llama.cpp Metal on apple-m3-ultra (60 GPU cores): Llama-2-7B Q4_0 pp512 (scoreboard #4167, 8e672ef)

COMMUNITY-SCOREBOARD (llama.cpp discussion #4167, ggerganov-maintained summary table). llama-bench -hf ggml-org/Llama-2-7B-GGUF -p 512 -n 128 -ngl 99 --delay 30, Metal backend, full GPU offload. Build commit 8e672ef (2023-11-21 baseline, before the later Metal flash-attention speedups the same thread documents on M2 Ultra). pp512 = prefill at batch 512; tg128 = batch-1 decode of 128 tokens. Mean over the default repeats. Per-request, NOT aggregate. This row is the LOWER GPU-core bin of the same chip; the higher-core bin is already stored and must not be averaged with this one.

Lower GPU-core bin from the #4167 summary table. The higher-core bin of this same chip is a separate already-stored record; do not treat these as duplicates. 8e672ef predates the Metal FA improvements, so a current build is faster — the thread's own M2 Ultra re-run shows that.