lhl 8060S gpt-oss-120b-F16 Vulkan tg1 decode
33.72 tok/s
Fields
id
benchmarks/lhl-8060s-gpt-oss-120b-f16-vulkan-tg1Metric
decode_tok_s
Value
33.72
Unit
tok/s
Model
gpt-oss-120b-F16
Methodology
lhl strix-halo-testing llm-bench (github.com/lhl/strix-halo-testing, gpt-oss-120b-F16/results.jsonl). llama.cpp llama.cpp-vulkan_amdvlk, backend Vulkan, no -fa, ngl=99, tg1 batch-1 decode at depth 0 on Ryzen AI Max+ 395 (Radeon 8060S gfx1151). System RAM peak 68169MiB, avg power 22W, max temp 52.0C. Per-request batch-1.
Measured by
third_party
Reproducible
yes