Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 3, 2022

• [spirv] Allow multiple tiles on M/N dimension for cooperative matrix (iree-org#11020)

  Show description for 6b421ca
  antiagainst

  authored
  6b421ca

  Copy full SHA for 6b421ca

• Handle case where SHARK results are empty (iree-org#11006)

  mariecwhite

  authored
  c20f592

  Copy full SHA for c20f592

• Relax condition to enable group reduction (iree-org#11015)

  Show description for c43f2b8
  ThomasRaoux

  authored
  c43f2b8

  Copy full SHA for c43f2b8

Commits on Nov 2, 2022

• Add support for generalizing unpack op when outer dim sizes are all 1s. (iree-org#10990)

  Show description for e75d7d1
  hanhanW

  authored
  e75d7d1

  Copy full SHA for e75d7d1

• Document PyPI deploy script a bit (iree-org#11016)

  Show description for d00784e
  GMNGeoffrey

  authored
  d00784e

  Copy full SHA for d00784e

• Bump actions/add-to-project version to use new API. (iree-org#11010)

  Show description for ab6ba22
  ScottTodd

  authored
  ab6ba22

  Copy full SHA for ab6ba22

• Add support for vectorizing pack ops that outer dims are not all 1s. (iree-org#10887)

  Show description for bec4d84
  hanhanW

  authored
  bec4d84

  Copy full SHA for bec4d84

• Remove transform.structured.print (iree-org#11007)

  Show description for 9008263
  matthias-springer

  authored
  9008263

  Copy full SHA for 9008263

• Prefetch LLVM @d4c4e49196cf17a1701bbd7337ffa2f8843908b3 and fix (iree-org#10995)

  nicolasvasilache

  authored
  e3826bd

  Copy full SHA for e3826bd

• Cherry-pick llvm/llvm-project@91f62f0e (iree-org#11004)

  Show description for 29bd703
  ThomasRaoux

  authored
  29bd703

  Copy full SHA for 29bd703

• [spirv] Promote C matrix to support various fused elementwise ops (iree-org#10980)

  Show description for 4221b2f
  antiagainst

  authored
  4221b2f

  Copy full SHA for 4221b2f

• [spirv] Support GPU subgroup MMA ops in SPIRVVectorizeLoadStore (iree-org#10979)

  Show description for 076f8e9
  antiagainst

  authored
  076f8e9

  Copy full SHA for 076f8e9

• [spirv] Fix K tile size deduction for cooperative matrix (iree-org#10998)

  Show description for 5cc4ae7
  antiagainst

  authored
  5cc4ae7

  Copy full SHA for 5cc4ae7

Commits on Nov 1, 2022

• Updating IPO to preserve base duplicate arguments. (iree-org#11001)

  Show description for 27960a3
  benvanik

  authored
  27960a3

  Copy full SHA for 27960a3

• Update e2e unpack tests to not rely on pack ops. (iree-org#10999)

  Show description for 5802ca7
  hanhanW

  authored
  5802ca7

  Copy full SHA for 5802ca7

• Fix mhlo.scatter implicit batch update dims (iree-org#10976)

  Show description for 16506ec
  rsuderman

  authored
  16506ec

  Copy full SHA for 16506ec

• Migrate to use linalg.transpose op and unify all the pack tests. (iree-org#10994)

  Show description for d30bd44
  hanhanW

  authored
  d30bd44

  Copy full SHA for d30bd44

• Cherry-pick llvm/llvm-project@b270fbe0 & llvm/llvm-project@c97035c4 (iree-org#10991)

  Show description for 76a4219
  antiagainst

  authored
  76a4219

  Copy full SHA for 76a4219

• Integrate llvm/llvm-project@cb84df210a47 (iree-org#10986)

  Show description for 131d96a
  

  antiagainst

  and

  Mahesh Ravishankar

  authored
  131d96a

  Copy full SHA for 131d96a

• NFC: Move matmul C matrix promotion related code to Common/ (iree-org#10989)

  Show description for acc70fb
  antiagainst

  authored
  acc70fb

  Copy full SHA for acc70fb

• Add tiling for cases where a broadcast is fused with a group reduction (iree-org#10984)

  Show description for bcf6fbd
  ThomasRaoux

  authored
  bcf6fbd

  Copy full SHA for bcf6fbd

• Adding flag for setting VMA preferredLargeHeapBlockSize. (iree-org#10983)

  Show description for fe3767d
  benvanik

  authored
  fe3767d

  Copy full SHA for fe3767d

• Fixing codegen warnings on MSVC.

  benvanik

  committed
  7cd0a8c

  Copy full SHA for 7cd0a8c

• Make the iree.bufferize transform op not call the IREEComprehensiveBufferizePass (iree-org#10964)

  Show description for e04d1e4
  nicolasvasilache

  authored
  e04d1e4

  Copy full SHA for e04d1e4

• Reenable transform dialect tests. (iree-org#10981)

  Show description for f4c9a01
  nicolasvasilache

  authored
  f4c9a01

  Copy full SHA for f4c9a01

• Allowing affine dialect in inline HAL conversion. (iree-org#10977)

  Show description for 1fadf32
  benvanik

  authored
  1fadf32

  Copy full SHA for 1fadf32

• Handle the case where PR trailers contain colons themselves (iree-org#10978)

  Show description for f646b2c
  GMNGeoffrey

  authored
  f646b2c

  Copy full SHA for f646b2c

Commits on Oct 31, 2022

• Fix parsing trailers when they're the only thing in the PR body (iree-org#10974)

  Show description for 00bc379
  GMNGeoffrey

  authored
  00bc379

  Copy full SHA for 00bc379

• Rework usage of testing templates (iree-org#10972)

  Show description for 0b8c3b8
  GMNGeoffrey

  authored
  0b8c3b8

  Copy full SHA for 0b8c3b8

• Fixing MSVC breakage from iree-org#10969.

  benvanik

  committed
  beeb0e8

  Copy full SHA for beeb0e8

• Tag runners to allow health checks (iree-org#10975)

  Show description for 458c784
  GMNGeoffrey

  authored
  458c784

  Copy full SHA for 458c784

• [spirv] Enable cooperative matrix flow for matmul-like generic (iree-org#10968)

  Show description for 467c675
  antiagainst

  authored
  467c675

  Copy full SHA for 467c675

• Tightening ukernel common definitions (iree-org#10950)

  Show description for f7c9bc8
  bjacob

  authored
  f7c9bc8

  Copy full SHA for f7c9bc8

• Cover all transposition cases of linalg_ext.unpack (iree-org#10941)

  bjacob

  authored
  ff68838

  Copy full SHA for ff68838

• Microkernel for linalg_ext.pack, take 2 (iree-org#10969)

  Show description for 36f6355
  bjacob

  authored
  36f6355

  Copy full SHA for 36f6355

Pagination

Previous

Next