discuss-archive  

Messages by Thread

• [PR] [IR] Make function attribute updates generic via reflected shallow copy [tvm] via GitHub
  • Re: [PR] [IR] Make function attribute updates generic via reflected shallow copy [tvm] via GitHub
• [GH] (tvm/fix_cudnn_cleanworkspace): Workflow run "Lint" failed! GitBox
• [PR] [DOC] Clarify UnchangedOr replacement and identity semantics [tvm-ffi] via GitHub
  • Re: [PR] [DOC] Clarify UnchangedOr replacement and identity semantics [tvm-ffi] via GitHub
• [PR] [REFACTOR][TIRX] Use single-inheritance typed StmtExprMutator [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRX] Use single-inheritance typed StmtExprMutator [tvm] via GitHub
• [PR] [REFACTOR][IR] Extract shared primitive operations and registry names [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Share primitive helpers and Python expressions [tvm] via GitHub
• [PR] [REFACTOR][RELAX] Own global device metadata in Relax [tvm] via GitHub
  • Re: [PR] [REFACTOR][RELAX] Own global device metadata in Relax [tvm] via GitHub
• [GH] (tvm/ir-shared-side-effect): Workflow run "CI" is working again! GitBox
• [PR] [CUDA] Fix workspace pool destruction order [tvm] via GitHub
  • Re: [PR] [CUDA] Fix workspace pool destruction order [tvm] via GitHub
  • Re: [PR] [CUDA] Fix workspace pool destruction order [tvm] via GitHub
  • Re: [PR] [Fix][cuDNN] Avoid thread-local workspace pool for ConvEntry [tvm] via GitHub
• [I] [Bug] [Relax][cuDNN] [tvm] via GitHub
  • Re: [I] [Bug] [Relax][cuDNN] CoreDump When cuDNN ConvEntry CleanWorkspace [tvm] via GitHub
• [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tests/trim-pytorch-frontend-tests): Workflow run "Lint" failed! GitBox
• [PR] [Refactor][Frontend] Consolidate PyTorch tests and fix importer semantics [tvm] via GitHub
  • Re: [PR] [Frontend][PyTorch] Simplify tests and support exported assertions [tvm] via GitHub
• [GH] (tvm/arith/remove-vscale-specialization): Workflow run "CI" is working again! GitBox
• [GH] (tvm/ir-shared-side-effect): Workflow run "Lint" is working again! GitBox
  • [GH] (tvm/ir-shared-side-effect): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "CI" failed! GitBox
  • [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "CI" failed! GitBox
  • [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "CI" failed! GitBox
• [GH] (tvm/ir-shared-side-effect): Workflow run "Lint" failed! GitBox
• [PR] [IR][TE] Share expression effects and register tensor loads [tvm] via GitHub
  • Re: [PR] [IR][TE] Share expression effects and register tensor loads [tvm] via GitHub
• [GH] (tvm/arith/remove-vscale-specialization): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/task467-tirx-mechanical-exprfunctor): Workflow run "Lint" failed! GitBox
• [PR] [REFACTOR][TIRX] Share generic expression functor dispatch [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRX] Share generic expression functor dispatch [tvm] via GitHub
• [GH] (tvm/arith/remove-vscale-specialization): Workflow run "Lint" failed! GitBox
• [PR] [Arith] Remove vscale-specific simplification and analysis [tvm] via GitHub
  • Re: [PR] [Arith] Remove vscale-specific simplification and analysis [tvm] via GitHub
• [GH] (tvm/task469-arith-var-evalset): Workflow run "CI" is working again! GitBox
• [GH] (tvm/prim-expr-deep-equal): Workflow run "Lint" is working again! GitBox
  • [GH] (tvm/prim-expr-deep-equal): Workflow run "Lint" is working again! GitBox
• [PR] [Refactor][IR] Move ExprDeepEqual into shared primitive expressions [tvm] via GitHub
  • Re: [PR] [Refactor][IR] Move ExprDeepEqual into shared primitive expressions [tvm] via GitHub
• [GH] (tvm/s-tir-conditional-bounds): Workflow run "Lint" failed! GitBox
