biondizzle ec8f292112 Fix: use self.mma_tiler_mnk (full K=64) for SMEM layout computation ...

SFA/SFB SMEM layouts need the full K dimension to compute the correct
number of K-tiles. self.mma_tiler has K=1 (placeholder for cute.slice_)
which gives 0 K-tiles and zero-dimension SMEM shapes.

2026-06-01 11:03:08 +00:00

..

__init__.py

Wire NVFP4 fused router kernel into e2e single-shot pipeline

2026-06-01 09:47:48 +00:00

_activation_topk.py

Rewrite NVFP4 fused router kernel: MoE-style epilogue replaces broken SMEM merge

2026-06-01 09:59:34 +00:00

dense_router_decode_kernel.py

Fix BF16 router mma_tiler: use cutlass.Int32 for CuTe DSL compatibility

2026-06-01 07:29:30 +00:00

dense_router_decode.py

Rewrite NVFP4 fused router kernel: MoE-style epilogue replaces broken SMEM merge

2026-06-01 09:59:34 +00:00

dense_router_prefill.py

Router: clean up dense_router_decode.py — realistic architecture, no fake code

2026-05-21 21:58:31 +00:00

nvfp4_fused_router_kernel.py

Fix: use self.mma_tiler_mnk (full K=64) for SMEM layout computation

2026-06-01 11:03:08 +00:00