commits  

Messages by Thread

• (tvm) branch main updated: [CMAKE] Upgrade TVM build baseline to C++20 (#19734) tqchen
• (tvm) branch main updated: [CPP_RPC] Bugfix race conditions and enhance print infos (#19778) tqchen
• (tvm) branch v0.25.0 updated: [CI] Merge PR against its target branch instead of main (#19712) (#19775) tqchen
• (tvm) branch nightly updated (d291513131 -> 21298c77ed) github-bot
• (tvm) branch main updated: [Relax][ONNX] Support exclusive option in CumSum (#19773) tlopex
• (tvm) branch main updated (4119c9fecd -> 44b55a05fb) tlopex
• (tvm) branch main updated: [REFACTOR][CUDA] Phase out l2 cache flush preproc test (#19768) tlopex
• (tvm) branch main updated: [MetaScheduler] Improve print info about builder/runner state (#19767) tqchen
• (tvm) branch main updated (e43555f739 -> 8b79dfdb80) bohan
• (tvm) branch main updated: [REFACTOR][DataType] Phase out target custom datatype support (#19760) tqchen
• (tvm) branch main updated: [CI] Drop redundant cmake/ninja install from the Linux wheel CUDA sidecar (#19761) tqchen
• (tvm) branch main updated: [Runtime][CoreML] Fix FFI casts in CoreML runtime (#19762) tqchen
• (tvm) branch nightly updated (6f9609f10a -> d291513131) github-bot
• (tvm) branch main updated: [REFACTOR][VM] Move CUDA graph VM builtin back under VM runtime (#19758) tlopex
• (tvm-ffi) branch main updated: Adopt AGENTS.md for agent guidance (#620) ruihangl
• (tvm) branch main updated: [Relax][PyTorch] Add logical_or and logical_xor converters (#19756) tlopex
• (tvm) branch main updated: [Tests][AArch64] Make SVE codegen assertions robust across LLVM versions (#19752) tqchen
• (tvm) branch main updated: [S-TIR][CUDA] Fix legacy predicated cp.async zero fill (#19741) tqchen
• (tvm) branch main updated: [TIRX][Tests] Fix LLVM version gate for vectorized lround (#19753) tqchen
• (tvm) branch main updated: [CI] Updated cibw to 4.1.0 (#19754) tqchen
• (tvm) branch nightly updated (2a16d98b9e -> 6f9609f10a) github-bot
• (tvm) branch main updated: [CI] Run s_tir/transform tests in the python-unittest stage (#19737) tlopex
• (tvm) branch main updated: [S-TIR][Tests] Mark test_cp_async_in_if_then_else as xfail (#19751) bohan
• (tvm-ffi) branch main updated: [REFACTOR][STUBGEN] refactor the stubgen logic. (#608) tqchen
• (tvm) branch main updated: [Tests][LLVM] Gate stepvector intrinsic rename on LLVM 20 (#19745) tqchen
• (tvm) branch main updated: [Runtime][Tests] Drop int4 from random_fill test, fix dtype error message (#19748) tqchen
• (tvm) branch main updated: [Runtime][Disco] Fix session attribute storage, NVSHMEM build, and test gating (#19736) tqchen
• (tvm) branch main updated: [Codegen][LLVM][Tests] Gate +v9a vscale_range expectation on LLVM version (#19744) tqchen
• (tvm) branch main updated: [CI] Align cuda-python with PyTorch cuda-bindings (#19738) tqchen
• (tvm) branch main updated: [Hexagon][Tests] Clean up stale hexagon tests (#19747) tqchen
• (tvm) branch main updated: [Tests] Fix flaky popen pool executor test (#19746) tqchen
• (tvm) branch nightly updated (a4a2bf5b89 -> 2a16d98b9e) github-bot
• (tvm) branch main updated: [S-TIR] Fix software pipeline offsets for legacy MMA intrinsics (#19742) tlopex
• (tvm) branch main updated: [Tests] Check WebGPU volatile allreduce annotation structurally (#19740) tlopex
• (tvm) branch main updated: [TIRx] Use canonical PTX async script API in s_tir test (#19739) tlopex
• (tvm) branch investigate-s-tir-nested-pipeline-mma updated (3705f396e4 -> e56ca66dcd) tlopex
• (tvm) branch investigate-s-tir-nested-pipeline-mma updated (adbd870e8f -> 3705f396e4) tlopex
