biondizzle 147d85a617 CRITICAL FIX: K GMEM slice (None,None,0,0) not (None,0,None,0) ...

K from QK MMA B-partition has GMEM iter at mode 1, NOT mode 2.
(None,0,None,0) hardcodes mode 1 to 0 → TMA always loads tile 0.
(None,None,0,0) keeps mode 1 free → correct multi-tile loading.

Proof: diag n=256 went from cos 0.711 → 0.999999 with this one change.

2026-05-22 17:59:57 +00:00

..

archive

Restructure: cutedsl/ -> dsv4/ with proper layering

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

e2e

Restructure: cutedsl/ -> dsv4/ with proper layering

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

integration

Restructure: cutedsl/ -> dsv4/ with proper layering

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

unit

CRITICAL FIX: K GMEM slice (None,None,0,0) not (None,0,None,0)

2026-05-22 17:59:57 +00:00

check_log.sh

Add check_log.sh convenience script

2026-05-22 17:07:23 +00:00

fmha_v3_stage_c_example1.py

restore tBgK to kh.count indexing (single-tile working), add TODO for multi-tile

2026-05-22 15:54:03 +00:00

fmha_v3_stage_c_example2.py

FMHA Stage-C multi-tile: combined K+V barrier, final_o_bar, acc_pipe producer

2026-05-22 16:23:36 +00:00

fmha_v3_stage_c_example3.py

Stage C: integrate example3 multi-tile fixes into unit test

2026-05-22 16:39:45 +00:00

fmha_v3_stage_c_example4.py

Add example4: manual kv_coord Int32 for GMEM tile indexing

2026-05-22 17:24:38 +00:00

fmha_v3_stage_c_example5.py

Add example5: use cutlass.range induction variable as TMA GMEM coord

2026-05-22 17:47:10 +00:00

native_stage_c_patch.py

more stuff

2026-05-22 08:57:38 +00:00

requirements.txt

test: add standalone layer 0 comparison test (no vLLM, no Docker)

2026-05-16 02:13:18 +00:00

run_test.sh

run_test.sh: SIGKILL all children of screen session on cleanup

2026-05-22 17:08:12 +00:00