Commits

Breadcrumbs

History for

onb253b77

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 27, 2021

• [DOCS] Improved documentation and integration in CI (triton-lang#139)

  ptillet

  committed
  b253b77

  Copy full SHA for b253b77

• [CI] Made build-wheels compatible with system LLVM setup (triton-lang#138)

  Show description for 76c6f24
  ptillet

  committed
  76c6f24

  Copy full SHA for 76c6f24

• [CI] Various improvements to CI (triton-lang#137)

  Show description for 8eb63bc
  ptillet

  committed
  8eb63bc

  Copy full SHA for 8eb63bc

• [FRONTEND] Fixed bugs in global symbols resolution of @triton.jit'd functions (triton-lang#136)

  ptillet

  committed
  298aead

  Copy full SHA for 298aead

• [FRONTEND] Added compilation flag to force use of .nc cache modifier (triton-lang#134)

  Show description for 01276b5
  ptillet

  committed
  01276b5

  Copy full SHA for 01276b5

• [LANGUAGE] Added cos/sin (triton-lang#132)

  ptillet

  committed
  2824345

  Copy full SHA for 2824345

• [PYTHON] Bugfix in trans_c for dsd in blocksparse matmul (triton-lang#131)

  ptillet

  committed
  3169e43

  Copy full SHA for 3169e43

• [IR] Preliminary support for BF16 (triton-lang#129)

  Show description for 8cea583
  ptillet

  committed
  8cea583

  Copy full SHA for 8cea583

• [PYTHON] Added isinstance to list of builtin symbols exposed (triton-lang#128)

  ptillet

  committed
  9b4e2ca

  Copy full SHA for 9b4e2ca

• [CODEGEN] Performance improvement on A100 (triton-lang#125)

  Show description for d8d6b71
  

  daadaada

  authored and

  ptillet

  committed
  d8d6b71

  Copy full SHA for d8d6b71

• [DRIVER] Now giving the option to use system ptxas through environment variable (triton-lang#123)

  ptillet

  committed
  b7b05a5

  Copy full SHA for b7b05a5

• [IR] Added IR and Codegen support for atomic_rmw (triton-lang#120)

  ptillet

  committed
  0274429

  Copy full SHA for 0274429

• [LANGUAGE] Added support for bitcast (triton-lang#119)

  ptillet

  committed
  59b0ac6

  Copy full SHA for 59b0ac6

• [PYTHON] Added support for tuples (triton-lang#116)

  ptillet

  committed
  3ab121d

  Copy full SHA for 3ab121d

• [CODEGEN] Fixed atomic_add issue (triton-lang#112)

  Show description for f81012a
  ptillet

  committed
  f81012a

  Copy full SHA for f81012a

• [PYTHON] Fixed bug in scoping mechanism (triton-lang#111)

  Show description for 325ee38
  ptillet

  committed
  325ee38

  Copy full SHA for 325ee38

• [GENERAL] Minor improvements: (triton-lang#110)

  Show description for 9f30af7
  ptillet

  committed
  9f30af7

  Copy full SHA for 9f30af7

• [PYTHON] Added frontend to print sass using turingas disasm.py (triton-lang#109)

  ptillet

  committed
  288b4f7

  Copy full SHA for 288b4f7

• [CI] Made setup.py more backwards-compatible (triton-lang#108)

  ptillet

  committed
  c91dd56

  Copy full SHA for c91dd56

• [triton-ops] Minor build improvements (triton-lang#106)

  ptillet

  committed
  1476759

  Copy full SHA for 1476759

• [PYTHON] Update TensorWrapper with device attribute (triton-lang#102)

  Show description for ac57812
  

  siemanko

  authored and

  ptillet

  committed
  ac57812

  Copy full SHA for ac57812

• [CODEGEN] Removed dedicated reassociate pass to merge it into LLVM isel (triton-lang#101)

  Show description for 840140b
  ptillet

  committed
  840140b

  Copy full SHA for 840140b

• [LANG] Preliminary FP8 support (triton-lang#96)

  ptillet

  committed
  7355efa

  Copy full SHA for 7355efa

• [PYTHON] Various minor codegen fixes (triton-lang#95)

  ptillet

  committed
  4290be1

  Copy full SHA for 4290be1

• [PYTHON] Added atomic_add (triton-lang#94)

  ptillet

  committed
  2b75158

  Copy full SHA for 2b75158

• [PYTHON] Added shape & device checks for inputs to sparse matmul op (triton-lang#93)

  

  norabelrose

  authored and

  ptillet

  committed
  d7f8792

  Copy full SHA for d7f8792

• [PYTHON] Renamed triton.core -> triton.language (triton-lang#92)

  ptillet

  committed
  bfc0a75

  Copy full SHA for bfc0a75

• [PYTHON] Fix crash in sparse matmul for tensors w/ less than 4 dims (triton-lang#91)

  

  norabelrose

  authored and

  ptillet

  committed
  4141001

  Copy full SHA for 4141001

• [PYTHON] Allow triton.code_gen.Binary to print Triton-IR asm. (triton-lang#89)

  

  daadaada

  authored and

  ptillet

  committed
  f668837

  Copy full SHA for f668837

• [PYTHON] Fix variable referenced before assignment error in blocksparse matmul backward (triton-lang#90)

  

  norabelrose

  authored and

  ptillet

  committed
  1112e25

  Copy full SHA for 1112e25

• [PYTHON] Now triton.code_gen.Binary can print PTX and LLIR (triton-lang#88)

  ptillet

  committed
  d911214

  Copy full SHA for d911214

• [DOCS] Updates and improvements (triton-lang#87)

  ptillet

  committed
  29e33e5

  Copy full SHA for 29e33e5

• Deprecation of Triton-C and Replacement by decorated Python functions (triton-lang#86)

  Show description for 39f4730
  ptillet

  committed
  39f4730

  Copy full SHA for 39f4730

• [DOCS] Various improvements and typo fixes

  ptillet

  committed
  1fdb465

  Copy full SHA for 1fdb465

• [PYTHON] Fixed formatting issue in conv.c

  ptillet

  committed
  167a2e4

  Copy full SHA for 167a2e4

Pagination

Previous

Next