Jason WittyandClaude Opus 5.5
698005c602
Standard benchmark, two fresh-restart runs: prompt -51% at 4K, -37% at 16K, decode -40% vs IQ3_S. 12/12 sanity checks. Kept active for husky-bench. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
75 lines
4.3 KiB
Markdown
75 lines
4.3 KiB
Markdown
# <user> standard benchmark: NVIDIA GeForce RTX 3090
|
|
|
|
Protocol `strata-qwen38-thinking-v1-draft.1` (draft), run `20261027T192209Z-58540dcc`, started 2026-10-27T19:22:09Z. **Status: nonconforming.**
|
|
|
|
Written by `<user> bench run`. The machine-readable result is [results.json](results.json); every request, its raw usage and timings and the client event offsets are in [outputs.jsonl](outputs.jsonl).
|
|
|
|
This run is not a complete standard result:
|
|
|
|
- no model file could be hashed
|
|
|
|
## Hardware and software
|
|
|
|
- GPU: NVIDIA GeForce RTX 3090 (24.0 GiB dedicated)
|
|
- CPU: Intel(R) Core(TM) i9-14900KS; RAM: 110.0 GiB
|
|
- Storage of the model files: not measured
|
|
- OS: NixOS 26.05 (Yarara) 6.18.55; drivers: gpu-driver 595.71.05; runtime: cuda 13.2
|
|
- <user> commit: `4b9e05a2a806a3fe24f7b2560d98626c37dade35`; engine 0.1.51; source build
|
|
- Background work: not measured
|
|
|
|
## Model and configuration
|
|
|
|
- Model: qwen3.8-flash-next-unsloth-ud-q4_k_xl; quantization UD-Q4_K_XL; revision not measured; source revision not measured
|
|
- Files: not measured
|
|
- Track: **custom** (UD-Q4_K_XL rep 1, 275 W, case closed)
|
|
- Context 131072 tokens; server capacity 1 slot(s)
|
|
- Sampling: temperature 1.0, top_p 0.95, top_k 20, min_p 0.0, presence_penalty 0.0, repetition_penalty 1.0, reasoning_effort xhigh; thinking on, preserved
|
|
- Engine options: `config.json`; chat template: `provenance/chat_template.jinja`
|
|
|
|
## Method
|
|
|
|
The frozen requests of `bench/prompts/standard-v1` (corpus `e75eff8b9feea04f...`) were sent over loopback HTTP, streaming, one at a time, in this order: 1 warm-up (excluded), 3 repetitions of the short, 4K and 16K prompts, and the twelve sanity checks. Prompt and decode speed are the engine's own `timings`; time to first token is the client's monotonic clock from dispatch to the first nonempty reasoning, answer or tool-call delta. No request was retried. Time after the model loaded: 2 min 39 s.
|
|
|
|
## Results
|
|
|
|
| Case | Prompt tok/s | Decode tok/s | TTFT s | First answer s | Total s | Capped |
|
|
| --- | --- | --- | --- | --- | --- | --- |
|
|
| short | n/a | 66.9 (66.6-70.1) n=3 | 1.14 (0.94-1.16) n=3 | 2.42 (2.33-2.93) n=3 | 5.9 (5.2-6.1) n=3 | 0/3 |
|
|
| prompt-4k | 845.8 (811.7-855.6) n=3 | 69.7 (67.2-73.5) n=3 | 4.86 (4.81-5.07) n=3 | 9.05 (8.27-9.82) n=2 | 12.1 (11.8-12.7) n=3 | 3/3 |
|
|
| prompt-16k | 1917.7 (1914.3-1919.4) n=3 | 67.8 (67.6-77.1) n=3 | 8.57 (8.57-8.60) n=3 | 15.11 (12.55-15.64) n=3 | 16.1 (15.2-16.1) n=3 | 3/3 |
|
|
|
|
Cells are median (min-max) n=samples. Short is a prompt of about 80 tokens: its prompt rate is dominated by the fixed cost of a request, so it is not shown; read prompt speed from prompt-4k and prompt-16k. Capped = requests that reached the 512-token cap, reasoning and answer together. A capped request's total time is the time to the cap, not to an answer, and it has no first-answer time when it was still reasoning.
|
|
|
|
| Phase | Planned | Completed | Truncated | Failed | Skipped |
|
|
| --- | ---: | ---: | ---: | ---: | ---: |
|
|
| warmup | 1 | 0 | 1 | 0 | 0 |
|
|
| performance | 9 | 3 | 6 | 0 | 0 |
|
|
| sanity | 12 | 12 | 0 | 0 | 0 |
|
|
| concurrency | 0 | 0 | 0 | 0 | 0 |
|
|
|
|
Sanity checks: **12 of 12**.
|
|
|
|
| Case | Result | Reason |
|
|
| --- | --- | --- |
|
|
| I1 | pass | the expected JSON value |
|
|
| I2 | pass | the expected JSON value |
|
|
| I3 | pass | the expected JSON value |
|
|
| I4 | pass | the exact expected text |
|
|
| R1 | pass | the exact expected text |
|
|
| R2 | pass | the exact expected text |
|
|
| R3 | pass | the exact expected text |
|
|
| R4 | pass | the expected JSON value |
|
|
| T1 | pass | the expected tool call |
|
|
| T2 | pass | the expected tool call |
|
|
| T3 | pass | the expected tool call |
|
|
| T4 | pass | the expected tool call |
|
|
|
|
## Failures, exclusions and limitations
|
|
|
|
- No request failed.
|
|
- Excluded from speed summaries: performance-prompt-4k-r3: first_answer_s is not available for this request
|
|
- Prompt-prefix reuse is the shared template prefix only; the cache was not emptied between requests, and the expert cache stays warm and evolves. These are not cold-model measurements.
|
|
- Output caps include reasoning; a request that reaches its cap is `truncated` and fails its sanity check.
|
|
- The runner cannot see other programs on the machine, the repeat-stop and loop-recovery switches of the server, or the quantization repository's revision; see `missing` in results.json for every value it could not read.
|
|
- The server had answered no request before this run (a freshly loaded engine).
|