Speed test matrix
NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 on RTX 3090 24GB
vllm · 0.32.0 · NVFP44 contexts (0–56K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 56K
Hardware
- Type
- Discrete GPU
- GPU
- RTX 3090
- VRAM
- 24 GB
- CPU
- Intel Core i7-9700K (8C)
- RAM
- 32 GB
- OS
- Ubuntu 24.04.5
Engine
- Engine
- vllm
- Version
- 0.32.0
- Quantization
- NVFP4
- Backend
- cuda
- Repository
- github.com/vllm-project/vllm
- Build
- vllm/vllm-openai:v0.32.0-x86_64
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: MTP, 3 tokens lookahead
# 1x RTX 3090 24GB (Zagitta/gpu-unlock), 420 W limit, PCIe Gen3 x16 | Intel Core i7-9700K 8C | 32 GB DDR4 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4 --max-model-len 65536 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --speculative-config {"method":"mtp","num_speculative_tokens":3} --max-cudagraph-capture-size 16Prefill
prompt tok/s by context length8K context8,1601,029 ms TTFT
16K context6,7532,462 ms TTFT
32K context7,7374,276 ms TTFT
56K context7,0268,221 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 | 167.1592 ms TTFT | 264.21,896 ms TTFT | 261.7Requests queued30,332 ms TTFT | 262.6Requests queued31,198 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity |
| 16K | 125.31,020 ms TTFT | 128.2Requests queued25,220 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 32K | 132.91,144 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 56K | 137.61,273 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
Requests queued: the server could not run all requests at once, so some waited in a queue. 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 | 8,160.0 | — | 1,029 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 6,753.0 | — | 2,462 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 7,737.0 | — | 4,276 ms | — | — | 1 | — | — |
| Prefill | 56K | 1 | Valid | 7,026.0 | — | 8,221 ms | — | — | 1 | — | — |
| Sustained decode | 0 | 1 | Valid | 167.1 | 163.7 | 592 ms | 6.1 ms | 1.0 | 0 | 5,013 | 30.0 s |
| Sustained decode | 0 | 2 | Valid | 264.2 | 134.1 | 1,896 ms | 7.5 ms | 2.0 | 0 | 7,921 | 30.0 s |
| Sustained decode | 0 | 4 | Requests queued | 261.7 | 132.1 | 30,332 ms | 7.6 ms | 2.0 | 2 | 7,851 | 30.0 s |
| Sustained decode | 0 | 8 | Requests queued | 262.6 | 131.9 | 31,198 ms | 7.6 ms | 2.0 | 2 | 7,874 | 30.0 s |
| Sustained decode | 0 | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 0 | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 0 | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 0 | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 1 | Valid | 125.3 | 128.0 | 1,020 ms | 7.8 ms | 1.0 | 0 | 3,759 | 30.0 s |
| Sustained decode | 16K | 2 | Requests queued | 128.2 | 151.8 | 25,220 ms | 6.7 ms | 1.0 | 1 | 3,845 | 30.0 s |
| Sustained decode | 16K | 4 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 1 | Valid | 132.9 | 134.3 | 1,144 ms | 7.4 ms | 1.0 | 0 | 3,987 | 30.0 s |
| Sustained decode | 32K | 2 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 4 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 32K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 1 | Valid | 137.6 | 138.9 | 1,273 ms | 7.2 ms | 1.0 | 0 | 4,129 | 30.0 s |
| Sustained decode | 56K | 2 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 4 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 56K | 128 | Over capacity | — | — | — | — | — | — | — | — |
Requests queued: the server could not run all requests at once, so some waited in a queue. Over capacity: not run because this many requests at this context would exceed the server's KV cache.