• [GH] (tvm/experiment/constr-visitor-main): Workflow run "Lint" failed! GitBox
• [GH] (tvm/feat/declared-atomic-ref): Workflow run "CI" failed! GitBox
• [GH] (tvm/feat/declared-atomic-ref): Workflow run "Lint" failed! GitBox
• [PR] [Refactor][Arith] Move conditional bounds into S-TIR [tvm] via GitHub
  • Re: [PR] [Refactor][Arith] Move conditional bounds into S-TIR [tvm] via GitHub
• [PR] [REFACTOR][Arith] Evaluate iterator domains through Var maps [tvm] via GitHub
  • Re: [PR] [REFACTOR][Arith] Evaluate iterator domains through Var maps [tvm] via GitHub
  • Re: [PR] [REFACTOR][Arith] Evaluate iterator domains through Var maps [tvm] via GitHub
• [PR] [TIRx][CUDA] Add the declared synchronization-word wait [tvm] via GitHub
  • Re: [PR] [TIRx][CUDA] Add the declared synchronization-word wait [tvm] via GitHub
• [I] [Bug] Hexagon codegen is broken on main and the only test covering it never runs [tvm] via GitHub
• [PR] [REFACTOR][S-TIR] Preserve schedule error payloads across FFI bridges [tvm] via GitHub
  • Re: [PR] [REFACTOR][S-TIR] Preserve schedule error payloads across FFI bridges [tvm] via GitHub
  • Re: [PR] [REFACTOR][S-TIR] Preserve schedule error payloads across FFI bridges [tvm] via GitHub
• [GH] (tvm/task466-mechanical-visitors): Workflow run "Lint" is working again! GitBox
• [GH] (tvm/task466-mechanical-visitors): Workflow run "Lint" failed! GitBox
• [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
  • Re: [PR] [REFACTOR][TIRx] Migrate visitors to shared expression traversal [tvm] via GitHub
• [PR] [REFACTOR][Arith] Inline bound-check Ramp expression [tvm] via GitHub
  • Re: [PR] [REFACTOR][Arith] Inline bound-check Ramp expression [tvm] via GitHub
• [PR] [REFACTOR][IR] Share checked PrimVar view across dialects [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Share checked PrimVar view across dialects [tvm] via GitHub
• [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
  • Re: [PR] [FIX][TIR] Conservatively analyze unsigned branch conditions [tvm] via GitHub
• [PR] chore(release): Version bump after release v0.1.14 [tvm-ffi] via GitHub
  • Re: [PR] chore(release): Version bump after release v0.1.14 [tvm-ffi] via GitHub
• [GH] (tvm/refactor-ir-exception-first-traversal): Workflow run "CI" is working again! GitBox
• [GH] (tvm/cleanup/exception-first-traversal): Workflow run "Lint" is working again! GitBox
• [I] [RESULT][VOTE] Release Apache TVM FFI v0.1.14-rc5 [tvm-ffi] via GitHub
  • Re: [I] [RESULT][VOTE] Release Apache TVM FFI v0.1.14-rc5 [tvm-ffi] via GitHub
• [PR] [CLEANUP][IR] Simplify exception-first traversal [tvm] via GitHub
  • Re: [PR] [CLEANUP][IR] Simplify exception-first traversal [tvm] via GitHub
• [GH] (tvm/tflite-quantized-avgpool): Workflow run "Lint" failed! GitBox
  • [GH] (tvm/tflite-quantized-avgpool): Workflow run "Lint" failed! GitBox
• [GH] (tvm/tflite-quantized-avgpool): Workflow run "CI" failed! GitBox
  • [GH] (tvm/tflite-quantized-avgpool): Workflow run "CI" failed! GitBox
• [PR] [FIX][Relax] Guard unknown tensor shapes in VMShapeLower [tvm] via GitHub
• [I] [Bug] Unsigned conditions crash block access-region analysis with an internally generated negative constant [tvm] via GitHub
• [PR] [FIX][Relax] Validate DeviceMesh cardinality without overflow [tvm] via GitHub
• [I] [Bug][Performance Regression][Relax][DLight] FP16 Exp producer inlining across broadcast causes 30-60% slowdown on RTX A6000 [tvm] via GitHub
  • Re: [I] [Bug][Performance Regression][Relax][DLight] FP16 Exp producer inlining across broadcast causes 30-60% slowdown on RTX A6000 [tvm] via GitHub
  • Re: [I] [Bug][Performance Regression][Relax][DLight] FP16 Exp producer inlining across broadcast causes 30-60% slowdown on RTX A6000 [tvm] via GitHub