• (tvm) branch investigate-s-tir-nested-pipeline-mma updated (f180256f0a -> adbd870e8f) tlopex
• (tvm) branch investigate-s-tir-nested-pipeline-mma created (now f180256f0a) tlopex
  • (tvm) 01/01: Fix software pipeline legacy MMA offsets tlopex
• (tvm) branch main updated: [S-TIR][Tests] Fix transform test failures after TIRx bringup (#19735) tlopex
• (tvm) branch main updated: [Script][Tests] Fix dialect redirect module re-execution and stray category-less tirx.intrin_test op (#19731) tqchen
• (tvm) branch main updated: [PYTHON] Autoload backends; simplify library loading; remove TVMError for native errors (#19727) bohan
• (tvm) branch main updated (3e201a6aa5 -> e9ead6e670) tqchen
• (tvm) branch main updated: [TIRx] Preserve Triton call_kernel compile options (#19728) tqchen
• (tvm) branch main updated: [Tests] Remove test_runtime_ndarray (covered by tvm-ffi) (#19715) tqchen
• (tvm) branch main updated: [Tests][S-TIR] Fix stale MetaSchedule sketch expectations and migrate let binds to T.let (#19729) tqchen
• (tvm) branch main updated: [Tests][NNAPI] Skip tests cleanly when remote environment is unavailable (#19730) tqchen
• (tvm) branch nightly updated (def37e30bd -> a4a2bf5b89) github-bot
• (tvm) branch main updated: [Python] Refactor pyproject.toml dependencies (#19723) tlopex
• (tvm) branch main updated: [Tests][Hexagon] Lazily import pytest plugin dependencies (#19726) tqchen
• (tvm-site) branch asf-site updated: deploying docs (apache/tvm@a1d46dc5ab77d33991d20cc3f6ef1486a8377a46) tqchen
• (tvm) branch main updated: [Tests][Relax] Gate multi-GPU VM test on three devices (#19725) tqchen
• (tvm) branch main updated: [Tests][Disco] Skip CCL tests when runtime support is absent (#19724) tqchen
• (tvm) branch main updated: [Runtime][Tests] Fix contrib wheel tests (#19714) tqchen
• (tvm) branch main updated: [REFACTOR][PYTHON] Remove tvm.ffi shim; import tvm_ffi directly (#19721) tqchen
• (tvm) branch main updated: [Codegen][NVPTX] Skip runtime execution in Vulkan codegen tests (#19717) tqchen
• (tvm-ffi) branch main updated: Robustify FromJSONGraph (#619) tqchen
• (tvm) branch main updated: [REFACTOR][PYTHON] Slim tvm.libinfo to info-only helpers (#19719) bohan
• (tvm) branch main updated: [DOCS] Clarify loading serialized artifacts requires a trusted source (#19720) tqchen
• (tvm) branch main updated: [Codegen][LLVM] Accept splat form in VLA broadcast test (#19716) tqchen
• (tvm-ffi) branch main updated: [FEAT] Add structural visitor and typed structural walk APIs (#601) tqchen
• (tvm) branch main updated: [CI] Fix CI script test subprocess environment (#19713) tqchen
• (tvm) branch main updated (def37e30bd -> 02280c1654) tqchen
• (tvm) branch nightly updated (15b1d98393 -> def37e30bd) github-bot
• (tvm) branch main updated: [Tests] Skip test modules cleanly when optional deps are missing (#19704) tlopex
• (tvm) branch main updated: [Relax][CoreML] Fix CoreML partition pass (#19711) tlopex
• (tvm) branch main updated: [Python] Bump apache-tvm-ffi floor to >=0.1.12 (#19710) tlopex
• (tvm) branch main updated: [Refactor][Meta-schedule] Remove meta-schedule as_string mechanism in favor of default representation (#19709) tlopex
• (tvm) branch ruihanglai/ci-merge-target created (now 13298638aa) ruihangl
• (tvm) branch main updated: [CI] Fix release verification script (#19700) tqchen
• (tvm) branch main updated: [Tests] Fix s_tir tests using removed T.block API in TIRx script (#19706) bohan
• (tvm-ffi) branch main updated (59da4c0 -> 7c077e4) tqchen
• (tvm) branch main updated: [CI] Pin actions by version tag, trim wheel perms (#19703) tqchen
• svn commit: r85127 - dev/tvm/tvm-v0.25.0-rc0 ruihangl
