Commit Graph

28 Commits

Author SHA1 Message Date
Sunny-bot1
479c8b85d3 [Optimize]support machete weight only gemm (#3561)
* support machete weight only gemm

* add generate

* update

* fix

* change file location

* add sm_version limit

* fix

* fix

* fix ci

* fix coverage

* fix xpu
2025-08-28 09:49:58 +08:00
Liumengyuan
11803e0907 fix undefined cuPointerGetAttribute symbol error (#3628) 2025-08-27 20:24:59 +08:00
Jiang-Jia-Jun
c694fa2879 Revert "[Feature] block sparse attention (#3209)" (#3647)
This reverts commit 646a0c2fd8.
2025-08-27 17:35:04 +08:00
yangjianfengo1
646a0c2fd8 [Feature] block sparse attention (#3209)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
* 支持稀疏attn

* fix bug

* code style

* fix moba attn get kv shape

* 修复a100编译

* codestyle

* code style

* code style

* code style

* fix conflict

* 增加单侧

* code style

* 增加eblite 加载时间

* fix bug

* for ci

* for ci

* for ci

* for ci

* 支持mlp block size 128

* 增加小算子单测

* fix 单测 mlp

* 将环境变量加入到config里面

* fix rollout config
2025-08-26 07:16:04 -07:00
Kane2011
2ae7ab28d2 [MetaxGPU] adapt to the latest fastdeploy on metax gpu (#3492) 2025-08-25 17:44:20 +08:00
yangjianfengo1
b047681c5d 【New Feature】支持Fp8 group Gemm 24稀疏 (#3463)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
Publish Job / publish_pre_check (push) Has been cancelled
Publish Job / print_publish_pre_check_outputs (push) Has been cancelled
Publish Job / FD-Clone-Linux (push) Has been cancelled
Publish Job / Show Code Archive Output (push) Has been cancelled
Publish Job / BUILD_SM8090 (push) Has been cancelled
Publish Job / BUILD_SM8689 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8090 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8689 (push) Has been cancelled
Publish Job / Run FastDeploy Unit Tests and Coverage (push) Has been cancelled
Publish Job / Run FastDeploy LogProb Tests (push) Has been cancelled
Publish Job / Extracted partial CE model tasks to run in CI. (push) Has been cancelled
Publish Job / Run Base Tests (push) Has been cancelled
Publish Job / Run Accuracy Tests (push) Has been cancelled
* 支持24稀疏

* code style

* 增加stmatrix 宏定义判断

* code style
2025-08-19 02:54:47 -07:00
chen
f0f00a6025 [OPs] Universal optimization and Fix early_stop cuda 700 (#3375)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
* delete nonzero

* delete setup_ops_base.py

* check if

* check gcp infer_seed.cpu()

* fix repetition_early_stopper_kernel cuda 700
2025-08-14 22:40:44 +08:00
Kane2011
b4fef2cf29 [MetaxGPU] Support FastDeploy on metax gpu (#3241)
* [MetaxGPU] Support FastDeploy on metax gpu

* Update metax_worker.py

1. change worker log;
2. remove custom allreduce, adapt it later;
3. remove cuda graph;

* Update __init__.py

1. remove metax's key work comment

* Update __init__.py

1. remove metax's key word comment;
2. add fused_moe_kernel_paddle import

---------

Co-authored-by: yongqiangma <xing.wo@163.com>
2025-08-13 11:11:54 +08:00
yzwu
fbdd6b0663 [Iluvatar GPU] Optimze attention and moe performance (#3234) 2025-08-08 10:51:24 +08:00
yangjianfengo1
89397516a8 [New Feature] Support W4Afp8 MoE GroupGemm (#3171)
* init

* 增加多线程编译

* fix bug

* fix bug

* code style

* 增加fp16

* 将print替换成assert

* 修复stmatrix

* 减小单测shape

* 减小单测shape
2025-08-06 10:34:05 +08:00
yangjianfengo1
40f7f3e0d8 [New Feature] fa3 支持flash mask (#3184)
* 支持flash mask

* 修改test_flash_mask

* 修改test.sh
2025-08-05 12:20:48 +08:00
AIbin
22fe695f1c 【Inference Optimize】Support automatic generation of marlin kernel (#3149)
* Support automatic generation of marlin kernel
2025-08-01 22:43:18 +08:00
yangjianfengo1
64d7a3194d 集中式支持fa3 (#3112) 2025-08-01 18:03:36 +08:00
ming1753
5acde4eb43 [Feature] Multimodal Scheduler V1 (#3019)
* [Feature] Support multimodal scheduler v1

* remove debug log

* fix bug

* fix format

* modify code

* fix bug

* fix bug

* fix bug

* modify code
2025-07-30 16:05:55 +08:00
JYChen
dafe02a7b9 [stop sequence] support stop sequence (#3025)
* stop seqs in multi-ends

* unittest for gpu stop op

* kernel tid==0
2025-07-29 14:17:37 +08:00
chenjian
85a78d695d [Feature] Support block scheduler v1 for FD (#2928)
* Support FD block scheduler v1

* Support FD block scheduler v1

* Support FD block scheduler v1

* Fix according to copilot review

* Fix according to review

* Remove is_dummy

* Fix bug when real_bsz=1

* Fix infer first token cost time

---------

Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
2025-07-23 20:31:31 +08:00
lizexu123
67990e0572 [Feature] support min_p_sampling (#2872)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
* Fastdeploy support min_p

