Repository navigation
IREE v3.14.0 Release Notes
Release Candidate: iree-3.14.0rc20261115
PRs: 109 since v3.13.0
VMFB Bytecode Version: 17.0 (unchanged from v3.13.0)
HAL Module Version: 0.7 (unchanged from v3.13.0)
Breaking changes
iree_linalg_ext.sorttakes its inputs asins(). The op used to sort itsouts()operands in place. Inputs now go inins()and destination tensors inouts(), like other LinalgExt ops. Update hand-written IR and custom lowerings that create the op. #24658iree_linalg_ext.sortis no longer stable by default. The O(n²) bubble sort lowering was replaced with faster algorithms. Without the newis_stable(true)attribute, elements that the comparator does not order end up in unspecified order. The Torch input conversion setsis_stable(true), and the StableHLO conversion forwardsis_stablefromstablehlo.sort. #25262- Tracy is updated to 0.15. Use a Tracy 0.15 profiler, or the
iree-tracy-capturefrom this release, to capture traces.IREE_TRACE_MESSAGElevels map to Tracy message severities, and the exportedTracyClientCMake target now enablesTRACY_ENABLE. #25241 - The
iree-global-opt-convert-batch-matmul-to-matmulpass was removed. Broadcasts now fold into the indexing maps ofbatch_matmul(iree-global-opt-fold-broadcast-into-batch-matmul), and dispatch creation collapses the result into a plain matmul. #25114
Notable changes
Compiler
- Quantization:
- Quantized convolutions convert to integer math. im2col is now one generic rewrite for any convolution (grouped, dilated, 1-D/3-D, dynamic and integer), and it runs after the QDQ conversion. #25011, #25099
- Constant 4-bit and 2-bit quantized weights are packed into
i4/i2storage. #25145 - Broadcast and collapsed quantized weights reach the integer QDQ rewrite. #25114, #25197, #25086
- Torch: torchao
quantize_affine/dequantize_affineconvert to LinalgExt, and float zero points of PT2E and torchao ops that are zero or integral are converted. #25104, #25112, #25136
- Data tiling:
- Experimental early data tiling during global optimization with
--iree-opt-experimental-early-data-tiling. Layouts propagate through the surrounding tensor graph. #24969, #24968 - Named contractions with user indexing maps are collapsed like generic ones, so data tiling no longer depends on the named or generic form. #25113
- Experimental early data tiling during global optimization with
- New LinalgExt ops
group_matmulandgroup_mmt4d, the counterpart of PyTorch'sgrouped_mm, for Mixture-of-Experts models. #24909 - CPU:
- RISC-V V: f16 (
vfmacc,vfwmacc, Zvfhmin) and bf16 (vfwmaccbf16) inner_tiled MMA. #25000, #25120 - Packed f32 layouts for Wasm SIMD128. #25199
math.roundevenstays a native instruction on AArch64, x86 with SSE4.1 and riscv64 with D, instead of a software sequence. #25219- Invariant reductions are no longer fused into reduction tile loops, so tiled softmax no longer recomputes the row maximum for each tile. #25201
- RISC-V V: f16 (
- Reproducible
.vmfbfiles: LLVMCPU no longer writes uninitialized source location lines, and multi-device modules get deterministic allocator-select global names. #25239, #25243 - Compiler plugins can handle StableHLO composites and custom calls before input conversion, and C and Python API callers can activate plugins through the session. #25067, #25226
- Torch: ONNX reductions with a runtime axes tensor compile. #25041
- WebGPU: fixes for client API detection, wide elementwise vectors, read-only bindings and executable export names. #25051, #25053, #25215, #25195
- Correctness fixes:
- Attention normalization used the wrong operand order, which gave wrong results for a division. #25170
- GPU integer reductions used wrong identities for narrow and unsigned types. #25176
- Mutable globals that share a splat initializer are no longer updated through the same resource. #25030
- VM branches no longer leak a reference, and merged
scf.index_switchops keep their selectors. #24781, #25037 - Executable canonicalization and CSE in the Flow pipeline actually run. #25160
Runtime
- Python:
from_dlpack_capsuletakesmemory_typeandmemory_access. DLPack imports validate shapes and offsets and keep the producer alive, and buffers that legacy DLPack cannot describe, such as read-only ones, are not exported. #25227- HAL enums support
__index__, so they work as integer arguments with nanobind 2.15. #25163, #25223 - Memory leaks on error paths are fixed. #25014
- VM: the bytecode verifier checks register ordinals against the function's declared register counts. #25088
- HAL: drivers handle a device group being reassigned consistently. #25174
- WebGPU and WebAssembly: wasm64 support in the WebGPU bridge, and fixes for resource lifetimes in command buffers, staging buffer mappings, copies within one buffer and inline completion draining. #25054, #25075, #25070, #25068, #25069, #25193, #25245
Developer tools
- Release candidates have build provenance attestations. Verify a file with
gh attestation verify <file> --repo iree-org/iree. Stable releases keep them for theiree-disttarballs, which are shipped unchanged with their.sigstore.jsonbundles. Theiree-base-compilerandiree-base-runtimewheels change when they are promoted to the final version, so they have no attestation yet; attesting them is in progress in #25204. #25202, #25220 - Bazel: HAL drivers from outside IREE (like
IREE_EXTERNAL_HAL_DRIVERSin CMake), Android detection through@platforms//os:android, and include paths when IREE is an external repository. #25238, #25107, #25101 - CMake: builds against an installed LLVM/MLIR SDK honor tool overrides and no longer need
FileCheckwithIREE_BUILD_TESTS=OFF. #25167, #25146 - AMDGPU device bitcode comes from the official ROCm distribution, pinned in
build_tools/third_party/amdgpu_device_libs/rocm_package.json. #25133
New contributors
- @jonathansong made their first contribution in #25037
- @professor-moody made their first contribution in #25088
- @ttm02 made their first contribution in #25170
- @riestmo-nxp made their first contribution in #25142
- @rpqt made their first contribution in #25117
- @agourakis82 made their first contribution in #25176
- @npitre made their first contribution in #25030
Full changelog
Commit history: v3.13.0...v3.14.0
Release notes were AI-generated and edited.