biondizzle f4b7bed463 Fix final normalize: use working 2D register tensor pattern from working_softmax_maybe.py ...

The make_rmem_tensor(tTMEM_LOADcO.shape) creates a 1D tensor that doesn't
match the paired atom layout. The working pattern uses a 2D register tensor
with sub-tile composition (tTMrO_i_ = tTMrO[None, i] + composition).

2026-05-23 00:25:16 +00:00

..

__init__.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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

cudagraph_test.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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

layertest.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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

test_cutedsl.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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

test_fmha_v3_stage_c.py

Fix final normalize: use working 2D register tensor pattern from working_softmax_maybe.py

2026-05-23 00:25:16 +00:00

test_fmha_v3.py

FIX: (None,0,None,0) for ALL tma_partition outputs — verified shapes on B200

2026-05-22 23:35:55 +00:00

test_fp4_roundtrip.py

Restructure: cutedsl/ -> dsv4/ with proper layering

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