llama.cpp ROCm Instinct MI60 llama-2-7b Q4_0 tg128 FA
91.46 tok/s (tg128, batch 1)
Fields
id
benchmarks/llamacpp-rocm-mi60-llama27b-q40-fa-tg128Metric
decode_tok_s
Value
91.46
Unit
tok/s (tg128, 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 504af20, reporter @Said-Akbar. Chip Instinct MI60 (32 GB HBM2 4096-bit). Per-request batch-1, NOT aggregate.
Measured by
third_party
Reproducible
yes