fa4-cute-runtime / CARD.md
liangsu9988's picture
Update CARD.md
ded1bf8 verified
|
Raw
History Blame
2.11 kB
metadata
library_name: kernels
license: bsd-3-clause
tags:
  - cuda
  - blackwell
  - flash-attention
  - cute-dsl
  - inference

flashrt/fa4-cute-runtime

Forward-only FlashAttention-4 CuTe DSL runtime used by FlashRT's GROOT N1.7 and PI0.5 Thor pipelines. The source is vendored under the private flashrt_fa4 namespace and does not shadow an installed flash_attn package.

This package adds the dedicated Blackwell D256 2CTA forward path required by PI0.5's 8-Q/1-KV-head PaliGemma encoder. The community FlashAttention-4 package already covers the D48/D72/D128 family; it does not currently expose D256 on SM100/SM110.

Functions

  • flash_attn_func
  • flash_attn_varlen_func
  • forward_static
from kernels import get_kernel

fa4 = get_kernel("flashrt/fa4-cute-runtime", version=1)
out = fa4.flash_attn_func(q, k, v, causal=False)

The vendored forward wrapper returns (out, lse); use result[0] when only the attention output is needed.

For a CUDA Graph hot path, preallocate the output and use the allocation-free entry point:

out = torch.empty_like(q)
fa4.forward_static(q, k, v, out, causal=False)

For a padded fixed-shape graph, pass the valid K/V length as a CUDA int32 tensor. PI0.5 uses this form for its encoder cache:

seqused_k = torch.tensor([valid_k], device="cuda", dtype=torch.int32)
fa4.forward_static(
    q, k_padded, v_padded, out,
    causal=False,
    pack_gqa=True,
    seqused_k=seqused_k,
)

Inputs follow FlashAttention's (batch, sequence, heads, head_dim) contract. Qualified model profiles include D72 MHA and D256 GQA (8 Q heads / 1 KV head), with both dense and seqused_k execution. This package targets SM100-family Blackwell forward inference and requires CUDA 13 plus nvidia-cutlass-dsl 4.4.x, 4.5.x, or 4.6.x. The wrapper selects Thor's accepted architecture alias according to the installed DSL version.

This is an execution backend rather than a universal SDPA replacement. Select it with model-shape profiling; the GROOT causal GQA profile benefits while some short vision profiles remain faster on PyTorch SDPA.