• (tvm) tag v0.25.0.rc0 updated (7e93b733be -> 5ec68440e5) ruihangl
• (tvm) branch v0.25.0 updated: [Python] Bump apache-tvm-ffi floor to >=0.1.12 on v0.25.0 (#19701) ruihangl
• (tvm) branch fix-tirx-cuda-sm-guards updated (a2ea8ea39d -> 911a9d7e29) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (b630aa07d0 -> a2ea8ea39d) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (7441e4be94 -> b630aa07d0) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (b480a8125d -> 7441e4be94) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (6916239ce4 -> b480a8125d) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (37cc404b02 -> 6916239ce4) bohan
• (tvm) branch fix-tirx-cuda-sm-guards updated (6c7fe202eb -> 37cc404b02) bohan
• (tvm) branch main updated: [Fix] CommReduce could handle 0-dim data (#19683) tlopex
• (tvm) branch fix-tirx-cuda-sm-guards updated (7b1c7124ae -> 6c7fe202eb) bohan
• (tvm) branch fix-tirx-cuda-sm-guards created (now 7b1c7124ae) bohan
  • (tvm) 01/01: [Tests] Guard TIRX CUDA tests by compute capability bohan
• svn commit: r85126 - release/tvm/tvm-ffi-v0.1.12 ruihangl
• svn commit: r85125 - release/tvm/tvm-ffi-v0.1.12 ruihangl
• (tvm-ffi) tag v0.1.12-rc0 deleted (was 3050b0a) ruihangl
• (tvm-ffi) annotated tag v0.1.12 updated (3050b0a -> 2ab1aaf) ruihangl
• svn commit: r85114 - dev/tvm/tvm-v0.25.0-rc0 ruihangl
• (tvm) tag v0.25.0.rc0 updated (13012485fa -> 7e93b733be) ruihangl
• (tvm) branch v0.25.0 updated: [CI] Target apache-tvm for PyPI wheel publishing (#19697) bohan
• (tvm) branch nightly updated (498b8ed673 -> 15b1d98393) github-bot
• (tvm) branch main updated: [Web] Bump tvmjs version to 0.25.0-dev1 (#19687) akaashrp
• (tvm) branch main updated: [CI] Target apache-tvm for PyPI wheel publishing (#19686) ruihangl
• (tvm) branch main updated: [CI] Remove PyPI-only tag ref guard from wheel publishing (#19685) ruihangl
• svn commit: r85102 - in dev/tvm: tvm-v0.25.0-rc0 tvm-v0.25.0.rc0 ruihangl
• (tvm) branch main updated: [Relax][PyTorch] Cast non-bool inputs to bool in logical_and converter (#19679) tlopex
• svn commit: r85101 - dev/tvm ruihangl
• svn commit: r85100 - dev/tvm/tvm-v0.25.0.rc0 ruihangl
• (tvm) tag v0.25.0.rc0 created (now 13012485fa) ruihangl
• (tvm) tag v0.25.0.rc0 deleted (was ede9267a5e) ruihangl
• (tvm) branch v0.25.0 updated: [RELEASE] Bump web npm version to 0.25.0 (#19684) tqchen
• (tvm) tag v0.25.0.rc0 created (now ede9267a5e) ruihangl
  • (tvm) 01/01: [RELEASE] Bump web npm version to 0.25.0 ruihangl
• (tvm) branch v0.25.0 created (now b172d5ea32) ruihangl
• (tvm) tag v0.26.dev0 created (now b172d5ea32) ruihangl
• (tvm) branch nightly updated (9db74c7cee -> 498b8ed673) github-bot
• (tvm) branch main updated (aa59644072 -> 498b8ed673) tlopex
• (tvm-ffi) branch main updated: [TEST] Relax test_shared_dag_hash_scaling_not_exponential ratio to 4x (#612) ruihangl
• (tvm) branch main updated: [Relax][ONNX] Preserve NaN in Sign to align with ONNX Runtime (#19674) tlopex
• (tvm) branch nightly updated (4d9d129c93 -> 9db74c7cee) github-bot
• svn commit: r84996 - dev/tvm/tvm-ffi-v0.1.12-rc0 ruihangl
• (tvm-ffi) tag v0.1.12-rc0 created (now 3050b0a) ruihangl
• svn commit: r84995 - release/tvm ruihangl
• (tvm-ffi) branch main updated: [FIX] Drop test_empty_tensor_attributes (numpy dlpack stride change) (#607) ruihangl
• (tvm) branch main updated (4d9d129c93 -> 9db74c7cee) tqchen
• (tvm-ffi) branch tvm-ffi-fix-sphinx-autosummary-warnings-for-tensor-methods-in-docs-reference-python-index-rst created (now d62818c) tqchen
  • (tvm-ffi) 01/01: [DOC] Move tensor methods (ndim, numel, size, is_contiguous) under Tensor autosummary tqchen
