coffee/DeepInfer
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepInfer.git synced 2026-10-28 14:57:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

Neuron attention: decode fp4 KV records in float arithmetic (Vector engine), not integer bit ops (GpSimd)

Browse Source 
The compressed-KV scopes are fp4 records; their decode was nibble masks and
shifts plus the integer e2m1 / e4m3 decoders on GpSimd (three such chains were
the largest cost of the ratio-2 attention graph, device profile bench72). Now:
nibbles from floor(b / 16), e2m1 from a piecewise-linear map of the magnitude
code, scales through the hardware fp8 conversion. Bit-identical to the integer
decode for every byte (host check). bench73 (vllm-sonnet 512, TP=64, nextn=5):
TPOT 243.5 -> 228.6 ms, TTFT 4.84 -> 4.35 s, step 699 -> 666 ms, tokens per
step identical.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
whn09 2026-10-26 08:18:39 +08:00
parent 94fa5c92c0
commit e7cb740b97
1 changed files with 19 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
deepinfer/neuron_ops/nki_attention.py
View File

@ -119,8 +119,27 @@ def decode_v41_fp8(records: torch.Tensor) -> torch.Tensor:
return _e4m3(data, sf).reshape(*lead, D_QK).to(torch.bfloat16)
def _decode_v41_fp4_fast(records: torch.Tensor) -> torch.Tensor:
"""``decode_v41_fp4`` in float arithmetic (Vector engine) instead of integer bit ops (GpSimd): the nibbles come
from floor(b / 16), e2m1 from a piecewise-linear map of the 3-bit magnitude code, the e4m3 scales from the
hardware fp8 conversion (see FAST_E4M3). Same values code by code."""
lead = records.shape[:-1]
b = records[..., : D_QK // 2].to(torch.float32)
hi = torch.floor(b * 0.0625)
lo = b - 16.0 * hi
codes = torch.stack((lo, hi), dim=-1).flatten(-2) # even element in the low nibble
neg = codes >= 8.0
c = torch.where(neg, codes - 8.0, codes)
mag = torch.where(c < 4.0, 0.5 * c, torch.where(c < 6.0, c - 2.0, 2.0 * c - 8.0))
values = torch.where(neg, -mag, mag).reshape(*lead, D_QK // 16, 16)
scales = records[..., D_QK // 2 : D_QK // 2 + D_QK // 16].contiguous().view(torch.float8_e4m3fn).to(torch.float32)
return (values * scales.unsqueeze(-1)).reshape(*lead, D_QK).to(torch.bfloat16)
def decode_v41_fp4(records: torch.Tensor) -> torch.Tensor:
"""``[..., 288]`` uint8 V4.1 fp4 records -> ``[..., 512]`` bf16."""
if FAST_E4M3:
return _decode_v41_fp4_fast(records)
v = records.to(torch.int32)
lead = v.shape[:-1]
values = _unpack_e2m1(v[..., : D_QK // 2]).reshape(*lead, D_QK // 16, 16)