWriteCacheWithRoPEKernel<paddle::bfloat16, paddle::bfloat16>(
@@ -526,11 +788,15 @@ SpeculateWriteCacheWithRoPEKernel<paddle::bfloat16, paddle::bfloat16>(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out);
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps);
 
 template void SpeculateWriteCacheWithRoPEKernel<paddle::float16, int>(
     const AppendAttnMetaData& meta_data,
@@ -551,11 +817,15 @@ template void SpeculateWriteCacheWithRoPEKernel<paddle::float16, int>(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out);
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps);
 
 
 template void
@@ -578,8 +848,12 @@ SpeculateWriteCacheWithRoPEKernel<paddle::float16, paddle::float16>(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out);
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps);

custom_ops/gpu_ops/append_attn/speculate_write_cache_with_rope_kernel.h

@@ -35,8 +35,12 @@ void SpeculateWriteCacheWithRoPEKernel(
     const paddle::optional<paddle::Tensor>& cache_v_zp,
     const std::string& cache_quant_type_str,
     const bool use_neox_rotary_style,
+    const bool rope_3d,
     const int max_seq_len,
     cudaStream_t& stream,
     paddle::Tensor* qkv_out,
     paddle::Tensor* key_cache_out,
-    paddle::Tensor* value_cache_out);
+    paddle::Tensor* value_cache_out,
+    const paddle::optional<paddle::Tensor>& q_norm_weight,
+    const paddle::optional<paddle::Tensor>& k_norm_weight,
+    const float rms_norm_eps);

