TensorRT-LLM fix: fix for cp

Fix the issue that kv_head_num becomes 0 when cp_size * tp_size > kv_head_num for MQA. Refine Ulysses code in AttentionOp

Mar 24 '25 03:03 DylanChen-NV

/bot run

Mar 24 '25 03:03 DylanChen-NV

PR_Github #221 [ run ] triggered by Bot

Mar 24 '25 03:03 niukuo

PR_Github #221 [ run ] completed with state SUCCESS /LLM/main/L0_MergeRequest_PR pipeline #228 completed with status: 'FAILURE'

Mar 24 '25 05:03 niukuo

/bot run

Mar 24 '25 08:03 DylanChen-NV

PR_Github #263 [ run ] triggered by Bot

Mar 24 '25 08:03 niukuo

PR_Github #263 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #253 completed with status: 'FAILURE'

Mar 24 '25 08:03 niukuo

/bot run

Mar 24 '25 08:03 DylanChen-NV

PR_Github #275 [ run ] triggered by Bot

Mar 24 '25 09:03 niukuo

PR_Github #275 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #265 completed with status: 'FAILURE'

Mar 24 '25 10:03 niukuo

/bot run

Mar 24 '25 10:03 DylanChen-NV

PR_Github #289 [ run ] triggered by Bot

Mar 24 '25 10:03 niukuo

PR_Github #289 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #278 completed with status: 'FAILURE'

Mar 24 '25 12:03 niukuo

/bot run

Mar 24 '25 13:03 DylanChen-NV

PR_Github #298 [ run ] triggered by Bot

Mar 24 '25 13:03 niukuo

PR_Github #298 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #285 completed with status: 'FAILURE'

Mar 24 '25 14:03 niukuo

/bot run

Mar 25 '25 08:03 DylanChen-NV

PR_Github #406 [ run ] triggered by Bot

Mar 25 '25 08:03 niukuo

PR_Github #406 [ run ] completed with state FAILURE /LLM/main/L0_MergeRequest_PR pipeline #358 completed with status: 'FAILURE'

Mar 25 '25 17:03 niukuo

/bot run --stage-lite "H100_PCIe-5"

Mar 26 '25 01:03 DylanChen-NV

PR_Github #485 Bot args parsing error!

Mar 26 '25 01:03 niukuo

/bot run --stage-list "H100_PCIe-5"

Mar 26 '25 01:03 DylanChen-NV

PR_Github #486 [ run ] triggered by Bot

Mar 26 '25 01:03 niukuo

PR_Github #486 [ run ] completed with state SUCCESS /LLM/main/L0_MergeRequest_PR pipeline #418 (Partly Tested) completed with status: 'SUCCESS'

Mar 26 '25 03:03 niukuo

/bot reuse-pipeline

Mar 26 '25 03:03 byshiue

PR_Github #507 [ reuse-pipeline ] triggered by Bot

Mar 26 '25 03:03 niukuo

PR_Github #507 [ reuse-pipeline ] completed with state SUCCESS Release Check Pipeline #119 failed Reusing PR_Github #486 (Partly Tested) for commit 79a2842

Mar 26 '25 03:03 niukuo

/bot reuse-pipeline

Mar 26 '25 04:03 byshiue

PR_Github #512 [ reuse-pipeline ] triggered by Bot

Mar 26 '25 04:03 niukuo

PR_Github #512 [ reuse-pipeline ] completed with state SUCCESS Reusing PR_Github #486 (Partly Tested) for commit 79a2842

Mar 26 '25 04:03 niukuo

fix: fix for cp > kvHeadNum