Issues

• [Stream] RefineUsage never terminates on a concretely typed await result

  bug 🐞
  Bug

  iree-org/iree#25048 · Filter by author ziereis opened on Oct 14, 2026

• [Stream] SplitParameterEncoder ignores the offset of parameter subranges

  bug 🐞
  Bug

  iree-org/iree#25047 · Filter by author ziereis opened on Oct 14, 2026

• [GlobalOptimization] Rebase quantized operands to the signedness the target's integer dot product wants

  enhancement ➕
  Enhancement

  iree-org/iree#25036 · Filter by author ziereis opened on Oct 13, 2026

• [GlobalOptimization] Pack constant QDQ weights with sub-byte quantization ranges into i4/2.. storage

  enhancement ➕
  Enhancement

  iree-org/iree#25035 · Filter by author ziereis opened on Oct 13, 2026

• [GlobalOptimization] Pass ordering keeps im2col convolutions in quantized conv models from lowering to integer math

  bug 🐞cleanup 🧹performance ⚡
  Bug

  iree-org/iree#25034 · Filter by author ziereis opened on Oct 13, 2026

• Move the early data-tiling layout decision into its own component

  iree-org/iree#25031 · Filter by author maxbartel opened on Oct 13, 2026

• Mutable globals with identical splat initializers alias after ElideAsyncCopies (silent miscompile)

  iree-org/iree#25029 · Filter by author npitre opened on Oct 12, 2026
  1

• [Codegen][CPU] inner_tiled does not scale past 1 thread: distribution tile is always 1.

  iree-org/iree#25028 · Filter by author TMahlatini opened on Oct 12, 2026

• StableHLO stable sort reverses equal-key payloads on llvm-cpu

  iree-org/iree#25022 · Filter by author vera-rubin opened on Oct 11, 2026
  1

• [Codegen] LLVMGPU Untiled outer reduction dimensions cause excessive IR expansion

  bug 🐞
  iree-org/iree#25016 · Filter by author Cerulime opened on Oct 9, 2026
  3

  

• [LLVM] Investigate VectorToGPU regression after llvm/llvm-project#221795

  iree-org/iree#25013 · Filter by author phemashekar opened on Oct 8, 2026

• [Codegen] LLVMGPU ContractToChainFMA skips scalar dot products and mispairs permuted reduction dimensions

  bug 🐞
  iree-org/iree#25007 · Filter by author Cerulime opened on Oct 7, 2026
  2

• [Codegen] Shared-memory padding produces invalid IR for loop-carried memrefs

  bug 🐞codegen
  iree-org/iree#25005 · Filter by author Cerulime opened on Oct 7, 2026
  3

• [Codegen] Static scf.forall workgroup counts can exceed the target limit after translation-info reconciliation

  bug 🐞codegen
  iree-org/iree#24999 · Filter by author Cerulime opened on Oct 6, 2026
  2

• Release Tracker v3.13.0 (2026-10-19)

  iree-org/iree#24996 · Filter by author devtbi opened on Oct 5, 2026
  

• [LLVMGPU] Missing args in calling createOptimizeVectorTransferPass

  iree-org/iree#24991 · Filter by author Cerulime opened on Oct 3, 2026
  3

• [gpu] NVIDIA FP8 mma.sync lowering hard-codes an eight-element lhs reshape

  iree-org/iree#24956 · Filter by author snarang181 opened on Sep 22, 2026
  3

• [codegen] NestedLayoutAttr::reshape changes thread ownership for non-contiguous shape casts

  iree-org/iree#24955 · Filter by author snarang181 opened on Sep 22, 2026
  2

• [VMVX] Support vector operands in Math → VM conversions

  iree-org/iree#24937 · Filter by author ayokunle321 opened on Sep 17, 2026
  3

• Tracking: [RVV] inner_tiled performance vs mmt4d ukernels

  iree-org/iree#24932 · Filter by author TMahlatini opened on Sep 14, 2026
  5

• Multi-device Zero-Copy Path: Missing lowering for stream.cmd.flush/invalidate and zero-copy path

  enhancement ➕
  iree-org/iree#24930 · Filter by author Manewing opened on Sep 14, 2026
  7

• [Stream] SplitParameterEncoding in consolidate mode doesn't work

  bug 🐞
  Bug

  iree-org/iree#24920 · Filter by author ziereis opened on Sep 11, 2026
  4

  

• [CPU] Eliminate cross-dispatch pack/unpack overhead when tensor.pad intervenes between producer and consumer

  enhancement ➕
  iree-org/iree#24914 · Filter by author Murari007 opened on Sep 10, 2026
  2

• [LLVMCPU] Data-tiling pack/unpack is vectorized on x86 but not on AArch64

  enhancement ➕
  Enhancement

  iree-org/iree#24888 · Filter by author FedericoBruzzone opened on Sep 7, 2026
  1

• [StableHLO] GatherToTorchIndexSelect drops the index clamp, so out-of-bounds gathers read arbitrary memory

  iree-org/iree#24881 · Filter by author sebffischer opened on Sep 5, 2026
  2