biondizzle

0 Followers · 0 Following

• Joined on 2025-12-10

Repositories

25

Projects Packages Public Activity Starred Repositories

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:37:36 +00:00

07bf2adf51 D1.2: TMEM budget probe with real tensor major modes

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:36:41 +00:00

6e351c276d fix: OperandMajorMode.MN not .M

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:35:56 +00:00

cabe8489aa fix: typo + OperandMajorMode for TMEM budget probe

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:35:04 +00:00

61b9dbb2d6 fix: LayoutEnum import from cutlass.utils

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:34:31 +00:00

4c35fa49a9 fix import path for tcgen05

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:33:28 +00:00

a2d0dec7bb D1.2: TMEM budget probe script for hd=64,128,256,512

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:32:55 +00:00

578d186c20 fix: add SwiGLU clamping to fused kernel (paper §4.2.3, CG-1)

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:31:40 +00:00

11c7e2c663 STAGE_D.md: restructure with correctness gaps, TMEM budget, execution order

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:04:46 +00:00

3d69215c4e D1.1: Fix make_fragment_A — use sP for SMEM source pv_mma

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:03:29 +00:00

d0567524e1 D1.1: Fix PV A-operand construction — compile-time branch for TMEM vs SMEM

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 06:01:03 +00:00

a3344ddd50 D1.1: Add SMEM-P path behind use_smem_p flag (stub: zero sP)

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:55:04 +00:00

27041964e3 D1.0: Replace HEAD_DIM=64 with self.head_dim constructor parameter

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:52:05 +00:00

e98f5e4f9e Add STAGE_D.md: step-by-step runbook and todo list for D1-D5

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:46:00 +00:00

0520d55ca6 Rename FmhaV3StageC → FmhaKernel — no dev stage artifacts in production API

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:42:44 +00:00

af925abe3b Update README: reflect Stage C migration, built indexer/router/compressor, SMEM-P path, CuTeDSL scoping lesson

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:36:23 +00:00

c92976b3cd Migrate Stage C kernel (proven cos 0.97) into module - exact copy, no modifications

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:18:40 +00:00

e397386ba2 Fix TMEM-P offset calc: match Stage C with p_cols_fp32 from pv_mma_tiler[2]

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:17:46 +00:00

a284580422 Add missing TMEM fence after P store in TMEM-P path

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:16:21 +00:00

0cd0e8b35f Fix p_cols_fp32: use pv_mma_tiler[2] (K-dim) not [1] (N-dim)

biondizzle pushed to master at biondizzle/nvfp4-megamoe-kernel 2026-05-23 05:14:10 +00:00

721bac4958 Fix PV A-operand major mode: K for TMEM-P, a_major for SMEM-P