kyuz0 Strix Halo 8060S Qwen3.5-122B-A10B-UD-Q5_K_XL ROCm 6.4.4 tg128
19.73 tok/s
Fields
id
benchmarks/kyuz0-8060s-qwen35-122b-a10b-ud-q5-k-xl-rocm644-tg128Metric
decode_tok_s
Value
19.73
Unit
tok/s
Model
Qwen3.5-122B-A10B-UD-Q5_K_XL
Methodology
kyuz0 amd-strix-halo-toolboxes llama-bench sweep (results.json 2026-05-16, Fedora 43 kernel 6.19.12, llama.cpp b9187/9193). ROCm 6.4.4, -fa 1 -ngl 99 --mmap 0, default ctx. tg128 batch-1 per-request.
Measured by
third_party
Reproducible
yes