Llama 2 7B on Tesla P40
59.4best gen tok/s
1,080best prompt tok/s
640max context tested
4results
Measured · all numbers below come from the linked sources
llama.cpp
| Quant | Gen tok/s | Prompt tok/s | Source | Context tested | TTFT | Build |
|---|---|---|---|---|---|---|
| Q4_0 | 59.4 | 488 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @N/A |
| Q4_0 | 59.2 | 484 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @N/A |
| Q4_0 | 54.7 | 1,007 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @c76b420 |
| Q4_0 | 53.7 | 1,080 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @c76b420 |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
"Context tested" is the largest prompt+generation length actually benchmarked (tokens), not a memory-fit estimate.