• [PR] [FIX][TIR] Fix fixed-point multiply legalization [tvm] via GitHub
• [I] [Bug] Fixed-point multiply legalization rejects runtime operands, zero-exponent scaling, and integer flags [tvm] via GitHub
• [GH] (tvm-ffi/feat/rust-structural-visit-layers): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/feat/rust-structural-visit-layers): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/rust-structural-visit-layers): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/rust-structural-visit-layers): Workflow run "CI" failed! GitBox
• [PR] [FEAT][Rust] Compose structural walk and visit layers [tvm-ffi] via GitHub
  • Re: [PR] [FEAT][Rust] Compose structural walk and visit layers [tvm-ffi] via GitHub
  • Re: [PR] [FEAT][Rust] Add composable policies for structural visit and walk [tvm-ffi] via GitHub
• [PR] fix: preserve lvalues in UnchangedOr value access [tvm-ffi] via GitHub
  • Re: [PR] fix: preserve lvalues in UnchangedOr value access [tvm-ffi] via GitHub
• [PR] [REFACTOR] Default mutators to exceptions and migrate arithmetic passes [tvm] via GitHub
  • Re: [PR] [REFACTOR] Default mutators to exceptions and migrate arithmetic passes [tvm] via GitHub
• [PR] [Fix][Rust] Align structural mutation remapping policy [tvm-ffi] via GitHub
  • Re: [PR] [Fix][Rust] Align structural mutation remapping policy [tvm-ffi] via GitHub
  • Re: [PR] [Fix][Rust] Align structural mutation remapping policy [tvm-ffi] via GitHub
• [GH] (tvm-ffi/ffi-optional-visit-return): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/ffi-unchangedor-casts): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/ffi-unchangedor-casts): Workflow run "CI" failed! GitBox
• [PR] [FEAT] Add consuming casts and raw adoption to UnchangedOr [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] Add consuming casts and raw adoption to UnchangedOr [tvm-ffi] via GitHub
• [PR] [FEAT] Support optional structural visit early returns [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] Support optional structural visit early returns [tvm-ffi] via GitHub
• [PR] feat: add Expected error-return checks [tvm-ffi] via GitHub
  • Re: [PR] feat: add Expected error-return checks [tvm-ffi] via GitHub
• [PR] [REFACTOR][IR] Unify structural mutation modes and native traversal entrypoints [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Unify structural mutation modes and native traversal entrypoints [tvm] via GitHub
• [PR] [REFACTOR] Remove the constant-tree structural kind [tvm-ffi] via GitHub
  • Re: [PR] [REFACTOR] Remove redundant BigInt structural-kind metadata [tvm-ffi] via GitHub
• [PR] [FEAT] Add strict narrowing for Expected values [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] Add strict narrowing for Expected values [tvm-ffi] via GitHub
• [GH] (tvm-ffi/feat/unchangedor-const-fallback): Workflow run "CI" is working again! GitBox
• [GH] (tvm-ffi/feat/unchangedor-const-fallback): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/unchangedor-const-fallback): Workflow run "CI" failed! GitBox
• [PR] [API] Consolidate structural mutation entrypoints and borrowed fallback [tvm-ffi] via GitHub
  • Re: [PR] [API] Consolidate structural mutation entrypoints and borrowed fallback [tvm-ffi] via GitHub
• [GH] (tvm/experiment/constr-visitor-main): Workflow run "CI" is working again! GitBox
• [PR] [Fix][Metal][WebGPU] Preserve variable bindings when resolving allocation bounds [tvm] via GitHub
  • Re: [PR] [Fix][Metal][WebGPU] Preserve variable bindings when resolving allocation bounds [tvm] via GitHub
  • Re: [PR] [Fix][Metal][WebGPU] Preserve variable bindings when resolving allocation bounds [tvm] via GitHub
• [PR] [Fix][Relax] Allow callers to specify scan index-width budget [tvm] via GitHub
