llama.cpp ROCm Radeon 8060S llama-2-7b Q4_0 tg128 FA
48.97 tok/s (tg128)
Fields
id
benchmarks/llamacpp-scoreboard-radeon-8060s-ryzen-ai-max-fa-tg128Metric
decode_tok_s
Value
48.97
Unit
tok/s (tg128)
Model
llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params)
Methodology
COMMUNITY-SCOREBOARD (llama.cpp discussion #15021, maintainer-curated ROCm backend). llama-bench llama-2-7b.Q4_0.gguf (ggml-org, 3.56 GiB, 6.74B params), pp512 and tg128, n_gpu_layers 99, batch 1, full GPU offload, ROCm. Per-request batch-1, not an aggregate across concurrent requests. Flash attention on (-fa 1). Build 1d0125b, reporter @hspak. Printed mean 48.97. Re-read of the discussion page on 2026-10-17.
Measured by
third_party
Reproducible
yes
notes
Radeon 8060S, System Shared / DDR5, build 1d0125b, reporter @hspak. token generation cell.
Notes
Radeon 8060S, System Shared / DDR5, build 1d0125b, reporter @hspak. token generation cell.