biondizzle c2664281c3 KV-1/KV-2: Fix quantize kernel — each thread handles 16-elem blocks independently ...

Previous version used __shfl_down_sync for group-level amax reduction,
but shuffles operate at warp level and crossed group boundaries.
Fix: each thread independently quantizes its assigned 16-element blocks
from shared memory. Simpler and correct.

2026-06-02 09:41:15 +00:00

..

attention

perf: skip MQA GQA expansion in FMHA (stride=0, no 128x K/V copy)

2026-06-02 03:54:03 +00:00

cache

fix: correct gather.py kernel_dir path

2026-05-30 21:12:09 +00:00

compressor

KV-1/KV-2: Fused compress+NVFP4 quantize kernels + dequant

2026-06-02 09:37:53 +00:00

cuda

KV-1/KV-2: Fix quantize kernel — each thread handles 16-elem blocks independently

2026-06-02 09:41:15 +00:00

gemm

fix: use cute.where() directly for clamp in fused SwiGLU

2026-06-02 08:16:41 +00:00

indexer

P0 COMPLETE: Eliminate ALL .item() CPU-GPU syncs from NVFP4 activation path

2026-06-01 21:05:03 +00:00

router

Switch router to Nvfp4Linear production GEMM (custom CuTeDSL kernel crashes MLIR)

2026-06-01 11:17:54 +00:00

__init__.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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