custom_ops/gpu_ops/append_attn/template_config.json

@@ -0,0 +1,144 @@
+{
+  "multiquery_attention_c8": {
+    "name": "multiquery_attention_c8",
+    "function_name": "MultiQueryAppendC8Attention",
+    "impl_file": "multiquery_attention_c8_impl.cuh",
+    "template_params": [
+      "T",
+      "GROUP_SIZE",
+      "HEAD_DIM",
+      "BLOCK_SIZE",
+      "CAUSAL",
+      "BLOCK_SHAPE_Q",
+      "NUM_WARP_Q",
+      "OutT",
+      "ENABLE_PREFILL",
+      "IsFP8",
+      "IsDynamicC8"
+    ],
+    "dispatch_params": {
+      "GROUP_SIZE": [1, 2, 4, 5, 6, 7, 8, 12, 14, 16],
+      "HEAD_DIM": [128],
+      "BLOCK_SIZE": [64],
+      "CAUSAL": [0, 1],
+      "BLOCK_SHAPE_Q": [16, 32, 64, 128],
+      "ENABLE_PREFILL": [0, 1],
+      "IsFP8": [0, 1],
+      "IsDynamicC8": [0, 1]
+    },
+    "data_types": [
+      ["paddle::float16", "paddle::float16", "float16_float16"],
+      ["paddle::float16", "paddle::float8_e4m3fn", "float16_fp8"],
+      ["paddle::float16", "int8_t", "float16_int8"],
+      ["paddle::bfloat16", "paddle::bfloat16", "bfloat16_bfloat16"],
+      ["paddle::bfloat16", "paddle::float8_e4m3fn", "bfloat16_fp8"],
+      ["paddle::bfloat16", "int8_t", "bfloat16_int8"]
+    ],
+    "max_instances_per_file": 80,
+    "file_prefix": "multiquery_attention_c8_",
+    "function_signature": "template void {function_name}{template_args}(\n    const AppendAttnMetaData &meta_data,\n    const paddle::Tensor &qkv,\n    const paddle::Tensor &cache_k,\n    const paddle::Tensor &cache_v,\n    const paddle::optional<paddle::Tensor> &attn_mask,\n    const paddle::Tensor &cache_k_scale,\n    const paddle::Tensor &cache_v_scale,\n    const paddle::optional<paddle::Tensor> &shift_bias,\n    const paddle::optional<paddle::Tensor> &smooth_weight,\n    const paddle::optional<paddle::Tensor> &sinks,\n    const paddle::Tensor &seq_lens_q,\n    const paddle::Tensor &seq_lens_kv,\n    const paddle::Tensor &seq_lens_encoder,\n    const paddle::Tensor &batch_id_per_token,\n    const paddle::Tensor &cu_seqlens_q,\n    const paddle::Tensor &block_table,\n    const paddle::Tensor &batch_ids,\n    const paddle::Tensor &tile_ids_per_batch,\n    const int num_blocks_x_cpu,\n    const int max_seq_len,\n    const int max_dec_len,\n    const float quant_max_bound,\n    const float quant_min_bound,\n    const float in_scale,\n    const int max_partition_size,\n    const int encoder_max_partition_size,\n    const int speculate_max_draft_token_num,\n    const bool is_decoder,\n    cudaStream_t &stream,\n    paddle::Tensor *out,\n    const int sliding_window);\n\n"
+  },
+  "multiquery_attention_c4": {
+    "name": "multiquery_attention_c4",
+    "function_name": "MultiQueryAppendC4Attention",
+    "impl_file": "multiquery_attention_c4_impl.cuh",
+    "template_params": [
+      "T",
+      "GROUP_SIZE",
+      "HEAD_DIM",
+      "BLOCK_SIZE",
+      "CAUSAL",
+      "BLOCK_SHAPE_Q",
+      "NUM_WARP_Q",
+      "OutT",
+      "ENABLE_PREFILL"
+    ],
+    "dispatch_params": {
+      "GROUP_SIZE": [1, 2, 4, 5, 6, 7, 8, 12, 14, 16],
+      "HEAD_DIM": [128],
+      "BLOCK_SIZE": [64],
+      "CAUSAL": [0, 1],
+      "BLOCK_SHAPE_Q": [16, 32, 64, 128],
+      "ENABLE_PREFILL": [0, 1]
+    },
+    "data_types": [
+      ["paddle::float16", "paddle::float16", "float16_float16"],
+      ["paddle::float16", "paddle::float8_e4m3fn", "float16_fp8"],
+      ["paddle::float16", "int8_t", "float16_int8"],
+      ["paddle::bfloat16", "paddle::bfloat16", "bfloat16_bfloat16"],
+      ["paddle::bfloat16", "paddle::float8_e4m3fn", "bfloat16_fp8"],
+      ["paddle::bfloat16", "int8_t", "bfloat16_int8"]
+    ],
+    "max_instances_per_file": 160,
+    "file_prefix": "multiquery_attention_c4_",
+    "function_signature": "template void {function_name}{template_args}(\n    const AppendAttnMetaData &meta_data,\n    const paddle::Tensor &qkv,\n    const paddle::Tensor &cache_k,\n    const paddle::Tensor &cache_v,\n    const paddle::optional<paddle::Tensor> &attn_mask,\n    const paddle::Tensor &cache_k_scale,\n    const paddle::Tensor &cache_v_scale,\n    const paddle::optional<paddle::Tensor> &cache_k_zp,\n    const paddle::optional<paddle::Tensor> &cache_v_zp,\n    const paddle::optional<paddle::Tensor> &shift_bias,\n    const paddle::optional<paddle::Tensor> &smooth_weight,\n    const paddle::optional<paddle::Tensor> &sinks,\n    const paddle::Tensor &seq_lens_q,\n    const paddle::Tensor &seq_lens_kv,\n    const paddle::Tensor &seq_lens_encoder,\n    const paddle::Tensor &batch_id_per_token,\n    const paddle::Tensor &cu_seqlens_q,\n    const paddle::Tensor &block_table,\n    const paddle::Tensor &batch_ids,\n    const paddle::Tensor &tile_ids_per_batch,\n    const int num_blocks_x_cpu,\n    const int max_seq_len,\n    const int max_dec_len,\n    const float quant_max_bound,\n    const float quant_min_bound,\n    const float in_scale,\n    const int max_partition_size,\n    const int encoder_max_partition_size,\n    const int speculate_max_draft_token_num,\n    const bool is_decoder,\n    cudaStream_t &stream,\n    paddle::Tensor *out,\n    const int sliding_window);\n\n"
