nvfp4-megamoe-kernel

Files

biondizzle b2d0417a46 NVFP4-1.1: Mark fp4_quant.py as toolchain-blocked, clean up test files

CuTeDSL MLIR pipeline cannot lower any float→int op. All approaches fail:
arith.fptosi, llvm.inline_asm, nvvm.inline_ptx, llvm.bitcast.

Production path: dsv4/kernels/cuda/quantize_nvfp4.cu (raw CUDA, works).
For NVFP4-1.1 fusion, use post-epilogue CUDA kernel approach.

Removed dead test files (test_ptx_*, test_fp4_isolate*, test_minimal_cmp*,
test_dtype_store, test_threshold_round).

2026-05-28 04:59:01 +00:00

__init__.py

Restructure: cutedsl/ -> dsv4/ with proper layering

2026-05-21 17:30:44 +00:00

dense.py

Restructure: cutedsl/ -> dsv4/ with proper layering

2026-05-21 17:30:44 +00:00

fp4_quant.py

NVFP4-1.1: Mark fp4_quant.py as toolchain-blocked, clean up test files

2026-05-28 04:59:01 +00:00

fused_swiglu.py

NVFP4-1.1: FP4 quantization primitives for CuTeDSL kernels