TensorRT-LLM icon indicating copy to clipboard operation
TensorRT-LLM copied to clipboard
verified publisher icon NVIDIA

feat: update allreduce benchmark

Open yizhang-nv opened this issue 9 months ago • 16 comments

This MR changes the current allreduce benchmark from TRT flow to PyTorch flow with cuda graph + norm fusion support

yizhang-nv avatar Mar 24 '25 04:03 yizhang-nv

/bot run

yizhang-nv avatar Mar 24 '25 05:03 yizhang-nv

/bot run

yizhang-nv avatar Mar 24 '25 05:03 yizhang-nv

PR_Github #239 [ run ] triggered by Bot

niukuo avatar Mar 24 '25 05:03 niukuo

PR_Github #241 [ run ] triggered by Bot

niukuo avatar Mar 24 '25 05:03 niukuo

PR_Github #239 [ run ] completed with state ABORTED

niukuo avatar Mar 24 '25 05:03 niukuo

/bot run

yizhang-nv avatar Mar 24 '25 06:03 yizhang-nv

PR_Github #252 [ run ] triggered by Bot

niukuo avatar Mar 24 '25 06:03 niukuo

PR_Github #241 [ run ] completed with state ABORTED

niukuo avatar Mar 24 '25 06:03 niukuo

/bot run

yizhang-nv avatar Mar 24 '25 09:03 yizhang-nv

PR_Github #280 [ run ] triggered by Bot

niukuo avatar Mar 24 '25 09:03 niukuo

PR_Github #252 [ run ] completed with state ABORTED

niukuo avatar Mar 24 '25 09:03 niukuo

PR_Github #280 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #270 completed with status: 'FAILURE'

niukuo avatar Mar 24 '25 11:03 niukuo

/bot run

yizhang-nv avatar Mar 24 '25 16:03 yizhang-nv

PR_Github #324 [ run ] triggered by Bot

niukuo avatar Mar 24 '25 18:03 niukuo

PR_Github #324 [ run ] completed with state SUCCESS /LLM/main/L0_MergeRequest_PR pipeline #304 completed with status: 'FAILURE'

niukuo avatar Mar 24 '25 19:03 niukuo

/bot run

yizhang-nv avatar Mar 25 '25 02:03 yizhang-nv

close due to out of sync

yizhang-nv avatar Aug 05 '25 07:08 yizhang-nv