viable/strict/1788040196: Fix static launcher 16-bit float scalars (#184065)
- PyTorch: 235 events in the last 90 days
- PyTorch: 232th Release in the last 90 days
- Previous: earlier the same day · viable/strict/1788035047
What happened
Triton reports fp16 and bf16 scalar arguments as distinct 16-bit types, but the static launcher mapped both to the 32-bit float ABI code. CUDA and HIP therefore wrote four-byte values into slots that kernels read as two-byte scalars, while XPU bound every imported-kernel argument as a pointer-sized value. Assign distinct fp16 and bf16 ABI codes and pack them with the same semantics as Triton's generated launchers: d…
Summary assembled by rule from the sources below