llama.cpp ROCm RX 7900 XTX llama-2-7b Q4_0 tg128 noFA
167.11 tok/s (tg128, batch 1)
Fields
id
benchmarks/llamacpp-rocm-xtx-llama27b-q40-nofa-tg128Metric
decode_tok_s
Value
167.11
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 2f0c2db, reporter @Diablo-D3. Chip RX 7900 XTX (24 GB GDDR6 384-bit). Per-request batch-1, NOT aggregate.
Measured by
third_party
Reproducible
yes