ggml CUDA on DGX Spark GB10: gpt-oss-120b MXFP4 tg32 d0 (b7941)
58.72 tok/s (tg32, batch 1, per-request)
Fields
id
benchmarks/ggml-cuda-gb10-spark-gptoss120b-tg32-d0-b7941Metric
decode_tok_s
Value
58.72
Unit
tok/s (tg32, batch 1, per-request)
Model
gpt-oss-120b MXFP4 MoE (ggml-org/gpt-oss-120b-GGUF)
Methodology
MAINTAINER bench on one NVIDIA DGX Spark (GB10), build 11fb327bf (b7941). CUDA backend, ngl 99, flash attention on, ubatch 2048, mmap off, direct-io on. pp2048 is prefill of a 2048-token prompt, batch 1, per-request, NOT aggregate. tg32 is batch-1 decode of 32 new tokens, per-request. Depth d=0 is the same measurement after 0 tokens of prior context (plain row is depth 0). Value is the published mean; stddev is in notes. Single device, unified memory, no tensor parallel.
Measured by
third_party
Reproducible
yes
notes
Official ggml DGX Spark table, not a forum paste. stddev +/- 0.2. File size 59.02 GiB, params 116.83 B. A concurrent sweep at batch 1 through 32 is also on the page; those figures are aggregate and are not this record. Distinct from the earlier third-party Spark gpt-oss-120b CUDA sweep (different build, lower tok/s).
Notes
Official ggml DGX Spark table, not a forum paste. stddev +/- 0.2. File size 59.02 GiB, params 116.83 B. A concurrent sweep at batch 1 through 32 is also on the page; those figures are aggregate and are not this record. Distinct from the earlier third-party Spark gpt-oss-120b CUDA sweep (different build, lower tok/s).