discuss-archive  

Messages by Thread

• [PR] [TIRx][Test] Gate tcgen05.mma collector certification on nvcc >= 13.4 [tvm] via GitHub
• [GH] (tvm/issue-20200): Workflow run "CI" is working again! GitBox
• [I] [Bug] Allocation/shape arithmetic wraps in storage planning and workspace lowering [tvm] via GitHub
• [PR] Main dev/2026 09 03/stubgen rust complete [tvm-ffi] via GitHub
• [GH] (tvm/issue-20200): Workflow run "Lint" is working again! GitBox
• [GH] (tvm-ffi/task-351-container-mutate-hook): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/task-351-container-mutate-hook): Workflow run "CI" failed! GitBox
• [PR] [PERF][FFI] Streamline container mutation hooks [tvm-ffi] via GitHub
  • Re: [PR] [REFACTOR][EXTRA] Streamline container mutation hooks [tvm-ffi] via GitHub
• [GH] (tvm/issue-20200): Workflow run "CI" failed! GitBox
• [PR] [Relax] Exclude R.null_value()-bound vars from KillAfterLastUse [tvm] via GitHub
• [GH] (tvm/issue-20200): Workflow run "Lint" failed! GitBox
• [PR] [CUDA][TIRx] Add collector-qualified tcgen05.mma block_scale forms and bind tcgen05.ld.red redval as output [tvm] via GitHub
  • Re: [PR] [CUDA][TIRx] Add collector-qualified tcgen05.mma block_scale forms and bind tcgen05.ld.red redval as output [tvm] via GitHub
• [PR] [PERF] Use adaptive identity table in StructuralMap [tvm-ffi] via GitHub
  • Re: [PR] [PERF] Use adaptive identity table in StructuralMap [tvm-ffi] via GitHub
• [GH] (tvm-ffi/main-dev/2026-09-03/stubgen_rust_opaque): Workflow run "CI" failed! GitBox
• [PR] [STUBGEN][RUST] Add the Rust backend with opaque object bindings [tvm-ffi] via GitHub
  • Re: [PR] [STUBGEN][RUST] Add the Rust backend with opaque object bindings [tvm-ffi] via GitHub
• [PR] [TIR] Ignore None-valued pragma annotations [tvm] via GitHub
• [PR] [STUBGEN] Generalize one-line directive handling [tvm-ffi] via GitHub
  • Re: [PR] [STUBGEN] Generalize one-line directive handling [tvm-ffi] via GitHub
• [I] [VOTE] Release Apache TVM FFI v0.1.14-rc2 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc2 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc2 [tvm-ffi] via GitHub
• [PR] [RUST] Add borrowed object node cast [tvm-ffi] via GitHub
  • Re: [PR] [RUST] Add borrowed object node cast [tvm-ffi] via GitHub
• [GH] (tvm-ffi/task-348-expected-move-access): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/task-348-expected-move-access): Workflow run "CI" failed! GitBox
• [PR] Optimize StructuralMap [tvm-ffi] via GitHub
  • Re: [PR] Optimize StructuralMap [tvm-ffi] via GitHub
  • Re: [PR] [REFACTOR][EXTRA] Optimize StructuralMap [tvm-ffi] via GitHub
• [I] [STUBGEN][QUESTION] Can we unify the usage of `ty-map` in stubgen? [tvm-ffi] via GitHub
• [GH] (tvm-ffi/ffi-object-unique-no-weak-refs): Workflow run "CI" is working again! GitBox
• [PR] [FFI] Make Object::unique() require no weak references [tvm-ffi] via GitHub
  • Re: [PR] [FFI] Make Object::unique() require no weak references [tvm-ffi] via GitHub
• [GH] (tvm-ffi/main-dev/2026-09-02/stubgen_layout_classify): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/main-dev/2026-09-02/stubgen_layout_classify): Workflow run "CI" failed! GitBox
• [PR] [STUBGEN] Classify native layouts from reflected byte facts [tvm-ffi] via GitHub
  • Re: [PR] [STUBGEN] Classify native layouts from reflected byte facts [tvm-ffi] via GitHub
• [PR] [CUDA][TIRx] PTX ISA 9.4 / CUDA 13.4 support for SM103a and SM107a (Rubin) [tvm] via GitHub
  • Re: [PR] [CUDA][TIRx] PTX ISA 9.4 / CUDA 13.4 support for SM103a and SM107a (Rubin) [tvm] via GitHub
• [I] [Bug][Relax] ONNX Slice with runtime negative starts gives wrong results (zeros) — dynamic_strided_slice mishandles negative begin [tvm] via GitHub
• [GH] (tvm/main): Workflow run "npm_and_yarn in /web for js-yaml - Update #1552104634" failed! GitBox
• [PR] Bump browserslist from 4.28.6 to 4.28.8 in /web [tvm] via GitHub
• [I] [VOTE] Release Apache TVM FFI v0.1.14-rc1 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc1 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc1 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc1 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc1 [tvm-ffi] via GitHub
