Commits

Branch selector

 main_perf

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 4, 2025

• Add a python script to plot linear layout (#899)

  Show description for cb89b61
  zhanglx13

  authored
  cb89b61

  Copy full SHA for cb89b61

Commits on Nov 3, 2025

• [plot tool] Fixed an issue related to tilesPerWarp (#898)

  zhanglx13

  authored
  bfe7baa

  Copy full SHA for bfe7baa

Commits on Oct 2, 2025

• Correct links in tune_gemm README (#886)

  Show description for 76076e1
  matthiasdiener

  authored
  76076e1

  Copy full SHA for 76076e1

Commits on Sep 15, 2025

• fix some syntax errors of the kernel in 06-fused-attention-transV.py (#876)

  Show description for 44e1d37
  scxiao

  authored
  44e1d37

  Copy full SHA for 44e1d37

Commits on Aug 1, 2025

• [PERF-CI] Set kpack==1 for gfx950 fallbacks to avoid warnings (#847)

  AlexAUT

  authored
  fc0620e

  Copy full SHA for fc0620e

Commits on Jul 23, 2025

• Fix perfCI for streamk/persistent gemm on gfx950 (#843)

  Show description for 7383938
  xiaohuguo2023

  authored
  7383938

  Copy full SHA for 7383938

Commits on Jun 29, 2025

• Use rocprofv3 and improve streamk benchmark (#835)

  Show description for ea5f7ad
  zhanglx13

  authored
  ea5f7ad

  Copy full SHA for ea5f7ad

Commits on Jun 25, 2025

• Update readme of layout plot tool (#831)

  Show description for 7ab7288
  AlexAUT

  authored
  7ab7288

  Copy full SHA for 7ab7288

Commits on May 12, 2025

• Add support for tilesPerWarp (#794)

  Show description for ba25d4d
  zhanglx13

  authored
  ba25d4d

  Copy full SHA for ba25d4d

Commits on Apr 23, 2025

• Compute max before mul QK_SCALE to fold sub into fma (#781)

  zhanglx13

  authored
  1468103

  Copy full SHA for 1468103

• Remove the if-else block in causal loop (#779)

  Show description for 34ee120
  zhanglx13

  authored
  34ee120

  Copy full SHA for 34ee120

Commits on Apr 22, 2025

• Optimize RMSNorm backward pass (#769)

  Show description for 44313cf
  brunomazzottiamd

  authored
  44313cf

  Copy full SHA for 44313cf

Commits on Apr 15, 2025

• plot_layout.py Refactoring (#775)

  Show description for 32a8fdd
  

  jtang10

  and

  Jingning Tang

  authored
  32a8fdd

  Copy full SHA for 32a8fdd

Commits on Apr 11, 2025

• Fix pid remapping logic when GRID_MN cannot divide NUM_XCDS (#722)

  zhanglx13

  authored
  a18fc87

  Copy full SHA for a18fc87

• Fix bwd Mode in flash-attention.py (#772)

  knwng

  authored
  c7fea1b

  Copy full SHA for c7fea1b

Commits on Mar 31, 2025

• Enable all perf CI configs (#771)

  AlexAUT

  authored
  9a32ed0

  Copy full SHA for 9a32ed0

Commits on Mar 28, 2025

• Rename instruction_sched_hint to schedule_hint (#768)

  AlexAUT

  authored
  5dd08e1

  Copy full SHA for 5dd08e1

Commits on Mar 24, 2025

• Tianxing/moe int8 w8a8 (#765)

  Show description for 8e42af9
  Chi-Chu319

  authored
  8e42af9

  Copy full SHA for 8e42af9

Commits on Mar 21, 2025

• Modify gemm to support block scaling with fp8 (#761)

  Show description for f669d30
  azaidy

  authored
  f669d30

  Copy full SHA for f669d30

Commits on Mar 18, 2025

• Merge pull request #757 from ROCm/ravil/rocm-triton-perf

  Show description for 0e3e441
  ravil-mobile

  authored
  0e3e441

  Copy full SHA for 0e3e441

Commits on Mar 17, 2025

• [AMD] Added a script which computes TFLOP/s using perf. counters

  ravil-mobile

  committed
  8e2a597

  Copy full SHA for 8e2a597

Commits on Mar 14, 2025

• Fix bug in FA ops calc

  azaidy

  authored
  f07da4a

  Copy full SHA for f07da4a

• Fix flash attention ops calculation (#758)

  azaidy

  authored
  aa1c8ff

  Copy full SHA for aa1c8ff

Commits on Mar 13, 2025

• remove production name (#755)

  xiaohuguo2023

  authored
  48455d1

  Copy full SHA for 48455d1

Commits on Mar 7, 2025

• Changes to FA benchmarking (#742)

  Show description for 5653590
  vgokhale

  authored
  5653590

  Copy full SHA for 5653590

• Fix correctness of streamk by adding a required barrier (#745)

  AlexAUT

  authored
  0d975d4

  Copy full SHA for 0d975d4

Commits on Mar 6, 2025

• bugfix (#749)

  vgokhale

  authored
  7adc012

  Copy full SHA for 7adc012

• Vinayak/gemm benchmarking (#746)

  Show description for b4924f1
  vgokhale

  authored
  b4924f1

  Copy full SHA for b4924f1

Commits on Mar 4, 2025

• Added support for lower precision b tensor in GEMMs (#738)

  Show description for 46ca63c
  azaidy

  authored
  46ca63c

  Copy full SHA for 46ca63c

Commits on Feb 27, 2025

• Tianxing/rope latent attention (#731)

  Show description for ade5449
  Chi-Chu319

  authored
  ade5449

  Copy full SHA for ade5449

Commits on Feb 26, 2025

• Change grouping calculation in gemm.py (#732)

  Show description for 5bb32e8
  azaidy

  authored
  5bb32e8

  Copy full SHA for 5bb32e8

• Clean up GEMM kernel (#730)

  Show description for 83871ea
  vgokhale

  authored
  83871ea

  Copy full SHA for 83871ea

Commits on Feb 25, 2025

• tuned few configs by hand (#733)

  xiaohuguo2023

  authored
  2a90b5b

  Copy full SHA for 2a90b5b

Commits on Feb 24, 2025

• moe quantization support int8 and fp8 (#702)

  Show description for 845d46e
  Chi-Chu319

  authored
  845d46e

  Copy full SHA for 845d46e

Commits on Feb 21, 2025

• Added compiler hints to enable buffer loads (#729)

  azaidy

  authored
  752d83c

  Copy full SHA for 752d83c

Pagination

Previous

Next