-
Notifications
You must be signed in to change notification settings - Fork 18
Closed
Description
Timeline
Released Date: Dec. 16th, 2023
Work Items
Major Improvement
-
- MSCCL++: integrated with MSCCL++ and removed dependency on
gpudma(Integrate with MSCCL++ #179)
- MSCCL++: integrated with MSCCL++ and removed dependency on
Platforms Support
-
- ROCm: add ROCm multi-GPU support (ROCm multi-GPU support #181)
Operators
-
- reduce:
keepdimssupport for reduction (keepdimssupport for reduction #173)
- reduce:
Optimization
-
- OpGraph: optimize OpGraph scheduling (Optimize OpGraph scheduling #182)
Examples
-
- Example: add Llama2 multi-GPU examples (Support multi-GPU inference for llama2 #170)
CI
-
[ ] Unit Tests: revise Python unit tests & add to the Azure pipeline(moved to the next version release plan)
-
[ ] Unit Tests: add ROCm Azure pipelines(moved to the next version release plan)
-
[ ] Code Coverage: add code coverage for Python code(moved to the next version release plan)
Metadata
Metadata
Assignees
Labels
No labels