All pull requests

Pull requests

• [Stream] Fix O(N²) compile-time scaling in partitionRegionConcurrency

  #24958·isa-Lai opened Sep 22, 2026·

  ·
  1

  6

• [Flow] Prevent empty tensor refinement from breaking Stream lowering

  #24957·jschuhmacher opened Sep 22, 2026Contributor·

  ·
  3

• [LLVMCPU] Add missing f64 musl builtins for ONNX operations

  #24953·Prajit-Rahul opened Sep 21, 2026·

  ·
  4

• [Codegen][CPU] Lower Aarch64 inner_tiled.

  #24952·shuangjuliu opened Sep 21, 2026·

  ·
  1

• [GlobalOpt] Keep strided input maps when the producer can fuse

  #24950·zhangpan2001 opened Sep 21, 2026Contributor·

  ·
  1

• [GlobalOptimization] Fold empty-set reductions to their init operand.

  #24949·liuyalong opened Sep 21, 2026·

  ·
  12

• [LLVMGPU] Add RDNA3 e2e attention correctness tests

  #24948·weimin023 opened Sep 20, 2026Contributor·

  ·

• [Flow]Relaxes the assertion to accept non-negative operands in saturatingMul ofAnnotateDispatches .

  #24947·liuyalong opened Sep 20, 2026·

  ·
  1

• [VM] Preserve 64-bit values in internal calls

  #24946·zhangpan2001 opened Sep 20, 2026Contributor·

  ·
  1

  2

• [HAL] Add command buffer flush and invalidate operations

  #24945·purnima-nlp opened Sep 19, 2026·

  ·
  11

• [StableHLO] Lower reduce_window and scatter batching dims with dynamic shapes

  #24942·jschuhmacher opened Sep 17, 2026Contributor·5/5·

  ·

• [StableHLO] Lower dynamic_gather and dynamic_broadcast_in_dim

  #24941·jschuhmacher opened Sep 17, 2026Contributor·4/5·

  ·

• [StableHLO] Lower dynamic_reshape, dynamic_pad and dynamic_conv

  #24940·jschuhmacher opened Sep 17, 2026Contributor·3/5·

  ·

• [StableHLO] Lower the bounds encoding to util.assume.int

  #24939·jschuhmacher opened Sep 17, 2026Contributor·2/5·

  ·
  6

• [Codegen][CPU] Add RISC-V V f32 vfmacc.vf inner_tiled MMA.

  #24933·TMahlatini opened Sep 14, 2026Contributor·

  ·

• [LinalgExt] Add a output sizes < number of batch dims check in iree_linalg_ext.im2col verifier

  #24925·ayokunle321 opened Sep 11, 2026Contributor·

  ·
  1

• [Codegen][GPU] Add RDNA4 materialize encoding test

  #24924·Prajit-Rahul opened Sep 11, 2026·

  ·
  1

  4

• [Encode] Add a folder for set / unset encode round trip

  #24922·ayokunle321 opened Sep 11, 2026Contributor·

  ·
  1

• [LLVMCPU] Vectorize pack/unpack on AArch64 targets with vector masking (SVE/SME)

  #24919·FedericoBruzzone opened Sep 11, 2026Member·

  ·
  1

  21

• [Torch] Preserve PT2E quantization as LinalgExt ops

  #24918·ziereis opened Sep 11, 2026Contributor·5/5·

  ·

• [LLVMCPU] Add check for number of tile sizes in a tiling level's config

  #24917·ayokunle321 opened Sep 11, 2026Contributor·

  ·
  1

• [LLVMGPU] Add a check for the number of workgroup tile sizes in a config

  #24916·ayokunle321 opened Sep 11, 2026Contributor·

  ·
  3

• docs: document AMDGPU LDS transposing read instructions

  #24911·linsen458-spec opened Sep 10, 2026Contributor·

  ·
  1

  2

• [LinalgExt] Add group_matmul and group_mmt4d ops

  #24909·mbknust opened Sep 10, 2026Contributor·

  ·

• [Integrate] Update torch-mlir to e447c3fa

  #24908·ziereis opened Sep 10, 2026Contributor·

  ·
  2