MindPattern
← The Wire
Entity trail

SDPA

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs
1
Findings
1
Edges
0
Sources
1

Corpus findings

  1. 2026-08-28 / github-pulse-researcherllama.cpp fixes a SYCL path that moved 4.56 GB per 2048-token prefill chunk by staging the KV cache instead of binding itBuild b10669 (2026-08-28) binds the f16 KV cache in place for the oneDNN SDPA path. The commit message does the arithmetic on Qwen3.8 27B Q4_K_S at a live KV length of 34,816: 71.3 MB per tensor, 142.6 MB staged per call for K and V, 285.2 MB of traffic per call, and 16 calls per ubatch equals 4.56 GB of memory traffic for a single 2048-token prefill chunk, scaling with live KV length (the first ubatch at seq=2048 moves only 0.27 GB). Build b10668 the same day adds Metal flash-attention vector tunings for M3 Max, M5 and M5 Pro, contributed by community members running the repo's own `ggml-metal-tuning` sweep.

Source trail

GitHub

Graph sources

entity graphfindings textkg entitiesnewsletter issues