llama.cpp ROCm AI PRO R9700 llama-2-7b Q4_0 pp512 FA
4773.07 tok/s (pp512, batch 1)
Fields
id
benchmarks/llamacpp-rocm-r9700-llama27b-q40-fa-pp512Metric
prefill_tok_s
Value
4,773.07
Unit
tok/s (pp512, batch 1)
Model
llama-2-7b.Q4_0.gguf (TheBloke, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (llama.cpp discussion #15021, maintainer-curated ROCm/HIP backend). llama-bench llama-2-7b.Q4_0.gguf (TheBloke, 3.56 GiB, 6.74B params), pp512 prefill + tg128 batch-1 decode, ngl=99 full offload, ROCm. Build commit bd4ef13, reporter @gogich77. Chip AI PRO R9700 (32 GB GDDR6 256-bit). Per-request batch-1, NOT aggregate.
Measured by
third_party
Reproducible
yes