Can the Tesla P40 run UserLM-2.0?
UserLM-2.0·8B·Tesla P40·24 GB VRAM
Yes. The Tesla P40 runs UserLM-2.0 fully in VRAM at FP16 / BF16 (about 17 GB of 24 GB), at roughly 12 tokens/s.
- best quant
- FP16 / BF16
- needs
- 17 GB
- usable
- 24 GB
- est. speed
- ~12 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the UserLM-2.0 page →
Every quantisation of UserLM-2.0 on the Tesla P40
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 4.1 GB | fits | ~72 tok/s |
| Q3_K_M | 5.3 GB | fits | ~49 tok/s |
| Q4_K_M | 6.3 GB | fits | ~39 tok/s |
| Q5_K_M | 7.1 GB | fits | ~33 tok/s |
| Q6_K | 8.0 GB | fits | ~29 tok/s |
| Q8_0 | 9.8 GB | fits | ~22 tok/s |
| FP16 / BF16 pick | 17 GB | fits | ~12 tok/s |
Speed is a bandwidth-based estimate for generating text (347 GB/s); prompt processing and long contexts are slower.
Measured speed on the Tesla P40
No community reports for this pair yet.
Report your own speed →Run UserLM-2.0 on the Tesla P40
Using Q8_0, the best fit for Tesla P40.
llama-server -hf liodon-ai/UserLM-2.0-imatrix-GGUF:Q8_0Downloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
vllm serve microsoft/UserLM-2.0 vllm serve liodon-ai/UserLM-2.0-FP8services:
vllm:
image: vllm/vllm-openai:latest
command: microsoft/UserLM-2.0
ports:
- "8000:8000"
ipc: host
volumes:
- ~/.cache/huggingface:/root/.cache/huggingface
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]API on port 8000.
sglang serve microsoft/UserLM-2.0 --host 0.0.0.0 --port 30000services:
sglang:
image: lmsysorg/sglang:latest
command: sglang serve microsoft/UserLM-2.0 --host 0.0.0.0 --port 30000
ports:
- "30000:30000"
ipc: host
volumes:
- ~/.cache/huggingface:/root/.cache/huggingface
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]API on port 30000.
from transformers import pipeline
pipe = pipeline("text-generation", model="microsoft/UserLM-2.0", dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "Explain self-hosting in one paragraph."}]
out = pipe(messages, max_new_tokens=512)
print(out[0]["generated_text"][-1]["content"])Chat models that fit the Tesla P40 comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 8B | Q8_0 | ≈9.8 GBFits in memory | ~22 tok/s | |
| 8.2B | Q8_0 | ≈10 GBFits in memory | ~22 tok/s | |
| 7.6B | Q8_0 | ≈8.9 GBFits in memory | ~24 tok/s | |
| 8B | Q8_0 | ≈9.8 GBFits in memory | ~22 tok/s | |
| 9.4B | Q8_0 | ≈11 GBFits in memory | ~19 tok/s | |
| 7.2B | Q8_0 | ≈9.0 GBFits in memory | ~25 tok/s |
Other GPUs that run UserLM-2.0
Frequently asked questions
- Can the Tesla P40 run UserLM-2.0?
- Yes. The Tesla P40 runs UserLM-2.0 fully in VRAM at FP16 / BF16 (about 17 GB of 24 GB), at roughly 12 tokens/s.
- How fast does UserLM-2.0 run on the Tesla P40?
- An estimate from memory bandwidth gives about 12 tokens/s at FP16 / BF16.
- Which quantisation of UserLM-2.0 should I use on the Tesla P40?
- FP16 / BF16: the largest one that fits. Expected quality: lossless compared with the full-precision weights.
Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.