• (tvm) branch nightly updated (1240649257 -> 4d9d129c93) github-bot
• (tvm-ffi) branch main updated: feat(python): add tensor methods to align with C++ APIs (#604) yaxingcai
• (tvm) branch main updated: [Relax][ONNX] Fix Cast operator float->int NaN/Inf handling (#19626) tlopex
• (tvm) branch main updated: [Arith] Gate canonical-simplify LT Case 2 on extra scale == +1 (#19669) tqchen
• (tvm) branch main updated: [Arith] Memoize IntervalSet variable relaxation to avoid exponential blowup (#19670) tqchen
• (tvm) branch nightly updated (a979b2f98c -> 1240649257) github-bot
• (tvm) branch main updated: [FFI][REFACTOR] Direct structural APIs to tvm-ffi (#19661) tlopex
• (tvm) branch main updated: [CI] Reformat the macOS repair-wheel-command as a multiline script (#19664) tlopex
• (tvm) branch main updated: [CI] Derive the version from Git tags via setuptools_scm (#19665) tqchen
• (tvm) branch main updated: [Relax][PyTorch] Decompose integer pow into repeated multiplication (#19660) tlopex
• (tvm) branch tvm-direct-tvm-ffi-structural-apis updated (670a5a32a4 -> b43dc89827) tqchen
• (tvm) branch main updated: [FFI][IR] Route JSON serialization through tvm-ffi (#19662) bohan
• (tvm) branch main updated: [REFACTOR][TIRX] Consolidate split host device stages (#19663) bohan
• (tvm) branch tvm-direct-tvm-ffi-structural-apis updated (3c42a3ef0d -> 670a5a32a4) tqchen
• (tvm) branch tvm-splithostdevice-staged-pass-consolidation updated (85cb3e77f8 -> 5d42bd6859) tqchen
• (tvm) branch tvm-direct-json-graph-serialization-apis updated (806bb47ada -> e55ae44cff) tqchen
• (tvm) branch tvm-direct-json-graph-serialization-apis updated (6a6f6639a5 -> 806bb47ada) tqchen
• (tvm) branch tvm-direct-json-graph-serialization-apis created (now 6a6f6639a5) tqchen
  • (tvm) 01/01: [FFI][IR] Route JSON serialization through tvm-ffi tqchen
• (tvm) branch tvm-splithostdevice-staged-pass-consolidation created (now 85cb3e77f8) tqchen
  • (tvm) 01/01: [REFACTOR][TIRX] Consolidate split host device stages tqchen
• (tvm) branch tvm-direct-tvm-ffi-structural-apis-test-permission deleted (was da9b5803ac) tqchen
• (tvm) branch tvm-direct-tvm-ffi-structural-apis created (now 3c42a3ef0d) tqchen
  • (tvm) 01/01: [FFI][REFACTOR] Direct structural APIs to tvm-ffi tqchen
• (tvm) branch tvm-direct-tvm-ffi-structural-apis-test-permission created (now da9b5803ac) tqchen
• (tvm) branch main updated: [CI] Wheel publishing follow-ups (#19659) tqchen
• (tvm) branch nightly updated (8039963c23 -> a979b2f98c) github-bot
• (tvm) branch main updated: [RPC] Import tvm.testing lazily in rpc.testing (#19658) tqchen
• (tvm) branch main updated: [CI] Add cibw-based wheel publishing to PyPI (#19656) tqchen
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add EMBEDDING_LOOKUP_SPARSE converter (#19652) tlopex
• (tvm) branch main updated: [Fix][Relax] Support ND batched matmul chains in AdjustMatmulOrder pass (#19650) tlopex
• (tvm) branch nightly updated (cf859b927a -> 8039963c23) github-bot
• (tvm) branch static-libxml2 updated (05506afc7f -> d7435df17f) tqchen
  • (tvm) 01/01: [BUILD] Static-libxml2 + LLVM static build cross-platform support tqchen
• (tvm) branch main updated: [Web][COS] Persist URL→hash mapping across page loads (#19569) akaashrp
• (tvm) branch main updated: [Relax][PyTorch] Cast non-bool inputs to bool in logical_not converter (#19645) tlopex
• (tvm) branch main updated: fix: Security Patch: Fix missing exported flag in AndroidManifest (#19648) tlopex
• (tvm) branch main updated: [Relax][Frontend][TFLite] Support STABLEHLO_RNG_BIT_GENERATOR (#19651) tlopex
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add HASHTABLE_LOOKUP converter (#19654) tlopex
