* support eplb in api_server
* update code
* add eplb test case
* update eplb
* support tp+dp eplb
* update test cese
* update code
* update code
* fix bug
* update copilot review
* update test case name
* Enhance run_ci_xpu.sh with caching and prefill options
* Update model path and configuration in run_ci_xpu.sh
* Add '北朝' keyword to assertion in run_45vl.py
* Enhance process termination logic in run_ci_xpu.sh
* Set timeout for CI_XPU job to 60 minutes
* Remove extra newline in stop_processes function
* Add unit tests for DeepEP buffer functionality
This file contains unit tests for the DeepEP buffer helpers and runners, including various test cases for buffer allocation, cleanup, and dispatching processes.
* Refactor DeepEP tests to use scoped stubs
* Add licensing information to test_ep.py
Added licensing information to the test file.
* update test utils
* Add comprehensive unit tests for DP scheduler functionality
- Add test_dp_scheduler.py with full-featured unit tests supporting both normal and standalone modes
- Add test_dp_scheduler_simple.py with lightweight mock-based tests for easy execution
- Add comprehensive README.md documenting test architecture and usage
- Tests cover DPLocalScheduler and DPScheduler classes with focus on:
- Request lifecycle management and TTL support
- Response handling and routing
- Resource-based scheduling and constraint handling
- Multi-threading and concurrent operations
- Splitwise role support (prefill vs decode)
- Error handling and edge cases
- Thread-safe operations with proper synchronization
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Remove tests/multimodal/test_utils.py
This file appears to be duplicate or misplaced, removing it to clean up the test structure.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* update
* fix
* rm unused file
---------
Co-authored-by: Claude <noreply@anthropic.com>
* [fix] fix v1 scheduler profile run for append attention in prefill node
* [fix] skip send_signal if kv signal not inited for gpu and xpu
* [fix] extend fix to flash_attn & mla_attn
* [fix] fix v1 pd run in ipc transfer protocol
* [ci] add test for v1 pd profile run using ipc transfer protocol
* [style] fix code style check
* [style] fix code style again
* [fix] fix profile run
* [update] remove --num-gpu-blocks-override in example script
* [chore] rename forward_meta is_profiling to is_dummy_or_profile_run
* unify max tokens
* modify and add unit test
* modify and add unit test
* modify and add unit tests
---------
Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com>
* [INTEL HPU] added hpu ci work flow support
Signed-off-by: Luo, Focus <focus.luo@intel.com>
* [INTEL HPU] added run ci hpu test scripts
Signed-off-by: Luo, Focus <focus.luo@intel.com>
* [INTEL HPU] enabled HPU ernie test case
Signed-off-by: Luo, Focus <focus.luo@intel.com>
* [INTEL HPU] updated Intel Gaudi Readme with Warmup disable cmdline
Signed-off-by: Luo, Focus <focus.luo@intel.com>
* Modify paddlepaddle installation command
Updated paddlepaddle installation command to use a specific index URL.
* Update run_ci_hpu.sh
* Rename json directory to nlohmann_json
Rename extracted json directory to nlohmann_json.
* Update ci_hpu.yml
* Set pip global index URL to Tsinghua mirror
* Update CI workflow to use self-hosted runner and paths
* Update Docker image in CI workflow
* Modify HPU installation URLs in run_ci_hpu.sh
Updated the installation URL for paddle_intel_hpu and added paddlenlp_ops installation.
* Fix paddle_intel_hpu installation URL
Corrected the URL for paddle_intel_hpu wheel installation.
---------
Signed-off-by: Luo, Focus <focus.luo@intel.com>
Co-authored-by: plusNew001 <95567040+plusNew001@users.noreply.github.com>
* Ignore markdown and text files in CI workflow
* Change GPU_ID to XPU_ID in run_ci_xpu.sh
* Change GPU_ID to XPU_ID in test configuration
* Change GPU_ID to XPU_ID for service port calculation
* Change GPU_ID to XPU_ID for device identification
* Change GPU_ID to XPU_ID in test_ep function
* Update run_w4a8.py
* Redirect stop_processes output to kill.log
Redirect output of stop_processes to kill.log to capture logs.
* Log server output for failed test cases
Added logging of server.log for failed tests.
* Add '-s' option to pytest commands in run_ci_xpu.sh
* Refactor assertion to validate multiple keywords
Updated assertion to check for multiple keywords in response.
* Fix assertany to assert any in run_45vl.py
* metrics use port the same as api_port
* Be tolerant to tests that monkeypatch/partially mock args.
* Reduce code redundancy
---------
Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com>
* Ignore markdown and text files in CI workflow
* Change GPU_ID to XPU_ID in run_ci_xpu.sh
* Change GPU_ID to XPU_ID in test configuration
* Change GPU_ID to XPU_ID for service port calculation
* Change GPU_ID to XPU_ID for device identification
* Change GPU_ID to XPU_ID in test_ep function
* Update run_w4a8.py
* Redirect stop_processes output to kill.log
Redirect output of stop_processes to kill.log to capture logs.
* Log server output for failed test cases
Added logging of server.log for failed tests.
* Add '-s' option to pytest commands in run_ci_xpu.sh
* change CI ep test from offline to online
* add ep all2all ci's changes, from offline to online
* change env var in ep-all2all ci test
* add expected response for ep8tp8 all2all
* Adapt to CI refactoring and support dual-concurrent code execution
* Adapt to CI refactoring and support dual-concurrent, second
* Explicitly specify the #port
* change the startup method of all2all
* Modify the command of all2all
* Update assertion to check multiple keywords
* Update assertion to check multiple keywords
* Update run_w4a8.py
* Update run_w4a8.py
---------
Co-authored-by: plusNew001 <95567040+plusNew001@users.noreply.github.com>
