ggml CUDA on DGX Spark GB10: Qwen2.5-Coder-7B Q8_0 tg32 d8192 (b7941)
27.38 tok/s (tg32, batch 1, per-request)
Fields
id
benchmarks/ggml-cuda-gb10-spark-qwen25-coder-7b-tg32-d8192-b7941Metric
decode_tok_s
Value
27.38
Unit
tok/s (tg32, batch 1, per-request)
Model
Qwen2.5-Coder-7B Q8_0 (ggml-org/Qwen2.5-Coder-7B-Q8_0-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=8192 is the same measurement after 8192 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 +/- 0.09. File size 7.54 GiB, params 7.62 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 +/- 0.09. File size 7.54 GiB, params 7.62 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).