• [I] [Bug][Relax] strided_slice with symbolic shape: out-of-bound begin/end are clipped for the loop extent but not for the index expression, producing out-of-bound reads and silent wrong values [tvm] via GitHub
• [GH] (tvm/pr/ptx-isa-9.4-cuda-13.4-sm103a-sm107a): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/pr/ptx-isa-9.4-cuda-13.4-sm103a-sm107a): Workflow run "Lint" failed! GitBox
• [I] [Bug][Relax] FoldConstant crashes on shape_to_tensor with symbolic shapes (AssertionError) and folds it into a Constant that breaks reshape [tvm] via GitHub
• [PR] [CUDA] Add NVIDIA Jetson AGX Thor target tag [tvm] via GitHub
  • Re: [PR] [CUDA] Add NVIDIA Jetson AGX Thor target tag [tvm] via GitHub
• [GH] (tvm/upstream/metal-symbolic-stack-allocations): Workflow run "CI" is working again! GitBox
• [PR] [REFACTOR][RUST] Keep mutate dispatch state on the dispatcher [tvm-ffi] via GitHub
  • Re: [PR] [REFACTOR][RUST] Keep mutate dispatch state on the dispatcher [tvm-ffi] via GitHub
• Re: [I] [Bug][Relax] `conv2d_transpose` produces wrong results on CUDA when `output_padding > 0` [tvm] via GitHub
• [GH] (tvm/fix/conv2d-transpose-output-padding-19524): Workflow run "CI" failed! GitBox
• [GH] (tvm/fix/conv2d-transpose-output-padding-19524): Workflow run "Lint" failed! GitBox
• [I] TVMFFIBacktrace crashes under concurrency on Windows: DbgHelp used without serialization, per-call SymInitialize/SymCleanup, unchecked SymInitialize [tvm-ffi] via GitHub
  • Re: [I] TVMFFIBacktrace crashes under concurrency on Windows: DbgHelp used without serialization, per-call SymInitialize/SymCleanup, unchecked SymInitialize [tvm-ffi] via GitHub
  • Re: [I] TVMFFIBacktrace crashes under concurrency on Windows: DbgHelp used without serialization, per-call SymInitialize/SymCleanup, unchecked SymInitialize [tvm-ffi] via GitHub
  • Re: [I] TVMFFIBacktrace crashes under concurrency on Windows: DbgHelp used without serialization, per-call SymInitialize/SymCleanup, unchecked SymInitialize [tvm-ffi] via GitHub
• [PR] Bump pip from 22.1.1 to 26.2 in /docker/python [tvm] via GitHub
• [I] [VOTE] Release Apache TVM FFI v0.1.14-rc0 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc0 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc0 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc0 [tvm-ffi] via GitHub
  • Re: [I] [VOTE] Release Apache TVM FFI v0.1.14-rc0 [tvm-ffi] via GitHub
• [GH] (tvm-ffi/main-dev/2026-08-31/stubgen_layout_facts): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/feat-ffi-flexible-structuralmap-returns): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat-ffi-flexible-structuralmap-returns): Workflow run "CI" failed! GitBox
• [PR] [FEAT][FFI] Accept flexible callback return types in C++ StructuralMap [tvm-ffi] via GitHub
  • Re: [PR] [FEAT][FFI] Accept flexible callback return types in C++ StructuralMap [tvm-ffi] via GitHub
• [PR] [REFACTOR][IR] Make expression subscription eager and remove SubscriptProxy [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Make expression subscription eager and remove SubscriptProxy [tvm] via GitHub
• [GH] (tvm/tvm-bufferregion-typed-expr-explicit-load): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/tvm-bufferregion-typed-expr-explicit-load): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tvm-bufferregion-typed-expr-explicit-load): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tvm-bufferregion-typed-expr-explicit-load): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tvm-bufferregion-typed-expr-explicit-load): Workflow run "Lint" failed! GitBox
• [GH] (tvm/tvm-ir-prim-expressions): Workflow run "Lint" is working again! GitBox
• [PR] [REFACTOR][TIRx] Make BufferRegion a typed expression [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Make BufferRegion a typed expression [tvm] via GitHub
• [PR] [Fix][Relax][Frontend][Torch] Keep zero-sized dims when reshaping [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Keep zero-sized dims when reshaping [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Keep zero-sized dims when reshaping [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Keep zero-sized dims when reshaping [tvm] via GitHub
