Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 18, 2025

• [Gluon][Dialect] More fixes and verifier improvements (#9018)

  Show description for 8f8064d
  

  Mogball

  and

  peterbell10

  authored
  8f8064d

  Copy full SHA for 8f8064d

• [BACKEND] Remove synchronisation in 2CTA mma (#8986)

  Show description for b447ef0
  lezcano

  authored
  b447ef0

  Copy full SHA for b447ef0

• [mxfp] fix mx8 w_scale mask on Hopper (#8974)

  Show description for e8ca3e9
  jongsoo-openai

  authored
  e8ca3e9

  Copy full SHA for e8ca3e9

• [Frontend] Fix desc.shape values for fp4_padded tensor descriptors (#9012)

  Show description for 39b1351
  peterbell10

  authored
  39b1351

  Copy full SHA for 39b1351

• [KERNELS] closure-based output mapping for peer shards (#8999)

  aeng-openai

  authored
  1df28b4

  Copy full SHA for 1df28b4

• [WS] store asyncOp per partition (#9007)

  Show description for cacbe59
  

  3gx

  and

  evghenii

  authored
  cacbe59

  Copy full SHA for cacbe59

• [AMD][gfx1250] Support 4 Warps Scheduling in MXFP GEMM Gluon Kernel (#9031)

  Show description for 08c9c4b
  knwng

  authored
  08c9c4b

  Copy full SHA for 08c9c4b

• [Nvidia] Revert ptxas 13.1 upgrade to 12.9.86 (#9016)

  Mogball

  authored
  c90b606

  Copy full SHA for c90b606

Commits on Dec 17, 2025

• [Reapply][BACKEND] Add missing waits in WGMMA rhs in register pipelining (#8997)

  Show description for 88e8a37
  peterbell10

  authored
  88e8a37

  Copy full SHA for 88e8a37

• [Backend] Delay wgmma wait(0) to first use of the accumulator (#9021)

  Show description for 0ff2d3d
  peterbell10

  authored
  0ff2d3d

  Copy full SHA for 0ff2d3d

• Use fresh_knobs_except_libraries for some tests when setting `TRITON_PTXAS_BLACKWELL_PATH` (#9011)

  Show description for 8700600
  CliveUnger

  authored
  8700600

  Copy full SHA for 8700600

• [AMD] Fix shared order selection for direct-to-LDS loads on GFX9 (#9028)

  Show description for f49784f
  AlexAUT

  authored
  f49784f

  Copy full SHA for f49784f

• [AMD] Make sure `tt.dot` dominates predecessors when pingpong (#9027)

  Show description for 3de87a5
  AlexAUT

  authored
  3de87a5

  Copy full SHA for 3de87a5

• [AMD][gfx1250] Support test_extract_slice_concat_op.py (#8960)

  Show description for a441a9a
  yiqian1

  authored
  a441a9a

  Copy full SHA for a441a9a

Commits on Dec 16, 2025

• [Backend] Add explicit semantics for async ops. (#8966)

  Show description for 6cc4505
  nzaghen

  authored
  6cc4505

  Copy full SHA for 6cc4505

• [AMD][gluon][gfx1250] Add async_copy_local_to_global (#8984)

  Show description for ffb6e95
  jammm

  authored
  ffb6e95

  Copy full SHA for ffb6e95

Commits on Dec 15, 2025

• [AMD][Gluon] Unroll loop to optimize gfx1250 F16 FA example (#8998)

  Show description for fe3db94
  raikonenfnu

  authored
  fe3db94

  Copy full SHA for fe3db94

• [AMD] Add check for cooperative groups support in driver (#8935)

  Show description for e8836cf
  nithinsubbiah

  authored
  e8836cf

  Copy full SHA for e8836cf

• [AMD][GFX1250] Persistent variant for WS f16 GEMM (#8990)

  Show description for 601486f
  PMylon

  authored
  601486f

  Copy full SHA for 601486f

• [AMD][gfx1250] Fix a bug about TDM when dim > 2 (#8994)

  Show description for f960e6d
  yangshuxin

  authored
  f960e6d

  Copy full SHA for f960e6d

Commits on Dec 14, 2025

• [Gluon][Dialect] Tighten verifiers, add more helpful error messages (#8981)

  Show description for 84214b6
  Mogball

  authored
  84214b6

  Copy full SHA for 84214b6

Commits on Dec 13, 2025

• [AMD] Support warp specialization on gfx1250 (#8947)

  Show description for f171598
  PMylon

  authored
  f171598

  Copy full SHA for f171598

• [AMD][GLUON] host TDM descriptor support for 1D-5D on gfx1250 (#8977)

  Show description for 278e956
  jammm

  authored
  278e956

  Copy full SHA for 278e956

Commits on Dec 12, 2025

• [PROTON] Drop invalid time kernels (#8961)

  Show description for 5a8358c
  spatel-oai

  authored
  5a8358c

  Copy full SHA for 5a8358c

• Reland #8586 [AMD] Add Warp-Pipeline Support, Gluon and LLVM lowering. (#8980)

  Show description for 4dd12eb
  jungpark-mlir

  authored
  4dd12eb

  Copy full SHA for 4dd12eb

• [WS] Variable naming fix in `LowerAref` (#8978)

  Show description for 8c20473
  

  3gx

  and

  evghenii

  authored
  8c20473

  Copy full SHA for 8c20473

• Revert "[AMD] Add Warp-Pipeline Support, Gluon and LLVM lowering." (#8975)

  Show description for 8378f19
  ThomasRaoux

  authored
  8378f19

  Copy full SHA for 8378f19

• [AMD] Update gfx1250 MXFP FA example kernel (#8972)

  borontion

  authored
  0b44c63

  Copy full SHA for 0b44c63

• [Backend] Consider memdesc_slice in the Membar analysis (#8755)

  Show description for af6b560
  nzaghen

  authored
  af6b560

  Copy full SHA for af6b560

• [PROTON] Introduce the `clear_data` API to remove profile data collected before deactivation (#8971)

  Jokeren

  authored
  037bb02

  Copy full SHA for 037bb02

• [mxfp] allow non persistent mx matmul w/ strided layout (#8808)

  Show description for d568940
  jongsoo-openai

  authored
  d568940

  Copy full SHA for d568940

• [AMD] Add Warp-Pipeline Support, Gluon and LLVM lowering. (#8586)

  Show description for 4899090
  jungpark-mlir

  authored
  4899090

  Copy full SHA for 4899090

• [Backend] Bump to llvm/llvm-project@a992f294 (#8919)

  Show description for 2446c2f
  antiagainst

  authored
  2446c2f

  Copy full SHA for 2446c2f

Commits on Dec 11, 2025

• [CONSAN] Adding support for aliasing (#8939)

  Show description for 9073370
  pawelszczerbuk

  authored
  9073370

  Copy full SHA for 9073370

• Upgrade ptxas to 13.1 for Blackwell (#8941)

  Show description for f90bf63
  adstraw

  authored
  f90bf63

  Copy full SHA for f90bf63

Pagination

Previous

Next