The pre-slice (None,0,None,0) hardcoded GMEM iteration to tile 0. Instead, keep the original tBgK and index with (None, kt, None, 0) inside the TMA loop, where kt selects the correct GMEM tile. This preserves 2D rank matching with the SMEM tensor.
DSV4 Inference Kernel
Architecture
DSV4 is not MLA. It uses CSA (Compressed Sparse Attention, m=4) and HCA (Heavily Compressed Attention, m′=128). KV latent is (T, 512) shared across all 128 heads. Sink weights merge sparse + SWA attention. vLLM misnames this as "MLA" — it is not. The architecture is fundamentally different.
DSV4 inference pipeline — component status
==========================================
Legend:
[✓] built and tested
[~] partial — reference or seam exists, native pending
[✗] to build
┌────────────────────────────────────┐
│ [✗] Embedding + mHC init │
│ token embed + n_hc=4 streams │
└────────────────┬───────────────────┘
│
▼
┌─ Transformer layer × L ──────────────────────────────────────────────┐
│ HCA on layers 0–1 of Pro, alternating CSA / HCA after │
│ │
│ ┌─ Attention sub-block ──────────────────────────────────────────┐ │
│ │ [✓] Residual mHC pre + post mix │ │
│ │ [~] Norms + RoPE RMSNorm + partial RoPE │ │
│ │ [✓] Q / KV projection NVFP4 linears + LoRA │ │
│ │ [~] Token compressor CSA m=4 / HCA m′=128 │ │
│ │ [✗] Indexer + top-k CSA only, FP4 QK │ │
│ │ [~] FMHA core QK → online softmax → PV │ │
│ │ + SWA branch + sink merge │ │
│ │ [✓] Output projection inv RoPE + wo_a grouped + wo_b │ │
│ └────────────────────────────────────────────────────────────────┘ │
│ │
│ ┌─ FFN sub-block ────────────────────────────────────────────────┐ │
│ │ [✓] Residual mHC pre + post mix │ │
│ │ [~] Pre-FFN norm RMSNorm │ │
│ │ [✗] Router sqrt(softplus) + topk + hash │ │
│ │ [✓] Routed MoE fused SwiGLU L1 + L2 │ │
│ │ [✓] Shared expert NVFP4 single-group GEMM │ │
│ └────────────────────────────────────────────────────────────────┘ │
└──────────────────────────────────┬───────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────────────┐
│ [✗] Final RMSNorm → [✗] LM head → [✗] MTP (depth=1) → [✗] Sampler │
└──────────────────────────────────────────────────────────────────────┘
┌─ Supporting infrastructure ──────────────────────────────────────────┐
│ [✗] KV cache management │
│ • state cache: SWA window + uncompressed tail per layer │
│ • classical paged cache: lcm(m, m′) = 128 tokens per block │
│ • heterogeneous layout per layer │
└──────────────────────────────────────────────────────────────────────┘
Summary
-------
Built [✓] : 6 — mHC ×2, Q/KV proj, output proj, routed MoE,
shared expert
Partial [~] : 4 — norms+RoPE, token compressor, FMHA core,
pre-FFN norm
To build [✗] : 8 — embedding+init, indexer+top-k, router,
final norm, LM head, MTP, sampler, KV cache
Status (May 22, 2026 — 09:40 UTC)
| Stage | Status | Description |
|---|---|---|
| A | ✅ COMPLETE | Q@K^T via tcgen05.mma → TMEM → GMEM |
| B | ✅ COMPLETE | QK → identity softmax → P@V pipeline (TMEM alias, KV-tile interleaving) |
| C | ✅ WORKING | Real online softmax: row_max (fmax), exp2 scaling, P store, row_sum, O normalization. Cosine 0.993-0.996 |
