llama.cpp Metal on Apple M3 Max (MBP 16, 40-core GPU, Low Power Mode): llama 7B F16 pp512 (discussion #4167, c1d0e7a)

COMMUNITY llama-bench posted to llama.cpp discussion #4167. llama-bench -p 512 -n 128 -ngl 99, mean of the reporter's repeats, single stream, batch size 1. The number is per-request, NOT an aggregate across concurrent requests. Model is the canonical llama-2-7b GGUF (6.74B) the thread standardises on. Reply by @TeeAaTeeUu (comment 18201257, 2026-08-29) to the original @ymcui M3 Max post (comment 7656623): MacBook Pro 16-inch, M3 Max, 12+4 CPU, 40-core GPU. Backend MTL,BLAS, 12 threads. Build c1d0e7a00. The reporter calls it 'per updated instructions', so this is the post-2026-08-25 harness, not the 2023 8e672ef numbers already filed for this chip. LOW POWER MODE, printed as a separate table in the same comment. Decode falls 27.82 to 19.89 (F16), 48.94 to 32.11 (Q8_0), 81.57 to 49.95 (Q4_0).

Same machine and harness as the other rows in this comment. Recorded mean only; the reporter's +/- is in the methodology. format_id is null for F16 because data/quantization has no F16 GGUF record.