+  },
+  "multiquery_attention_c16": {
+    "name": "multiquery_attention_c16",
+    "function_name": "MultiQueryAppendAttention",
+    "impl_file": "multiquery_attention_c16_impl.cuh",
+    "template_params": [
+      "T",
+      "GROUP_SIZE",
+      "HEAD_DIM",
+      "BLOCK_SIZE",
+      "CAUSAL",
+      "BLOCK_SHAPE_Q",
+      "NUM_WARP_Q",
+      "OutT",
+      "ENABLE_PREFILL"
+    ],
+    "dispatch_params": {
+      "GROUP_SIZE": [1, 2, 4, 5, 6, 7, 8, 12, 14, 16],
+      "HEAD_DIM": [64,128],
+      "BLOCK_SIZE": [64],
+      "CAUSAL": [0, 1],
+      "BLOCK_SHAPE_Q": [16, 32, 64, 128],
+      "ENABLE_PREFILL": [0, 1]
+    },
+    "data_types": [
+      ["paddle::float16", "paddle::float16", "float16_float16"],
+      ["paddle::float16", "paddle::float8_e4m3fn", "float16_fp8"],
+      ["paddle::float16", "int8_t", "float16_int8"],
+      ["paddle::bfloat16", "paddle::bfloat16", "bfloat16_bfloat16"],
+      ["paddle::bfloat16", "paddle::float8_e4m3fn", "bfloat16_fp8"],
+      ["paddle::bfloat16", "int8_t", "bfloat16_int8"]
+    ],
+    "max_instances_per_file": 160,
+    "file_prefix": "multiquery_attention_c16_",
+    "function_signature": "template void {function_name}{template_args}(\n    const AppendAttnMetaData &meta_data,\n    const paddle::Tensor &qkv,\n    const paddle::Tensor &cache_k,\n    const paddle::Tensor &cache_v,\n    const paddle::optional<paddle::Tensor> &attn_mask,\n    const paddle::optional<paddle::Tensor> &shift_bias,\n    const paddle::optional<paddle::Tensor> &smooth_weight,\n    const paddle::optional<paddle::Tensor> &sinks,\n    const paddle::Tensor &seq_lens_q,\n    const paddle::Tensor &seq_lens_kv,\n    const paddle::Tensor &seq_lens_encoder,\n    const paddle::Tensor &batch_id_per_token,\n    const paddle::Tensor &cu_seqlens_q,\n    const paddle::Tensor &block_table,\n    const paddle::Tensor &batch_ids,\n    const paddle::Tensor &tile_ids_per_batch,\n    const int num_blocks_x_cpu,\n    const int max_seq_len,\n    const int max_dec_len,\n    const float quant_max_bound,\n    const float quant_min_bound,\n    const float in_scale,\n    const int max_partition_size,\n    const int encoder_max_partition_size,\n    const int speculate_max_draft_token_num,\n    const bool is_decoder,\n    cudaStream_t &stream,\n    paddle::Tensor *out,\n    const int sliding_window);\n\n"
+  },
+  "multiquery_decoder_attention": {
+    "name": "multiquery_decoder_attention",
+    "function_name": "MultiQueryDecoderAttention",
+    "impl_file": "multiquery_decoder_attention_impl.cuh",
+    "template_params": [
+      "T",
+      "GROUP_SIZE",
+      "HEAD_DIM_QK",
+      "HEAD_DIM_V",
+      "BLOCK_SIZE",
+      "CAUSAL",
+      "NUM_STAGE",
+      "cache_bytes",
+      "DEAL_EACH_TIME"
+    ],
+    "dispatch_params": {
+      "GROUP_SIZE": [8, 16, 128],
+      "HEAD_DIM_QK": [128, 192, 512, 576],
+      "HEAD_DIM_V": [128, 192, 512, 576],
+      "BLOCK_SIZE": [64],
+      "CAUSAL": [0, 1],
+      "NUM_STAGE": [2],
+      "cache_bytes": [16],
+      "DEAL_EACH_TIME": [32, 64]
+    },
+    "data_types": [
+      ["paddle::float16", "", "float16"],
+      ["paddle::bfloat16", "", "bfloat16"]
+    ],
+    "max_instances_per_file": 60,
+    "file_prefix": "multiquery_decoder_attention_",
+    "function_signature": "template void {function_name}{template_args}(\n    const AppendAttnMetaData& meta_data,\n    cudaStream_t &stream,\n    const paddle::Tensor &q,\n    const paddle::Tensor &cache_k,\n    const paddle::Tensor &cache_v,\n    const paddle::optional<paddle::Tensor>& attn_mask,\n    const paddle::optional<paddle::Tensor>& shift_bias,\n    const paddle::optional<paddle::Tensor>& smooth_weight,\n    const paddle::Tensor &seq_lens_q,\n    const paddle::Tensor &seq_lens_kv,\n    const paddle::Tensor &batch_id_per_token,\n    const paddle::Tensor &cu_seqlens_q,\n    const paddle::Tensor &block_table,\n    const int max_seq_len,\n    const int max_dec_len,\n    const float rope_scale,\n    const float rope_theta,\n    const float softmax_scale,\n    const float in_scale,\n    paddle::Tensor *out);\n\n"
+  }
+}