| C' | 🔨 NEXT | Cross-warp reduction, correction warps, 12-warp production pipeline, multi-tile KV |
| D | TODO | Full decode attention: paged KV cache, multi-query, causal mask |
| E | TODO | Production kernel: extract into dsv4/kernels/attention/, PyTorch custom op, vLLM bridge |
Package Structure
dsv4/
├── kernels/ Pure GPU code (CuTeDSL @cute.jit, .cu files)
│ ├── gemm/ NVFP4 MoE GEMM kernels (grouped, fused_swiglu, dense, scheduler)
│ ├── attention/ FMHA kernel (stub — extraction is Stage E)
│ ├── compressor/ CSA/HCA token-level compressor
│ ├── decode/ Decode-time attention (sparse, SWA — future)
│ └── cuda/ Raw .cu files (deinterleave_quantize, sparse_topk_metadata)
├── ops/ PyTorch ↔ kernel bridges
│ ├── quantize.py BF16 ↔ NVFP4 conversion, scale factors
│ ├── layouts.py Scale swizzle, gate/up interleave, K-major, offsets
│ ├── gemm_runner.py Warmup, compile, run grouped/fused GEMMs
│ ├── custom_ops.py torch.library.custom_op registrations
│ ├── decode_sparse.py native_sparse_decode dispatcher
│ ├── decode_swa.py native_swa_decode dispatcher
│ ├── rope.py Forward + inverse RoPE
│ └── topk.py Python wrapper for sparse_topk_metadata.cu
├── layers/ nn.Module-style components
│ ├── linear.py Nvfp4Linear
│ ├── grouped_linear.py Nvfp4GroupedLinear
│ ├── moe.py Nvfp4MoE
│ ├── shared_expert.py Nvfp4SharedExpert
│ ├── mhc.py mHCLayer
│ └── (stubs: attention, ffn, router, norm, embedding)
├── model/ Model assembly (stubs — Phase 1)
├── cache/ KV cache infra (stubs — Phase 3)
├── loader/ Checkpoint I/O (stubs — Phase 1)
└── reference/ Slow PyTorch oracles (never imported by production code)
├── attention.py RoPE, KV cache, causal attention, SWA
├── csa_attention.py CSA/HCA sparse attention
├── compressor.py Compressor PyTorch example
└── moe_pipeline.py MoE pipeline reference
Mental model: kernels/ → ops/ → layers/ → model/ (dependency flows left to right). reference/ and loader/ are sidecars.
Active Test Files
FMHA (Stages A/B/C) — in tests/unit/
| File | Stage | Status |
|---|---|---|
test_fmha_v3.py |
A+B | ✅ Full QK→identity softmax→PV, cosine 0.999999 |
test_fmha_v3_12w.py |
A+B | ✅ 12-warp QK→PV, cosine 0.999999 |
test_fmha_v3_stage_c_full.py |
C | ✅ Real online softmax + O normalization, cosine 0.993-0.996 |
test_fmha_v3_stage_c_min.py |
C | 🔨 Early 12-warp pipeline (broken pipeline state) |
test_pv64_with_softmax.py |
B | ✅ (128,64) PV, single AB pipeline |
test_128_128_vdiag.py |
A+B | ✅ (128,128) PV baseline |
test_qkonly.py |
A | ✅ QK with split Q/KV pipelines |
test_qk_softmax.py |
A+B | ✅ QK + identity softmax, no PV |
MoE / GEMM — in tests/unit/
| File | What |
|---|---|
test_cutedsl.py |
NVFP4 grouped GEMM kernel |
cudagraph_test.py |
Cudagraph capture + replay |
layertest.py |
Per-layer correctness |
test_custom_op.py |
torch.library custom ops |
test_compile_custom_op.py |
Compile + warmup |
test_fp4_roundtrip.py |
BF16 → NVFP4 → BF16 roundtrip |
test_interleave.py |
Gate/up weight interleaving |
test_interleave_gemm.py |
Interleaved GEMM correctness |
test_fused_step1.py |
Fused SwiGLU GEMM |
Archived Tests
tests/archive/ contains ~190 debug files from Stages A/B. Not maintained. Can be deleted.
