Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Feb 13, 2026

• Integrate llvm/llvm-project@b4e4616342ca4c7a3c53b99d8d082fd17bfbbb10 (#23473)

  Show description for 21df09d
  

  rkayaith

  and

  claude

  authored
  21df09d

  Copy full SHA for 21df09d

Commits on Feb 12, 2026

• [Tests] Use llvm-cpu backend for transients_test.mlir (#23474)

  Show description for 0007d07
  

  rkayaith

  and

  claude

  authored
  0007d07

  Copy full SHA for 0007d07

• Integrate llvm/llvm-project@c6964b1b4dcda0f19f91cef186314f4cc2e0feb8 (#23462)

  Show description for 1baab97
  rkayaith

  authored
  1baab97

  Copy full SHA for 1baab97

• [Target][NFC] Mark internal classes final and wrap within anonymous namespaces (#23453)

  Show description for ef30f6d
  hanhanW

  authored
  ef30f6d

  Copy full SHA for ef30f6d

• [VectorExt] Add iree_vector_ext.arg_compare operation (#23386)

  Show description for e65d4a5
  bangtianliu

  authored
  e65d4a5

  Copy full SHA for e65d4a5

• Fix wrong output_shape of tensor.expand_shape ops. (#23469)

  Show description for 5c43c4d
  hanhanW

  authored
  5c43c4d

  Copy full SHA for 5c43c4d

• [Stream] Fix EmplaceTransientsPass to handle zero allocas case. (#23392)

  Show description for b960de5
  

  MaheshRavishankar

  and

  claude

  authored
  b960de5

  Copy full SHA for b960de5

Commits on Feb 11, 2026

• [NFC] Make ExpandShapeOp output_shape dynamic dims match result type in tests. (#23465)

  Show description for bd6c02d
  hanhanW

  authored
  bd6c02d

  Copy full SHA for bd6c02d

• Move Convert1X1FilterConv2DToMatmul pass from GlobalOptimization to Preprocessing (#23445)

  Show description for 9223106
  yzhang93

  authored
  9223106

  Copy full SHA for 9223106

• [DispatchCreation] Update split reduction heuristics for GEMM (#23423)

  Show description for 955917d
  yzhang93

  authored
  955917d

  Copy full SHA for 955917d

• [LinalgExt][NFC] Drop redundant function argument comments. (#23461)

  Show description for ea8ee31
  hanhanW

  authored
  ea8ee31

  Copy full SHA for ea8ee31

• Revert "Enable TensorDynamicDimAnalysis to propagate through `iree_codegen.load_from_buffer`" (#23467)

  Show description for a01814b
  hanhanW

  authored
  a01814b

  Copy full SHA for a01814b

• [LLVMCPU][AArch64] New heuristic for matmul vector tile sizes (#22932)

  Show description for 3af8a10
  momchil-velikov

  authored
  3af8a10

  Copy full SHA for 3af8a10

• Reapply "[Codegen][Common] Allow generic conv ops to decompose to lower dim ops (#23294)" (#23383) (#23455)

  Show description for 9c4c42a
  Abhishek-Varma

  authored
  9c4c42a

  Copy full SHA for 9c4c42a

• Enable TensorDynamicDimAnalysis to propagate through `iree_codegen.load_from_buffer` (#23431)

  Show description for 9748673
  ziereis

  authored
  9748673

  Copy full SHA for 9748673

• [LinalgExt] Cleanup TD definitions to avoid variadic everywhere (#23459)

  Show description for dad36f9
  

  qedawkins

  and

  claude

  authored
  dad36f9

  Copy full SHA for dad36f9

• [VectorDistribute] add support padding integer attention masks (#23430)

  Show description for b7437d4
  ziereis

  authored
  b7437d4

  Copy full SHA for b7437d4

Commits on Feb 10, 2026

• [Codegen][Tuner] Expose XOR shuffle bounds and validation functions in CAPI (#23442)

  Show description for b1d73d0
  Muzammiluddin-Syed-ECE

  authored
  b1d73d0

  Copy full SHA for b1d73d0

• Integrate llvm/llvm-project@a88274f008604dcc7397e8f231ef8eebe983793f (#23444)

  Show description for 2ffd825
  

  rkayaith

  and

  claude

  authored
  2ffd825

  Copy full SHA for 2ffd825

• Integrate llvm/llvm-project@fbed673f5238f219e43af0f88802d5422a77218b (#23439)

  Show description for 4b33ff5
  

  rkayaith

  and

  claude

  authored
  4b33ff5

  Copy full SHA for 4b33ff5

• [Compiler] Remove shard dialect dependency (#23133)

  Show description for 32a7159
  Groverkss

  authored
  32a7159

  Copy full SHA for 32a7159

• [e2e] Fix missed iree-hip-* to iree-rocm-* flag renames (#23443)

  Show description for 6f93dfc
  Yu-Zhewen

  authored
  6f93dfc

  Copy full SHA for 6f93dfc

• [Codegen] Create CombineSourceLayoutTransformation pass for MapGatherOp (#23165)

  Show description for 1320245
  Abhishek-Varma

  authored
  1320245

  Copy full SHA for 1320245

• [ROCM] Add workgroup reordering for data-tiling ukernels (#23358)

  Show description for 5579e4c
  Yu-Zhewen

  authored
  5579e4c

  Copy full SHA for 5579e4c

• Add additional search paths for tools and bitcode (#23399)

  Show description for 4802bac
  AaronStGeorge

  authored
  4802bac

  Copy full SHA for 4802bac

• [CPU] Plumb through tuning spec support for LLVMCPU. (#23424)

  Show description for dcde0c7
  hanhanW

  authored
  dcde0c7

  Copy full SHA for dcde0c7

• [CodeGen][NFCI] Introduce bass CodegenOption and move tuningSpecPath into it. (#23404)

  Show description for fa0fd44
  hanhanW

  authored
  fa0fd44

  Copy full SHA for fa0fd44

Commits on Feb 9, 2026

• [CodeGen] Generalize ConvertBf16ToUInt16Buffers to support fp8 types. (#23389)

  Show description for 209a486
  hanhanW

  authored
  209a486

  Copy full SHA for 209a486

• Allow the duplicate registeration if they point to the same storage. (#23438)

  Show description for 6744f1b
  hanhanW

  authored
  6744f1b

  Copy full SHA for 6744f1b

• [LLVMGPU] Fix barrier handling when pipelining with non-transfer_read ops (#23435)

  Show description for e13bc59
  Max191

  authored
  e13bc59

  Copy full SHA for e13bc59

• [CPU] Config LinalgExt::GatherOp's tiling sizes based on semantics. (#23419)

  Show description for 50e5b63
  hanhanW

  authored
  50e5b63

  Copy full SHA for 50e5b63

• [GPU] Introduce and plumb through GPUCodegenOptions. (#23290)

  Show description for 6f2612b
  hanhanW

  authored
  6f2612b

  Copy full SHA for 6f2612b

• [LinalgExt] Add mixed tensor-buffer semantics for map_scatter (#23334)

  Show description for 0df5f36
  qedawkins

  authored
  0df5f36

  Copy full SHA for 0df5f36

• [DT] Use overflow<nsw> in the storage size calculation. (#23422)

  Show description for d7de350
  hanhanW

  authored
  d7de350

  Copy full SHA for d7de350

• [Codegen][GPU] Modify constraint on swizzle_hint ops to allow for multiple types of swizzles (#23309)

  Show description for b62fac6
  Muzammiluddin-Syed-ECE

  authored
  b62fac6

  Copy full SHA for b62fac6