  • Re: [PR] [Fix][Relax] Allow callers to specify scan index-width budget [tvm] via GitHub
• [I] [Bug][Relax] BundleModelParams output (tuple param with symbolic-shape fields) segfaults VMShapeLower [tvm] via GitHub
• Re: [PR] [Relax][ONNX] Handle empty Slice with negative step [tvm] via GitHub
• [PR] [Arith] Add constraint visitor snapshots [tvm] via GitHub
• [GH] (tvm-ffi/feat/rust-structural-unchanged): Workflow run "CI" is working again! GitBox
• [PR] [FEAT][Rust] Preserve unchanged structural mutation results [tvm-ffi] via GitHub
  • Re: [PR] [FEAT][Rust] Preserve unchanged structural mutation results [tvm-ffi] via GitHub
• [PR] [Fix][Rust] Map children of inline pre-order replacements [tvm-ffi] via GitHub
  • Re: [PR] [Fix][Rust] Map children of inline pre-order replacements [tvm-ffi] via GitHub
• [PR] [FIX][Relax] Preserve symbolic shape dependencies in FuseOps [tvm] via GitHub
• [GH] (tvm-ffi/feat/bigint): Workflow run "CI" is working again! GitBox
• [PR] [FIX] Forward implicitly convertible values into UnchangedOr [tvm-ffi] via GitHub
  • Re: [PR] [FIX] Forward implicitly convertible values into UnchangedOr [tvm-ffi] via GitHub
• [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] BigInt support for arbitrary-precision integer value [tvm-ffi] via GitHub
• [GH] (tvm-ffi/feat/bigint): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/bigint): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/bigint): Workflow run "CI" failed! GitBox
  • [GH] (tvm-ffi/feat/bigint): Workflow run "CI" failed! GitBox
• [PR] Bump js-yaml from 3.15.0 to 3.15.2 in /web [tvm] via GitHub
  • Re: [PR] Bump js-yaml from 3.15.0 to 3.15.2 in /web [tvm] via GitHub
  • Re: [PR] Bump js-yaml from 3.15.0 to 3.15.2 in /web [tvm] via GitHub
• [PR] [BugFix][Relax] Preserve out_dtype in matmul rewrite passes [tvm] via GitHub
• [PR] [DLight][CUDA] Fall back from unsupported GEMV spatial tiles [tvm] via GitHub
• [PR] [FEAT] Add optional in-place permission to default structural descent [tvm-ffi] via GitHub
  • Re: [PR] [FEAT] Add optional in-place permission to default structural descent [tvm-ffi] via GitHub
• [PR] [FIX][TIRx][CUDA] Support SM100 weight-stationary B collectors [tvm] via GitHub
  • Re: [PR] [FIX][TIRx][CUDA] Support SM100 weight-stationary B collectors [tvm] via GitHub
• [I] [TIRx][CUDA] tcgen05.mma.ws has no collector_b modifier on sm_100a [tvm] via GitHub
  • Re: [I] [TIRx][CUDA] tcgen05.mma.ws has no collector_b modifier on sm_100a [tvm] via GitHub
• [PR] [REFACTOR][IR] Object and expression functors [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Object and expression functors [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] New Functor/Visitor/Mutator Infra at Base Layer [tvm] via GitHub
• [PR] [Test] Restore shadowed duplicate test functions in relax/runtime tests [tvm] via GitHub
• [PR] [REFACTOR][IR] Colocate type hooks and registrations [tvm] via GitHub
  • Re: [PR] [REFACTOR][IR] Colocate type hooks and registrations [tvm] via GitHub
• [PR] [PERF] Reduce structural traversal result overhead [tvm-ffi] via GitHub
  • Re: [PR] [PERF] Reduce structural traversal result overhead [tvm-ffi] via GitHub
  • Re: [PR] [PERF] Reduce structural traversal result overhead [tvm-ffi] via GitHub
• [PR] [ONNX] Preserve integer Div constant-fold precision [tvm] via GitHub
  • Re: [PR] [ONNX] Preserve integer Div constant-fold precision [tvm] via GitHub
• [PR] [ONNX] Fix initializer prefix stripping [tvm] via GitHub
  • Re: [PR] [ONNX] Fix initializer prefix stripping [tvm] via GitHub
• [GH] (tvm-ffi/prerequisite/structural-map-remap-state): Workflow run "CI" is working again! GitBox

• Earlier messages
• Later messages