Stage C: Online Softmax — WORKING
What We Have
Working real softmax in test_fmha_v3_stage_c_full.py: cosine 0.993–0.996 across 3 seeds.
Current Architecture (6-warp)
Warps 0-3: Softmax + Epilogue — load S, real softmax, P store, O normalize, epilogue Warp 4: MMA (QK→S, PV→O) Warp 5: TMA (Q/K/V load)
Target Architecture (12-warp, production)
Warps 0-3: Softmax — S→softmax→P, broadcast vec=[old_max, new_max] Warps 4-7: Correction — O rescale (TMEM), final normalization, SMEM write Warp 8: MMA — QK→S, PV→O with pipeline chaining Warp 9: TMA — Q/K/V load Warp 10: Epilogue — O SMEM→GMEM via TMA Warp 11: Empty — tmem dealloc mbar init
Pipeline chain: MMA → Softmax → Correction → Epilogue (plus MMA → Correction)
CuTeDSL Constraints (hard-won)
vectorize=Trueloops: ONLY load/store/print — no fmax, no cmpf, no inner loops, no carry.reduce(cute.ReductionOp.MAX): reduces ENTIRE C-fragment to scalar — global max, not per-row. Usecute.arch.fmaxelement-wise instead- Dynamic control flow: variables need initial values BEFORE the flow starts
cute.arch.fmax: impure for vectorizer — use plainrange()loop- Carry variables (row_max, row_sum): cannot use
vectorize=True
Remaining for C' (Production Stage C)
- Cross-warp reduction for row_max and row_sum
- Correction warps for multi-tile KV (online O rescale in TMEM)
- 12-warp layout with separate softmax/correction/epilogue warps
- Per-row O normalization
TMEM Layout
Col 0-127: S (QK acc, 128 FP32) | Col 32-95: P (Softmax, 64 FP32) | Col 128+: O (PV acc, 64 FP32)
Row_max/row_sum are per-thread FP32 scalars. Correction warps will use TMEM-backed vec buffer.
Stage E: Production Kernel Extraction
When ready, extract from test_fmha_v3.py → dsv4/kernels/attention/fmha.py:
- Clean
FmhaKernelclass with@cute.jit __call__, no hardcoded dimensions - Add real softmax (Stage C)
- Add paged KV cache (Stage D)
- Wrap as
torch.library.custom_opindsv4/ops/ - Integrate with vLLM
Key Lessons
- NEVER use
find_tmem_tensor_col_offset()as TMEM placement. It returns footprint size, not a safe offset. - FMHA never trusts DLPack tensor layouts. Reconstruct V as (hd, s_k) MN-major inside CuTe.
- TMEM allocation must be power of 2.
- Square hides bugs. (128,128) worked for every wrong approach. Always test non-square.
- St32x32bOp MUST use Float32, NOT BFloat16. BFloat16 causes illegal memory access.
- First PV ACCUMULATE=False. Otherwise adds uninitialized TMEM to output.
- FMHA P store uses QK C-fragment composition, NOT PV A-fragment. Two aliases, same TMEM.
- Register bridge: FP32 backing (store partition) + BF16 view (QK-load layout). Do not skip this.
Environment
- Server: root@45.76.247.107 (B200, 180 GiB HBM3e per GPU)
- venv:
source /root/dsv4-nvfp4-workspace/venv/bin/activate - PYTHONPATH:
/root/dsv4-nvfp4-workspace/kernel - Model:
/root/nvidia-meeting/DeepSeek-V4-Pro-NVFP4 - vLLM repo:
/root/dsv4-nvfp4-workspace/vllm(modified for Blackwell) - CUTLASS FMHA reference:
/root/cutlass/examples/python/CuTeDSL/cute/blackwell/kernel/attention/fmha/fmha.py