tvm · Issues· 245 open

[CI Problem] test_solution_consistency fails intentionally-nondeterministic check in solve_linear_equations

[Bug][Performance Regression][Relax][DLight] FP16 Exp producer inlining across broadcast causes 30-60% slowdown on RTX A6000

[Bug][TIRx][FP8] DTypeConversion mis-encodes float8_e5m2 zero/subnormal values when converting to/from float32

[Bug] Hexagon codegen is broken on main and the only test covering it never runs

[Bug] Unsigned conditions crash block access-region analysis with an internally generated negative constant

[Bug] Fixed-point multiply legalization rejects runtime operands, zero-exponent scaling, and integer flags

[Bug][Relax] BundleModelParams output (tuple param with symbolic-shape fields) segfaults VMShapeLower

[Bug] `relax.build` crashes with an `AttributeError` during the dlight GPU scheduling pass

[Bug][Relax] split with out-of-range indices: type layer clamps but topi computes a negative extent ? bad_alloc / uint error / silent oversized (OOB-read) output depending on config

[Relax][Frontend][Torch] `_squeeze` does not validate `dim`: out-of-bounds tuple dims are silently converted to `squeeze(None)` via `from_fx` (native PyTorch raises `IndexError`)

Bug][Relax] Unchecked indices in gather/scatter ops: silent OOB reads and memory-corrupting OOB writes (reachable through the ONNX importer)

[Bug] MetaSchedule tuning exits successfully with zero candidates measured on Windows

[RFC] Expanding ONNX operator coverage in the Relax frontend

[Bug][CUDA][dlight] GPU Fallback calls sch.fuse on a zero loop extent and aborts with "Divide by zero"

[Bug][CUDA] A tensor with a leading zero extent compiles, and the generated kernel faults with cudaErrorIllegalAddress at the first StreamSync