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_bfloat16_bfloat16_kernel.cu

@@ -1,59 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-
-template void CascadeAppendAttentionC16Kernel<paddle::bfloat16, paddle::bfloat16>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_bfloat16_fp8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-template void CascadeAppendAttentionC16Kernel<paddle::bfloat16, paddle::float8_e4m3fn>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_bfloat16_int8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-template void CascadeAppendAttentionC16Kernel<paddle::bfloat16, int8_t>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_float16_float16_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-template void CascadeAppendAttentionC16Kernel<paddle::float16, paddle::float16>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_float16_fp8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-template void CascadeAppendAttentionC16Kernel<paddle::float16, paddle::float8_e4m3fn>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c16_float16_int8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c16_impl.cuh"
-
-template void CascadeAppendAttentionC16Kernel<paddle::float16, int8_t>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c4_bfloat16_bfloat16_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c4_impl.cuh"
-
-template void CascadeAppendAttentionC4Kernel<paddle::bfloat16, paddle::bfloat16>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c4_bfloat16_fp8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c4_impl.cuh"
-
-template void CascadeAppendAttentionC4Kernel<paddle::bfloat16, paddle::float8_e4m3fn>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);

custom_ops/gpu_ops/append_attn/template_instantiation/append_attention_c4_bfloat16_int8_kernel.cu

@@ -1,58 +0,0 @@
-// Copyright (c) 2024 PaddlePaddle Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-#include "../append_attention_c4_impl.cuh"
-
-template void CascadeAppendAttentionC4Kernel<paddle::bfloat16, int8_t>(
-    const AppendAttnMetaData& meta_data,
-    const paddle::Tensor& qkv,  // [token_num, num_heads, head_dim]
-    const paddle::Tensor&
-        cache_k,  // [max_block_num, num_heads, block_size, head_dim]
-    const paddle::Tensor&
-        cache_v,  // [max_block_num, num_heads, head_dim, block_size]
-    const paddle::optional<paddle::Tensor>& attn_mask,
-    const paddle::optional<paddle::Tensor>&
-        cache_k_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_scale,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_k_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        cache_v_zp,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        shift_bias,  // [num_kv_heads, head_dim]
-    const paddle::optional<paddle::Tensor>&
-        smooth_weight,  // [num_kv_heads, head_dim]
-    const paddle::Tensor& seq_lens_q,
-    const paddle::Tensor& seq_lens_kv,
-    const paddle::Tensor& seq_lens_encoder,
-    const paddle::Tensor& batch_id_per_token,
-    const paddle::Tensor& cu_seqlens_q,
-    const paddle::Tensor& block_table,
-    const paddle::Tensor& batch_ids,
-    const paddle::Tensor& tile_ids_per_batch,
-    const int num_blocks,
-    const int block_shape_q,
-    const int max_seq_len,
-    const int max_dec_len,
-    const float quant_max_bound,
-    const float quant_min_bound,
-    const float in_scale,
-    const int max_partition_size,
-    const int encoder_max_partition_size,
-    const int speculate_max_draft_token_num,
-    const bool causal,
-    const bool is_decoder,
-    const bool enable_prefill,
-    cudaStream_t& stream,
-    paddle::Tensor* out);