e4e63b0331
Revert TMEM layout change (caused error)
2026-05-23 20:07:02 +00:00
b450a82b4c
SMEM-P: try using PV A-operand layout directly for TMEM-P
2026-05-23 20:06:23 +00:00
74b3cc041d
SMEM-P: debug compute offset for known coordinate
2026-05-23 20:05:02 +00:00
5cfde8d3db
SMEM-P: add crd2idx debug attempt
2026-05-23 20:04:28 +00:00
98e0428840
SMEM-P: test pattern based on fragment indices (k,j)
2026-05-23 20:03:22 +00:00
8ca07722ae
SMEM-P: add shape debug prints
2026-05-23 20:02:32 +00:00
d008505330
SMEM-P: add debug prints for coordinates
2026-05-23 20:00:33 +00:00
638e8b862b
SMEM-P: implement CUTLASS LLM fixes - dynamic frg_tile, local coordinate conversion
2026-05-23 19:59:52 +00:00
f9f1381030
SMEM-P: try transposed mapping (swap m/n)
2026-05-23 19:53:42 +00:00
eb8084bb5f
SMEM-P: current state - working but mapping wrong (cos 0.02)
2026-05-23 19:53:25 +00:00
ebd086bc88
SMEM-P: debug with linear index pattern m*128+n
2026-05-23 19:52:46 +00:00
07973f88ab
SMEM-P: debug with test pattern (k+j)*0.01
2026-05-23 19:52:02 +00:00
90c62d0aac
SMEM-P: fix coordinate access - qk_coord is (m,n) not ((m,n),0,0)
2026-05-23 19:38:11 +00:00
b78b7f7a71
SMEM-P: fix scoping - define tTMEM_LOADcS_frg unconditionally
2026-05-23 19:37:34 +00:00
3fe7634643
SMEM-P: implement full 128-value write in softmax loop using coordinate mapping
2026-05-23 19:36:56 +00:00
9378df9e05
SMEM-P: fix BF16 value creation (use constant)
2026-05-23 19:33:29 +00:00
d4522e1245
SMEM-P: implement CUTLASS LLM coordinate mapping pattern (minimal test)
2026-05-23 19:32:11 +00:00
59b086451c
SMEM-P: fix thread_idx tuple access
2026-05-23 19:30:09 +00:00
e0252025fd
SMEM-P: add debug to understand thread partitioning
2026-05-23 19:29:27 +00:00
5ddaea6dad
SMEM-P: implement simple test pattern instead of coord lookup
2026-05-23 19:21:31 +00:00
97c04eab63
Start implementing manual SMEM-P addressing (helpers are a trap)
2026-05-23 19:20:40 +00:00
d4bdc70134
Implement manual SMEM-P copy instead of cute.copy (helpers are a trap)
2026-05-23 19:14:44 +00:00
2ec464824e
Try flattening sP and rP_bf16_qk with group_modes to fix rank mismatch
2026-05-23 19:13:59 +00:00
0e8f8ca2cf
Add debug prints for SMEM-P partition layouts
2026-05-23 19:13:07 +00:00
99566ef8a3
Fix SMEM-P: use BF16 copy atom and BF16 source with QK C-fragment layout
2026-05-23 19:12:13 +00:00
f77264f684
Fix SMEM-P copy: use tcgen05.copy.St32x32bOp with Float32 and copy from rP_words (Float32) not rP_bf16
2026-05-23 19:11:08 +00:00
65a52d20fd
Fix rP scope issue: use rP_bf16.iterator instead of rP.iterator
2026-05-23 09:36:22 +00:00
d5d259b214
Fix duplicate else: line in SMEM-P block
2026-05-23 09:35:47 +00:00
5aee06b991
SMEM-P: Use QK C-fragment layout instead of TMEM layout to fix rank mismatch
2026-05-23 09:35:24 +00:00
8ffd1154fb
Fix sP_2d definition for tSMEM_CPYsP
2026-05-23 09:34:50 +00:00
ba810e96c1
Remove debug print lines referencing deleted sP_2d
2026-05-23 09:34:09 +00:00
1b8bf4996f
Remove duplicate sP_2d line causing indentation error
2026-05-23 09:33:40 +00:00
7a06723cef
SMEM-P: Implement rank mismatch fix by reshaping source tensor
2026-05-23 09:33:24 +00:00
21be1d4355
Add more debug prints for sP shapes
2026-05-23 09:26:30 +00:00
69f458ac30
Add debug prints to SMEM-P path to understand rank mismatch
2026-05-23 09:25:48 +00:00
f6b43227e5
Fix SMEM-P copy rank mismatch (use rP_bf16 directly instead of group_modes)
2026-05-23 09:21:13 +00:00
83a7dd0679
D1.3: Implement SMEM-P path (write P to SMEM via tiled_smem_copy instead of zeroing sP)
2026-05-23 09:20:37 +00:00
d9349bd127
Stage D1: Multi-PV-tile support for hd>256 (tcgen05 MMA max N=256)
2026-05-23 09:04:01 +00:00
568840ffc9
Fix v_fmha layout to use pv_n_tile instead of head_dim for multi-PV-tile support
2026-05-23 09:02:01 +00:00
6934b9b4b6
D1.4: Add pv_n_tile and n_pv_tiles for multi-PV-tile support (tcgen05 MMA max N=256)
2026-05-23 09:00:18 +00:00
1816965db8
fix: import ceil_div in quantize.py (was NameError at runtime)
2026-05-23 08:40:24 +00:00
4eccbb05c1
shit carmine left dangling
2026-05-23 06:55:22 +00:00
cdfafe50de
fix: add SwiGLU clamping to fused kernel (paper §4.2.3, CG-1)
...
The fused SwiGLU kernel stored swiglu_limit but never applied it.
Paper §4.2.3: gate capped at swiglu_limit, linear clamped to [-limit, +limit].
Non-fused reference path already applies clamping correctly.
Fix: add fmin/fmax clamping in FP32 before BF16 conversion.
2026-05-23 06:32:54 +00:00
17b40eb3f8
STAGE_D.md: restructure with correctness gaps, TMEM budget, execution order
2026-05-23 06:31:37 +00:00
f30e7f5f43
D1.1: Fix make_fragment_A — use sP for SMEM source pv_mma
2026-05-23 06:04:44 +00:00
8b7b124295
D1.1: Fix PV A-operand construction — compile-time branch for TMEM vs SMEM
2026-05-23 06:03:27 +00:00
d7361fe9ae
D1.1: Add SMEM-P path behind use_smem_p flag (stub: zero sP)
2026-05-23 06:01:02 +00:00
e19a5d02d2
D1.0: Replace HEAD_DIM=64 with self.head_dim constructor parameter
2026-05-23 05:55:03 +00:00
a9adeb6e4a
Rename FmhaV3StageC → FmhaKernel — no dev stage artifacts in production API
2026-05-23 05:45:58 +00:00
0d3db00c44
Migrate Stage C kernel (proven cos 0.97) into module - exact copy, no modifications
2026-05-23 05:36:22 +00:00