• (tvm) branch static-libxml2 updated (755e96275a -> 05506afc7f) tqchen
  • (tvm) 01/01: [BUILD] Static-libxml2 + LLVM static build cross-platform support tqchen
• (tvm) branch static-libxml2 updated: [BUILD] Simplify static libxml2 substitution comment tqchen
• (tvm) branch static-libxml2 updated: [BUILD] Expand comment around static libxml2 substitution tqchen
• (tvm) branch static-libxml2 created (now 609ab46c8f) tqchen
  • (tvm) 01/01: [BUILD] Static-libxml2 + LLVM static build cross-platform support tqchen
• (tvm) branch main updated: [REFACTOR][PYTHON] Revisit lifted support modules from tvm.contrib (#19653) tqchen
• (tvm) branch nightly updated (fca86a6872 -> cf859b927a) github-bot
• (tvm) branch main updated: [Relax][Frontend][TFLite] Support STABLEHLO_CUSTOM_CALL (#19649) tlopex
• (tvm) branch main updated: [Relax][IR] Skip in-place multiply when two operands are views of the same tensor (#19644) tlopex
• (tvm) branch nightly updated (b971a75de4 -> fca86a6872) github-bot
• (tvm) branch main updated: [Fix] Stabilize layer_norm variance computation with two-pass reduction (#19643) tlopex
• (tvm) branch main updated: [Relax][Frontend][TFLite] Support STABLEHLO_WHILE (#19646) tlopex
• (tvm) branch pypi deleted (was 188cb53337) tlopex
• (tvm) branch pypi created (now 188cb53337) tlopex
• (tvm-ffi) branch main updated: [CORE] Make AnyView trivially copyable — match C-ABI struct passing (#602) junrushao
• (tvm) branch main updated: [Relax][Frontend][TFLite] Support sequence LSTM and RNN operators (#19634) tlopex
• (tvm) branch main updated: [TIRx] Fix stale Simplify import in lowering test (#19642) guanmingchiu
• (tvm) branch nightly updated (e89570fa83 -> b971a75de4) github-bot
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add TFLite Resource Variable and Static Hashtable Import Support (#19639) tlopex
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add LSTM and SVDF converter (#19633) tlopex
• (tvm) branch nightly updated (61ae85b9d1 -> e89570fa83) github-bot
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add RNN converter (#19632) tlopex
• (tvm) branch main updated: [Relax][Frontend][TFLite] Add REDUCE_WINDOW support (#19637) tlopex
• (tvm) branch tir-bench deleted (was cd8790bc54) bohan
• (tvm) branch tir-bench created (now cd8790bc54) bohan
  • (tvm) 22/22: feat(infra): util-gated GPU selection for tir-bench + per-config warmup/repeat bohan
  • (tvm) 07/22: feat(tvmscript): add @Tx.jit decorator, Tx.constexpr params, Tx.wg_reg_tile (#635) bohan
  • (tvm) 04/22: refactor(op-dispatch): rename copy/warp_matrix.py to copy/ld_stmatrix.py (#631) bohan
  • (tvm) 18/22: feat(gemm_async): accept Layout F C operand for M=64 MMAs (#648) bohan
  • (tvm) 21/22: fix(arith): memoize IntervalSet variable relaxation to avoid exponential blowup (#652) bohan
  • (tvm) 16/22: feat(tirx): add M=128 dispatch + layout for .16x*b tcgen05.ld/st (#646) bohan
  • (tvm) 03/22: feat(op-dispatch): add warp ldmatrix/stmatrix dispatch for Tx.copy (#630) bohan
  • (tvm) 14/22: feat(tirx): add .16x{64,128,256}b tcgen05.ld/st dispatch + factory (#644) bohan
  • (tvm) 17/22: revert(submodules): undo accidental 3rdparty pointer bumps in #646 (#647) bohan
  • (tvm) 06/22: refactor(codegen): remove tirx.entry_cluster_sync codegen attribute (#634) bohan
  • (tvm) 05/22: test(op-dispatch): add wg-scope both-sides-permuted invariance test (#632) (#633) bohan
  • (tvm) 11/22: feat(tirx): add typed pointer byte-offset intrinsic (#641) bohan
  • (tvm) 09/22: fix(tirx/stmt_functor): add ScopeIdDefStmt to Python StmtFunctor dispatch (#637) bohan

• Earlier messages