• [PR] [Fix][Relax][Frontend][Torch] Emit int64 indices for sort and argsort [tvm] via GitHub
• [GH] (tvm/ci/restore-ruff-format-c-host): Workflow run "Lint" is working again! GitBox
• [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
  • Re: [PR] [CI] Restore ruff-format on test_target_codegen_c_host.py [tvm] via GitHub
• [GH] (tvm/test/parametrize-round-ties-to-even): Workflow run "Lint" failed! GitBox
• [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
  • Re: [PR] [Test] Run round ties-to-even test on every backend that implements it [tvm] via GitHub
• [GH] (tvm/fix/c-round-ties-to-even): Workflow run "Lint" failed! GitBox
• [GH] (tvm/fix/c-round-ties-to-even): Workflow run "CI" failed! GitBox
• [GH] (tvm/codex/fix-subscript-proxy-regressions): Workflow run "CI" is working again! GitBox
• [GH] (tvm/tvm-ir-prim-expressions): Workflow run "CI" failed! GitBox
  • [GH] (tvm/tvm-ir-prim-expressions): Workflow run "CI" failed! GitBox
• [GH] (tvm/codex/fix-subscript-proxy-regressions): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/codex/fix-subscript-proxy-regressions): Workflow run "Lint" failed! GitBox
• [PR] [FIX][IR] Complete lazy subscript realization [tvm] via GitHub
  • Re: [PR] [FIX][IR] Complete lazy subscript realization [tvm] via GitHub
  • Re: [PR] [FIX][IR] Complete lazy subscript realization [tvm] via GitHub
• [PR] [FIX][RUST] Seal structural callback result traits and update return-value docs [tvm-ffi] via GitHub
  • Re: [PR] [FIX][RUST] Seal structural callback result traits and update return-value docs [tvm-ffi] via GitHub
• [GH] (tvm/codex/fix-mbarrier-local-predicate): Workflow run "Lint" is working again! GitBox
• [PR] [TIRx][CUDA] Preserve local mbarrier predicate and count [tvm] via GitHub
  • Re: [PR] [TIRx][CUDA] Preserve local mbarrier predicate and count [tvm] via GitHub
• [GH] (tvm/tvm-ir-prim-expressions): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tvm-ir-prim-expressions): Workflow run "Lint" failed! GitBox
• [PR] [REFACTOR][IR] Lift primitive expressions into core IR [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Lift primitive expressions into core IR [tvm] via GitHub
• [PR] [Rust] Decouple structural mutation dispatch from recursion control [tvm-ffi] via GitHub
  • Re: [PR] [Rust] Decouple structural mutation dispatch from recursion control [tvm-ffi] via GitHub
• [PR] [REFACTOR][IR] Consolidate expression operator overloading into the base layer [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Consolidate expression operator overloading into the base layer [tvm] via GitHub
• [PR] [FEAT][RUST] Accept any Into<Any> return value in structural map/mutate callbacks [tvm-ffi] via GitHub
  • Re: [PR] [FEAT][RUST] Accept any Into<Any> return value in structural map/mutate callbacks [tvm-ffi] via GitHub
• [GH] (tvm-ffi/main-dev/2026-08-31/refl_type_final_polymorphic): Workflow run "CI" is working again! GitBox
• [PR] [REFACTOR][IR] Introduce TensorLoad in core IR [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Introduce TensorLoad in core IR [tvm] via GitHub
• [GH] (tvm-ffi/main-dev/2026-08-31/refl_type_final_polymorphic): Workflow run "CI" failed! GitBox
• [PR] [CORE] Publish native layout traits from ObjectDef [tvm-ffi] via GitHub
  • Re: [PR] [CORE] Publish native layout traits from ObjectDef [tvm-ffi] via GitHub
• [PR] [REFACTOR][IR] Unify expression subscription realization [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Unify expression subscription realization [tvm] via GitHub
• [GH] (tvm-ffi/main-dev/2026-08-31/stubgen_layout_facts): Workflow run "CI" failed! GitBox
• [PR] [STUBGEN] Carry native layout facts through the language-neutral data. [tvm-ffi] via GitHub
  • Re: [PR] [STUBGEN] Carry native layout facts through the language-neutral data. [tvm-ffi] via GitHub
• [PR] [Fix][Relax][Frontend][Torch] Validate flatten dims in `from_fx` [tvm] via GitHub
• [PR] [REFACTOR][TIR] Split masked buffer access into special calls [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIR] Split masked buffer access into special calls [tvm] via GitHub
