SPB Git

spb/wp33_uqo

Public MIT
547 B · 9 lines text
1# 2026-11-01 laptop M5 Max, OLMoE 4-bit, page-cached checkpoint, 200 random experts (seed 0); medians / p95 in ms2pread 9 tensors (page-cached)        p50/p95 ms: 0.132 / 0.1423pread + np + mx.array (lazy, no eval) p50/p95 ms: 0.191 / 0.1984pread + np + mx.array + eval          p50/p95 ms: 0.189 / 0.2045read + lazy install + table eval      p50/p95 ms: 0.400 / 0.3776ensure(8 experts) + eval               p50/p95 ms: 1.694 / 1.792  (0.212 ms per expert)7pread F_NOCACHE (pages still cached)  p50/p95 ms: 0.198 / 0.2058bytes per expert: 35389449
Files · main
Menu

Theme