#2293 Community benchmark on NVIDIA GeForce RTX 4080 SUPER

Description

# Community benchmark on NVIDIA GeForce RTX 4080 SUPER

Measured on 2026-10-14 with `tools/strata_bench.py` (suite 1, 3 runs per test, medians with ranges). Speeds are the engine's own timings; first-token times are measured by the client over HTTP. Speed only: no correctness check was run.

## Hardware and software

- GPU and VRAM: NVIDIA GeForce RTX 4080 SUPER, 16 GiB VRAM
- CPU: Intel(R) Core(TM) i7-9700K CPU @ 3.60GHz (8 cores, 8 threads); installed RAM: 63.9 GiB; storage: not measured
- OS: Windows 11 (10.0.26200); GPU driver: 617.42; CUDA: not read
- Strata: project 0.1.43.1, commit a0116b27; engine 0.1.43.1
- Background workloads and power limits: not recorded

## Model and configuration

- Model and quantization: qwen3.8-flash-next-iq2_xs
- Context: 131072 tokens; parallel slots: 1
- Settings: the install's own configuration, unchanged; thinking off for the suite; temperature 0

```text
strata.exe --pack iq2_xs --native Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00001-of-00002.gguf --ple-gguf Qwen3.8-Flash-Next-GSQ-RCO-IQ2_XS-00002-of-00002.gguf --expert-profile expert-profile.bin --expert-cache auto --prefill auto --spec 4 --spec-min-p 0.5 --mtp rt --max-context 131072 --kv int8 --kv-resident 32768
```

## Method

Run by `python tools/strata_bench.py` against the HTTP API of a Strata server (already running); model loading is not timed. One warm-up chat and a 1K prompt are discarded first. Every prompt starts with a unique tag, so the conversation cache cannot help the fresh-prompt tests.

- short chat: the three questions of `tools/ab_engine.py`, up to 100 new tokens, 3 runs
- 4K / 16K prompts: generated word lists sized by a calibration request, 24 new tokens, 3 fresh prompts each
- agent session: system prompt and four tools, then 5 turns that each add about 1500 words of a project file and continue from the model's own earlier answers (the way agent clients resend a conversation), 64 new tokens per turn, 3 sessions; the cache numbers come from the answer's usage (`cached_tokens`) and timings (`prompt_n`)
- concurrency (when run): N requests sent at the same moment, total tokens over the wall time of the burst, latency over all requests

## Results

| Configuration | Actual prompt tokens | Reused tokens | Generated tokens | Runs | Prompt tok/s median and range | Decode tok/s median and range | TTFT seconds median and range |
| --- | ---: | ---: | ---: | ---: | --- | --- | --- |
| short chat | 46 (43-51) | 0 | 100 (100-100) | 9 | not measured | 67.9 (44.4-77.9) | 0.45 (0.39-0.68) |
| 4K prompt, fresh | 4019 (4019-4019) | 0 (0-0) | 24 cap | 3 | 1405.7 (1382.5-1409.1) | not measured | 2.88 (2.88-2.92) |
| 16K prompt, fresh | 16058 (16058-16058) | 0 (0-0) | 24 cap | 3 | 2507.9 (2497.1-2510.3) | not measured | 6.45 (6.43-6.47) |
| agent session, follow-up turns | 14467 (14454-14519) at the last turn | see below | up to 64 | 3 | see below | 82.2 (54.0-108.9) | 3.02 (2.74-3.08) |

Ranges are the lowest and highest of the runs; `Runs` counts requests that returned a token.

### Agent session (conversation cache)

Cache hit over the follow-up turns: 68.7 % (68.6-68.8); earlier-turn tokens read again per session: 72 (3-74) (earlier turns that the cache did not hold; the 13634 (13633-13634) new tokens of the files are read once); first turn (nothing cached) first token: 2.28 s (2.25-2.29).

| Turn (run 1) | Prompt tokens | Reused | Read | Prompt tok/s | First token (s) | Decode tok/s |
| ---: | ---: | ---: | ---: | ---: | ---: | ---: |
| 0 | 609 | 0 | 609 | 271.9 | 2.28 | 47.1 |
| 1 | 3721 | 641 | 3080 | 1033.5 | 3.016 | 54.0 |
| 2 | 5254 | 3785 | 1469 | 541.4 | 2.738 | 75.9 |
| 3 | 8698 | 5318 | 3380 | 1118.9 | 3.061 | 71.4 |
| 4 | 11622 | 8691 | 2931 | 980.3 | 3.019 | 96.2 |
| 5 | 14519 | 11673 | 2846 | 956.7 | 3.005 | 89.0 |

Per-request data: `results.json`. Engine log lines from the start: `engine-log-excerpt.txt` (when this run started the server).

## Correctness and limitations

Not measured: this is a speed suite. The conversation uses the model's own replies, so token counts differ a little between runs. Machine details come from the server, `nvidia-smi`/`rocm-smi` and the OS; background load and power limits are not recorded. Total time of the suite: 118 s.