inference/research
verifiedmedium · 0.70updated 2026-10-15

kyuz0 Strix Halo 8060S NVIDIA-Nemotron-3-Super-120B-A12B-UD-Q4_K_XL ROCm 7.2.3 tg128

14.85 tok/s

Fields

id
benchmarks/kyuz0-8060s-nvidia-nemotron-3-super-120b-a12b-rocm-723-tg128
Metric
decode_tok_s
Value
14.85
Unit
tok/s
Model
NVIDIA-Nemotron-3-Super-120B-A12B-UD-Q4_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 7.2.3, -fa 1 -ngl 99 --mmap 0, default ctx. tg128 batch-1 per-request.
Measured by
third_party
Reproducible
yes

Cross-references

Accelerators (1)

Engine (1)