Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-30 · source-backed
Build b10687 defaults the xmem F16xF32 GEMM on for Adreno X2E after measuring kernel_mul_mm_f16_f32_l4_lm running gpt-oss-20b attention projections at about a quarter the speed of the tuned dense q4_0 GEMM. Worth 25% prefill. The route already existed but was opt-in, so nobody hit it. Gated to X2E because the Adreno 840 measures neutral. The same commit bypasses the tiled f32 GEMM on the Adreno A7X, whose compiler allocates 488 B/WI of private memory against 304 on the next generation and spills in the K-loop. (GitHub)
Each link below shares sources, entities, or timing with this story.
Shared entities / Same source domain / Shared topic / Earlier coverage
Both cover GitHub, GPU; reported by the same outlet (github.com); overlapping topics (already, attention, dense).
Both cover Build, GitHub; reported by the same outlet (github.com); overlapping topics (commit, prefill).
Shared entities / Same source domain / Earlier coverage / Tension
Both cover GitHub, GPU; reported by the same outlet (github.com); earlier GitHub coverage from 2026-08-28.
Both cover GitHub, GPU; reported by the same outlet (github.com); earlier GitHub coverage from 2026-05-14.
Both cover GitHub, GPU; reported by the same outlet (github.com); earlier GitHub coverage from 2026-04-30.
Shared entities / Shared topic / Earlier coverage / Tension
Both cover GitHub, GPU; overlapping topics (already, dense); earlier GitHub coverage from 2026-04-23.
Shared entities / Same source domain / Earlier coverage / Tension
Both cover GitHub, GPU; reported by the same outlet (github.com); earlier GitHub coverage from 2026-02-22.
Shared entity: GitHub / Same source domain / Shared topic / Earlier coverage / Tension
Both cover GitHub; reported by the same outlet (github.com); overlapping topics (against, commit).