e637d3ae73
D1: Add K sub-tile loop for hd=512 (const_expr guarded, hd≤256 path unchanged)
2026-05-24 04:51:51 +00:00
24b9310682
D1: Debug TMA partition shapes at hd=512
2026-05-24 04:43:12 +00:00
9201a844dd
D1: K sub-tiling - qk_mma_tiler K-dim = k_tile=256, SMEM fits at hd=512
2026-05-24 04:41:12 +00:00
d234297712
D1: Remove debug prints, clean up
2026-05-24 04:06:26 +00:00
3b63405ad4
D1: const_expr for sP layout selection (CuTeDSL)
2026-05-24 04:05:17 +00:00
1c8b043702
D1: Python if for sP layout (trace-time, not MLIR)
2026-05-24 04:04:27 +00:00
3aa8e5185a
D1: Tiny 4-mode sP placeholder for TMEM-P path
2026-05-24 04:03:28 +00:00
03ad730a9b
D1: Conditional sP allocation (saves 64KB SMEM for TMEM-P at hd=256)
2026-05-24 04:02:02 +00:00
975829e5c7
D1: Fix sP dummy allocation
2026-05-24 04:00:19 +00:00
5fda73b53b
D1: Skip sP allocation when use_smem_p=False (saves 64KB at hd=256)
2026-05-24 03:59:27 +00:00
93590eb1ad
D1: Fix syntax (separate kv_stage line)
2026-05-24 03:58:12 +00:00
2958cad75d
D1: Reduce kv_stage to 1 at hd>128 to avoid SMEM overflow
2026-05-24 03:55:44 +00:00
d6f7d9009d
D1: FIX qk_mma_tiler K-dim = head_dim (was hardcoded to 64, broke hd>64)
2026-05-24 03:53:19 +00:00
b4bf6818c6
D1: Print qk_ik in _setup
2026-05-24 03:51:40 +00:00
0953708f2c
D1: Add more debug prints (QK/PV mode2 sizes)
2026-05-24 03:49:55 +00:00
d9bc430570
D1: Add sP shape debug print
2026-05-24 03:46:27 +00:00
0f50933f69
D1: Fix SMEM-P (coordinate store), LSE (FP32), add TMEM-P-only test
2026-05-24 03:27:14 +00:00
c995a2ca46
D1: Fix SMEM-P - coordinate-indexed store (replaces make_tiled_copy_C)
2026-05-24 03:24:44 +00:00
0de0f20799
feat: SMEM-P make_tiled_copy_C + zero-fill dest tensor
2026-05-24 03:23:53 +00:00
54915f6b56
feat: SMEM-P using make_tiled_copy_C(qk_mma) approach
2026-05-24 03:22:57 +00:00
ea4b6b10bc
fix: LSE type mismatch Float32→BFloat16
2026-05-24 03:20:26 +00:00
53bc54ed17
D1.5: Fix SMEM-P - use coordinate-indexed store (same proven pattern)
2026-05-24 03:19:32 +00:00
93e7fe97f7
D1.5: Always output un-normalized O + LSE (epilogue_tma_store only, no TMEM round-trip normalize)
2026-05-24 03:18:38 +00:00
b22ab84f1a
feat: SMEM-P using make_tiled_copy_A from PV MMA
2026-05-24 03:16:34 +00:00
7357b1a866
fix: fence_proxy not fence
2026-05-24 02:45:01 +00:00
6a909e16d0
D1.5: Replace broken make_cotiled_copy SMEM-P with coordinate-indexed store
2026-05-24 02:43:42 +00:00
29231fc75e
Merge branch 'master' of ssh://sweetapi.com:2222/biondizzle/nvfp4-megamoe-kernel
2026-05-24 02:41:39 +00:00
3803847900
D1.5: Use tCtO_fake layout for epilogue_tma_store (needs STAGE dim)
2026-05-24 02:41:32 +00:00
688ca6f737
feat: SMEM-P using make_cotiled_copy (one-row-per-thread)
2026-05-24 02:41:19 +00:00
34db0bf802
D1.5: Remove duplicate tTMrO definition (keep unconditional one)
2026-05-24 02:40:40 +00:00
af8303ba64
fix: reorder tTMrO definition after tTMEM_LOADcO
2026-05-24 02:36:08 +00:00
e41cf07f50
fix: tTMrO scoping + restore SMEM-P coordinate write
2026-05-24 02:34:53 +00:00
bfd598b937
D1.5: Move tTMrO def before softmax loop (CuTeDSL scoping)
2026-05-24 02:32:51 +00:00
1e8f784bb0
debug: zero-fill sP to check deadlock
2026-05-24 02:31:12 +00:00
ece137b2c2
D1.5: Use proven Stage C approach - normalize via TMEM round-trip + epilogue_tma_store
2026-05-24 02:30:25 +00:00
f2d95da4aa
test: SMEM-P coordinate verification test
2026-05-24 01:58:32 +00:00
699c646497
D1.5: Fix bSG_gC slicing - group trailing modes (CUTLASS pattern)
2026-05-24 01:41:52 +00:00
f2ab5790e8
D1.5: Dynamic slicing for tTR_gC (variable rest dims)
2026-05-24 01:40:44 +00:00
d28257185f
D1.5: Fix flat_divide slice coordinates (4 modes, no STAGE dim)
2026-05-24 01:39:21 +00:00
86971dda81
D1.5: Rewrite correction epilogue using CUTLASS pattern (transform_partitioned, flat_divide, paired atoms)
2026-05-24 01:37:53 +00:00
ec250eccd6
D1.5: Fix TMA store - group_modes on bSG_gC, use flat indexing
2026-05-24 01:36:01 +00:00
61c4e107da
D1.5: Fix TMA store - use flat_divide on tCgC instead of local_tile on mC
2026-05-24 01:35:10 +00:00
6de0d316ef
D1.5: Fix TMA store - use 3D tile for local_tile on 3D mC
2026-05-24 01:20:33 +00:00
f59fd07ba7
D1.5: Use group_modes on sC for 2D TMA view (preserves swizzle)
2026-05-24 00:52:57 +00:00
577066bb7f
D1.5: Use 2D sC_epi layout from c_smem_s for TMA partition
2026-05-24 00:51:18 +00:00
534327f31b
D1.5: Fix TMA store - use group_modes on sC and tCgC
2026-05-24 00:48:18 +00:00
8048aa4be6
D1.5: Simplify TMA store - use 2D sC_epi and gC_epi views
2026-05-24 00:46:52 +00:00
88a5f55120
D1.5: Fix TMA store - use existing gC partition
2026-05-24 00:43:35 +00:00
b9e91853fb
D1.5: Implement correction epilog with paired atoms (get_tmem_load_op + get_smem_store_op)
...
One-way: TMEM → registers (normalize) → SMEM → GMEM
Based on CUTLASS FMHA reference's correction_epilog pattern.
Eliminates TMEM round-trip error for O normalization.
O rescale (kt>0) still uses old atoms (separate fix).
2026-05-24 00:41:27 +00:00
e632490682
D1.5: Add TODO for correction epilog - keeping working TMEM round-trip for now
2026-05-24 00:37:36 +00:00
