Speed test matrix
Qwen3-0.6B-FP8 on CMP 170HX 64GB
vllm · 0.32.0rc1.dev249+g18305948a · FP83 contexts (0–32K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K
Hardware
- Type
- Discrete GPU
- GPU
- CMP 170HX
- VRAM
- 64 GB
- CPU
- 2x Intel Xeon E5-2630 v3
- RAM
- 64 GB
- OS
- Ubuntu 26.04.1
Engine
- Engine
- vllm
- Version
- 0.32.0rc1.dev249+g18305948a
- Quantization
- FP8
- Backend
- cuda
- Repository
- github.com/vllm-project/vllm
- Build
- ghcr.io/morrowmake/vllm-cmp170hx@sha256:5b45c4d6feafefe2eba962375b0a9a0b64fd22b9e05abb4d1966f14a32a2db79
Run
- Served model
- bench
- Revision
- main
- Benchmark tool
- llm-inference-bench v0.8.0
- Decode test
- Fixed duration
- Prefill mode
- Measured during decode warm-up
- Time per measurement
- 30 s
- Max output
- 8,192 tokens
Launch command
# Untuned out-of-the-box baseline. Spec decode: none # 1x NVIDIA CMP 170HX 64GB (cmpunlocker), 180 W limit, PCIe Gen2 x4 | 2x Intel Xeon E5-2630 v3 | 64 GB DDR4 | Ubuntu 26.04.1, NVIDIA driver 610.57.04 | vLLM: Morrowmake vllm-cmp170hx image vllm serve Qwen/Qwen3-0.6B-FP8 --max-model-len 40960 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code
Prefill
prompt tok/s by context length8K context27.5K298 ms TTFT
16K context23.3K696 ms TTFT
32K context17.4K1,863 ms TTFT
Sustained decode
total tok/s · rows are context length, columns are simultaneous requests| Context | 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128 |
|---|---|---|---|---|---|---|---|---|
| 0 | 341.753 ms TTFT | 593.489 ms TTFT | 904.278 ms TTFT | 1,352.795 ms TTFT | 2,001.4154 ms TTFT | 2,832.2281 ms TTFT | 3,926.4390 ms TTFT | Over capacity |
| 16K | 225.2164 ms TTFT | 337.2203 ms TTFT | 464.1309 ms TTFT | 578.3408 ms TTFT | 663.5759 ms TTFT | Over capacity | Over capacity | Over capacity |
| 32K | 167.7234 ms TTFT | 228.9341 ms TTFT | 285.4503 ms TTFT | 330.7764 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity |
Over capacity: not run because this many requests at this context would exceed the server's KV cache.
TTFT is measured after the benchmark warms the server's prompt cache, so it mostly excludes prefill.
All measurements
Latencies are medians. Total tok/s is output across all simultaneous requests.| Phase | Context | Requests at once | Status | Total tok/s | Per request tok/s | TTFT | Time between tokens | Average active requests | Completed requests | Output tokens | Measured for |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Prefill | 8K | 1 | Valid | 27.5K | — | 298 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 23.3K | — | 696 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 17.4K | — | 1,863 ms | — | — | 1 | — | — |
| Sustained decode | 0 | 1 | Valid | 341.7 | 352.8 | 53 ms | 2.8 ms | 1.0 | 1 | 10,251 | 30.0 s |
| Sustained decode | 0 | 2 | Valid | 593.4 | 321.1 | 89 ms | 3.1 ms | 2.0 | 2 | 17,800 | 30.0 s |
| Sustained decode | 0 | 4 | Valid | 904.2 | 303.2 | 78 ms | 3.5 ms | 4.0 | 4 | 27,132 | 30.0 s |
| Sustained decode | 0 | 8 | Valid | 1,352.7 | 189.8 | 95 ms | 5.3 ms | 8.0 | 0 | 40,576 | 30.0 s |
| Sustained decode | 0 | 16 | Valid | 2,001.4 | 145.8 | 154 ms | 6.9 ms | 16.0 | 0 | 60,032 | 30.0 s |
| Sustained decode | 0 | 32 | Valid | 2,832.2 | 105.5 | 281 ms | 9.5 ms | 32.0 | 0 | 84,992 | 30.0 s |
| Sustained decode | 0 | 64 | Valid | 3,926.4 | 73.9 | 390 ms | 13.5 ms | 64.0 | 0 | 117,786 | 30.0 s |
| Sustained decode | 0 | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 1 | Valid | 225.2 | 227.6 | 164 ms | 4.4 ms | 1.0 | 0 | 6,754 | 30.0 s |
| Sustained decode | 16K | 2 | Valid | 337.2 | 171.2 | 203 ms | 5.8 ms | 2.0 | 0 | 10,114 | 30.0 s |
| Sustained decode | 16K | 4 | Valid | 464.1 | 117.7 | 309 ms | 8.5 ms | 4.0 | 0 | 13,920 | 30.0 s |
| Sustained decode | 16K | 8 | Valid | 578.3 | 73.0 | 408 ms | 13.7 ms | 8.0 | 0 | 17,344 | 30.0 s |
| Sustained decode | 16K | 16 | Valid | 663.5 | 41.7 | 759 ms | 24.0 ms | 16.0 | 0 | 19,904 | 30.0 s |
| Sustained decode | 16K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 1 | Valid | 167.7 | 169.1 | 234 ms | 5.9 ms | 1.0 | 0 | 5,031 | 30.0 s |
| Sustained decode | 32K | 2 | Valid | 228.9 | 115.2 | 341 ms | 8.7 ms | 2.0 | 0 | 6,864 | 30.0 s |
| Sustained decode | 32K | 4 | Valid | 285.4 | 71.8 | 503 ms | 13.9 ms | 4.0 | 0 | 8,560 | 30.0 s |
| Sustained decode | 32K | 8 | Valid | 330.7 | 41.5 | 764 ms | 24.1 ms | 8.0 | 0 | 9,920 | 30.0 s |
| Sustained decode | 32K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 128 | Over capacity | — | — | — | — | — | — | — | — |
Over capacity: not run because this many requests at this context would exceed the server's KV cache.