Commit Graph

1049 Commits

Author SHA1 Message Date
0c6c3ea956 D1.4: Add hd=512 QK-only and standalone test for compilation debugging 2026-05-24 14:19:26 +00:00
59195c7d17 D1.4: Reduce pv_n_tile to 128 for hd=512 to fit SMEM budget (192KB) 2026-05-24 08:07:32 +00:00
f77feb6e0f D1: Unrolled k_sub path (hardcoded k_sub=0,1) to avoid cutlass.range IR explosion 2026-05-24 07:03:14 +00:00
555d953eff D1: Use cutlass.range for k_sub loops (CuTeDSL immutable handle) 2026-05-24 06:43:30 +00:00
978c789641 D1: Fix kvh scoping - define before loops, consume V via pipeline 2026-05-24 06:42:26 +00:00
c9eff9a60f D1: Fix kvb→kvh typo in PV GEMM 2026-05-24 06:41:25 +00:00
7acd5f10f3 D1: Remove qh.commit() - pipeline handles commit internally 2026-05-24 06:40:10 +00:00
cf6d7b1750 D1: TMA producer uses acquire_and_advance + commit (no wait_and_advance) 2026-05-24 06:38:15 +00:00
24ed5d311f D1: Use same pipeline API as working code (acquire_and_advance) for k_sub path 2026-05-24 06:36:19 +00:00
888af57f7b D1: Add PipelineState for k_sub TMA path 2026-05-24 05:02:17 +00:00
0db580f18a D1: Fix pipeline API for K sub-tile path (producer_acquire/commit) 2026-05-24 04:59:41 +00:00
9afef9ed7d D1: K sub-tile MMA path using pipeline barriers 2026-05-24 04:57:08 +00:00
8c82de7f61 D1: Fix TMA copies in k_sub path (no mbarrier, use cp_async wait) 2026-05-24 04:53:46 +00:00
8f0c9e0dad D1: Add K sub-tile loop for hd=512 (const_expr guarded, hd≤256 path unchanged) 2026-05-24 04:51:51 +00:00
9ed927f86a D1: Debug TMA partition shapes at hd=512 2026-05-24 04:43:12 +00:00
e58ec5ab41 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
f3d93bc810 Docs: Update STAGE_D.md, README.md status for D1 hd≤256 milestone 2026-05-24 04:32:43 +00:00
a325f4ca5f D1: Full test with TMEM-P at hd=64,128,256,512 2026-05-24 04:07:40 +00:00
1f38d54206 D1: Remove debug prints, clean up 2026-05-24 04:06:26 +00:00
bc97bf67ee D1: const_expr for sP layout selection (CuTeDSL) 2026-05-24 04:05:17 +00:00
95cafebad4 D1: Python if for sP layout (trace-time, not MLIR) 2026-05-24 04:04:27 +00:00
317d3209ac D1: Tiny 4-mode sP placeholder for TMEM-P path 2026-05-24 04:03:28 +00:00
8d565efd7a D1: Conditional sP allocation (saves 64KB SMEM for TMEM-P at hd=256) 2026-05-24 04:02:02 +00:00
65aaec5df0 D1: Fix sP dummy allocation 2026-05-24 04:00:19 +00:00
86064a2ffa D1: Skip sP allocation when use_smem_p=False (saves 64KB at hd=256) 2026-05-24 03:59:27 +00:00
9bab7fe659 D1: Fix syntax (separate kv_stage line) 2026-05-24 03:58:12 +00:00
15eef33728 D1: Reduce kv_stage to 1 at hd>128 to avoid SMEM overflow 2026-05-24 03:55:44 +00:00
e288fea062 D1: FIX qk_mma_tiler K-dim = head_dim (was hardcoded to 64, broke hd>64) 2026-05-24 03:53:19 +00:00
b408fa4718 D1: Print qk_ik in _setup 2026-05-24 03:51:40 +00:00
aa8b145035 D1: Add more debug prints (QK/PV mode2 sizes) 2026-05-24 03:49:55 +00:00
d3a3a6988a D1: SMEM-P test at hd=128 2026-05-24 03:48:37 +00:00
073119e9ca D1: Add sP shape debug print 2026-05-24 03:46:27 +00:00
44b72cc570 D1: Fix SMEM-P (coordinate store), LSE (FP32), add TMEM-P-only test 2026-05-24 03:27:14 +00:00
1effbd9484 D1: Fix SMEM-P - coordinate-indexed store (replaces make_tiled_copy_C) 2026-05-24 03:24:44 +00:00
0105037d99 feat: SMEM-P make_tiled_copy_C + zero-fill dest tensor 2026-05-24 03:23:53 +00:00
26cf1fbaed Merge branch 'master' of ssh://sweetapi.com:2222/biondizzle/nvfp4-megamoe-kernel 2026-05-24 03:23:22 +00:00
66aaadadbf D1: LSE diagnostic at various hd 2026-05-24 03:23:16 +00:00
143a9ae5c7 feat: SMEM-P using make_tiled_copy_C(qk_mma) approach 2026-05-24 03:22:57 +00:00
cadc11d5d2 D1: Add diagnostic test (TMEM-P vs SMEM-P at various hd) 2026-05-24 03:22:23 +00:00
47369c4d4d Merge branch 'master' of ssh://sweetapi.com:2222/biondizzle/nvfp4-megamoe-kernel 2026-05-24 03:21:06 +00:00
64e11caf97 D1 test: compare un-norm O + norm using ref row_sum + LSE verification 2026-05-24 03:21:01 +00:00
8eef860947 fix: LSE type mismatch Float32→BFloat16 2026-05-24 03:20:26 +00:00
aaf32ad8ea merge: keep our fmha.py (coordinate-indexed SMEM-P + epilogue_tma_store) 2026-05-24 03:19:52 +00:00
304429ee3e D1.5: Fix SMEM-P - use coordinate-indexed store (same proven pattern) 2026-05-24 03:19:32 +00:00
e6ed497bdd feat: SMEM-P with make_tiled_copy_tv + partition_S 2026-05-24 03:19:18 +00:00
b471579140 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
940141a480 feat: SMEM-P using make_tiled_copy_A from PV MMA 2026-05-24 03:16:34 +00:00
641af108c3 fix: fence_proxy not fence 2026-05-24 02:45:01 +00:00
8f4b494bbe merge: resolve conflict (keep our version) 2026-05-24 02:44:09 +00:00
6b47cea90f D1.5: Replace broken make_cotiled_copy SMEM-P with coordinate-indexed store 2026-05-24 02:43:42 +00:00