* add test_min_p

* fix

* min_p_sampling

* update

* delete vl_gpu_model_runner.py

* fix

* Align usage of min_p with vLLM

* fix

* modified unit test

* fix test_min_sampling

* pre-commit all files

* fix

* fix

* fix

* fix xpu_model_runner.py
2025-07-20 23:17:59 -07:00
Zero Rains
25698d56d1 polish code with new pre-commit rule (#2923) 2025-07-19 23:19:27 +08:00
Sunny-bot1
e45050cae3 [Feature] support top_k_top_p sampling (#2753)
* support top_k_top_p sampling

* fix

* add api param

* add api para

* fix

* fix

* fix

* fix

* fix

* fix

* fix
2025-07-09 20:58:58 -07:00
lifulll
1f28bdf994 dcu adapter ernie45t (#2756)
Co-authored-by: lifu <lifu@sugon.com>
Co-authored-by: yongqiangma <xing.wo@163.com>
2025-07-09 18:56:27 +08:00
zhink
b89180f1cd [Feature] support custom all-reduce (#2758)
* [Feature] support custom all-reduce

* add vllm adapted
2025-07-09 16:00:27 +08:00
celsowm
771e71a24d Feat/blackwell sm100 support (#2670)
* Add initial support for NVIDIA Blackwell (SM100) architecture

This change introduces initial support for the NVIDIA Blackwell GPU
architecture, specifically targeting SM100 (Compute Capability 10.x)
with '100a' architecture-specific features (e.g., for CUTLASS).

Key changes:
- Updated custom_ops/setup_ops.py to generate appropriate gencode
  flags (arch=compute_100a,code=sm_100a) when '100' is specified
  in FD_BUILDING_ARCS. Requires CUDA 12.9+.
- Updated custom_ops/gpu_ops/cutlass_extensions/gemm_configs.h:
    - Added CutlassTileConfigSM100 enum (with placeholder tile shapes).
    - Added BLACKWELL to CandidateConfigTypeParam.
    - Updated CutlassGemmConfig struct with is_sm100 flag,
      tile_config_sm100, and new constructor for SM100.
    - Modified toString() and fromString() for SM100 support.
- Updated custom_ops/gpu_ops/cutlass_kernels/cutlass_heuristic.cu:
    - Added get_candidate_tiles_sm100() (with placeholder tiles).
    - Added placeholder mcast support functions for SM100.
    - Updated get_candidate_configs() to include SM100 paths using
      the BLACKWELL flag and new SM100 config types.
- Updated build.sh with comments to guide users on specifying '100'
  for Blackwell in FD_BUILDING_ARCS.

Further work:
- Optimal CUTLASS tile configurations for SM100 need to be researched
  and updated in cutlass_heuristic.cu.
- Kernel auto-generation scripts in custom_ops/utils/ may need
  SM100-specific versions if Blackwell's hardware features for FP8/TMA
  differ significantly from SM90.
- Compatibility of third-party libraries (CUTLASS v3.8.0, DeepGEMM)
  with Blackwell should be fully verified.

* Feat: Implement detailed Blackwell (SM100) CUTLASS heuristics

This change integrates specific, expert-provided CUTLASS heuristic
configurations for the NVIDIA Blackwell (SM100) GPU architecture,
replacing previous placeholders. This includes:

- Updated `custom_ops/gpu_ops/cutlass_extensions/gemm_configs.h`:
    - Populated `CutlassTileConfigSM100` enum with specific tile shapes
      (e.g., CtaShape64x64x128B, CtaShape128x128x128B) suitable for SM100.
    - Added `FP4_ONLY` to `CandidateConfigTypeParam` for new FP4 paths.

- Updated `custom_ops/gpu_ops/cutlass_kernels/cutlass_heuristic.cu`:
    - Implemented `get_candidate_tiles_sm100` with detailed logic for
      selecting tile configurations based on GROUPED_GEMM and FP4_ONLY flags,
      using the new SM100 tile enums.
    - Implemented `supports_mcast_along_m_sm100` and
      `supports_mcast_along_n_sm100` with specific tile checks for Blackwell.
    - Updated the `sm == 100` (Blackwell) block in `get_candidate_configs`
      to use these new helper functions and accurately populate candidate
      kernel configurations for various cluster shapes.

- `custom_ops/setup_ops.py` remains configured to compile for
  `arch=compute_100a,code=sm_100a` with CUDA 12.9+ for these features.

This aligns the codebase with heuristic configurations similar to those
in upstream TensorRT-LLM / CUTLASS for Blackwell, enabling more
performant kernel selection on this new architecture.

---------

Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>
Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
2025-07-09 15:29:42 +08:00
EnflameGCU
d0f4d6ba3a [GCU] Support gcu platform (#2702)
baseline: e7fa57ebae

Co-authored-by: yongqiangma <xing.wo@163.com>
2025-07-08 13:00:52 +08:00
ming1753
ef6649a577 [Optimize] Optimize tensorwise fp8 performance (#2729)
Some checks failed
Deploy GitHub Pages / deploy (push) Has been cancelled
* [Optimize] Optimize tensorwise fp8 performance
2025-07-07 20:06:28 +08:00
liddk1121
1b54a2831e Adapt for iluvatar gpu (#2684) 2025-07-07 16:53:14 +08:00
Jiang-Jia-Jun
05c670e593 [Sync] Update to latest code (#2679)
* [Sync] Update to latest code

* Add new code files

* Add new code files

* update code

* Try to fix build.sh

* Try to fix build.sh

* Update code

* Update requirements.txt

* Update code

---------

Co-authored-by: Jiang-Jia-Jun <jiangjiajun@baidu.com>
2025-07-03 15:43:53 +08:00
Jiang-Jia-Jun
92c2cfa2e7 Sync v2.0 version of code to github repo 2025-06-29 23:29:37 +00:00
jiangjiajun
684703fd72 [LLM] First commit the llm deployment code 2025-06-09 19:20:15 +08:00