f28223c9da
Replaces all Python PyTorch fallback attention with native ixformer kernels: Decode path: - ≤32K: paged_attention_v1 (5D KV layout, x=8) — verified on real BI-V100 - >32K: paged_attention_v2 (5D→4D permute) — verified 65K+ on real BI-V100 - Removes _forward_decode_pytorch Python fallback entirely Prefill path (profiling): - _run_sdpa_fallback now uses ixformer.flash_attn_varlen_func - head_dim=256 verified correct (diff<0.004) and 1.7x faster than PyTorch - Falls back to Q-tiling pure-math if ixformer unavailable Also includes: MoE kernel integration, GDN C++ kernels, diagnostic scripts, xllm upstream layer/kernel references, .dockerignore cleanup. All changes verified on real BI-V100 hardware (single card).
5 lines
176 B
Python
5 lines
176 B
Python
#!/usr/bin/env python3
|
|
"""Print ixformer vllm.py source code."""
|
|
with open("/usr/local/corex/lib64/python3/dist-packages/ixformer/functions/vllm.py") as f:
|
|
print(f.read())
|