Llama 2 7B on P106-100
30.8best gen tok/s
438best 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 | 30.8 | 183 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @23bc779 |
| Q4_0 | 30.6 | 438 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @5fd160b |
| Q4_0 | 30.4 | 407 | llama.cpp (CUDA) ↗ | 512+128 | — | llama.cpp CUDA @5fd160b |
| Q4_0 | 29.8 | 184 | llama.cpp (Vulkan) ↗ | 512+128 | — | llama.cpp Vulkan @23bc779 |
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.