• [GH] (tvm/fix/relax-torch-split-int-split-size): Workflow run "CI" is working again! GitBox
• [PR] [Fix][Relax] Validate non-concat axes before fusing parallel matmul biases [tvm] via GitHub
  • Re: [PR] [Fix][Relax] Validate non-concat axes before fusing parallel matmul biases [tvm] via GitHub
  • Re: [PR] [Fix][Relax] Validate non-concat axes before fusing parallel matmul biases [tvm] via GitHub
  • Re: [PR] [Fix][Relax] Validate non-concat axes before fusing parallel matmul biases [tvm] via GitHub
• [I] [Bug][Relax] CPU default pipeline: fused PrimFunc keeps symbolic height/width unbound (MakePackedAPI ICHECK) [tvm] via GitHub
  • Re: [I] [Bug][Relax] CPU default pipeline: fused PrimFunc keeps symbolic height/width unbound (MakePackedAPI ICHECK) [tvm] via GitHub
  • Re: [I] [Bug][Relax] CPU default pipeline: fused PrimFunc keeps symbolic height/width unbound (MakePackedAPI ICHECK) [tvm] via GitHub
• [GH] (tvm/fix/relax-torch-round): Workflow run "Lint" failed! GitBox
• [PR] [CORE] Propagate static object type bounds in Any conversions [tvm-ffi] via GitHub
  • Re: [PR] [CORE] Propagate static object type bounds in Any conversions [tvm-ffi] via GitHub
  • Re: [PR] [CORE] Propagate static object type bounds in Any conversions [tvm-ffi] via GitHub
• [PR] [Fix][S-TIR] Preserve general reduction predicates [tvm] via GitHub
  • Re: [PR] [Fix][S-TIR] Preserve general reduction predicates [tvm] via GitHub
• [PR] [Relax][Frontend][Torch] Fix `torch.round(x, decimals)` via `from_exported_program` and negative-decimals rounding [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Fix `torch.round(x, decimals)` via `from_exported_program` and negative-decimals rounding [tvm] via GitHub
• [GH] (tvm/fix/relax-torch-split-int-split-size): Workflow run "Lint" is working again! GitBox
• [PR] [Relax][Frontend][PyTorch] Fix x.split(int) with a non-divisible split_size [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][PyTorch] Fix `x.split(int)` with a non-divisible `split_size` [tvm] via GitHub
• [GH] (tvm/fix/relax-torch-einsum-diagonal): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/fix/relax-torch-einsum-diagonal): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/fix/relax-torch-einsum-diagonal): Workflow run "Lint" failed! GitBox
• [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Support `aten.diagonal` from decomposed repeated-subscript einsum [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Frontend][Torch] Honor the `dtype` argument of `aten.mean` (`torch.Tensor.mean` / `torch.mean`) [tvm] via GitHub
• [PR] [Fix][Metal] Bound symbolic stack allocations [tvm] via GitHub
  • Re: [PR] [Fix][Metal] Bound symbolic stack allocations [tvm] via GitHub
  • Re: [PR] [Fix][Metal] Bound symbolic stack allocations [tvm] via GitHub
  • Re: [PR] [Fix][Metal] Bound symbolic stack allocations [tvm] via GitHub
  • Re: [PR] [Fix][Metal] Bound symbolic stack allocations [tvm] via GitHub
• [PR] [Fix][Relax][Metal] Constrain wide-head prefill tiling [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Metal] Constrain wide-head prefill tiling [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Metal] Constrain wide-head prefill tiling [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Metal] Constrain wide-head prefill tiling [tvm] via GitHub
  • Re: [PR] [Fix][Relax][Metal] Constrain wide-head prefill tiling [tvm] via GitHub
• Re: [I] [Bug][S-TIR][DLight] Reduction schedule rule crashes with ScheduleError on 4D tensor non-last-axis reduction [tvm] via GitHub
  • Re: [I] [Bug][S-TIR][DLight] Reduction schedule rule crashes with ScheduleError on 4D tensor non-last-axis reduction [tvm] via GitHub
• [GH] (tvm/fix/relax-onnx-pad-wrap-axes): Workflow run "CI" failed! GitBox
• [GH] (tvm/fix/relax-match-cast-storage-liveness): Workflow run "Lint" failed! GitBox
• [GH] (tvm/fix/relax-match-cast-storage-liveness): Workflow run "CI" failed! GitBox
• [PR] [TIRx] Restore BufferRegion as typed expression [tvm] via GitHub
  • Re: [PR] [TIRx] Restore BufferRegion as typed expression [tvm] via GitHub
• [PR] [CUDA][TIRx] Preserve device state during cleanup and skip invalid Top-K references [tvm] via GitHub

• Earlier messages