Home › GPUs & Macs › NVIDIA RTX 4080 SUPER vs NVIDIA Tesla P100
NVIDIA RTX 4080 SUPER vs NVIDIA Tesla P100 for local LLMs
Across 4 tests measured on both, the NVIDIA RTX 4080 SUPER is faster in 4 and generates a median 2.16× the tokens/s of the NVIDIA Tesla P100.
On Llama 2 7B (Q4_0, llama-bench) the NVIDIA RTX 4080 SUPER generates 150 tokens/s and the NVIDIA Tesla P100 66.5 tokens/s (measured).
Same model, same quant, same test
| Model | Quant / test | RTX 4080 SUPER tok/s | Tesla P100 tok/s | Faster | Sources and run details |
|---|---|---|---|---|---|
| Llama 2 7B | Q4_0 llama-bench | 150 Vulkan | 66.5 Vulkan | RTX 4080 SUPER +126% | llama.cpp #10879 RTX 4080 SUPER · llama.cpp #10879 Tesla P100 |
| Llama 3.1 8B Instruct | Q4_K_M LocalScore | 66.6 llamafile | 32.3 llamafile | RTX 4080 SUPER +106% | LocalScore RTX 4080 SUPER · LocalScore Tesla P100 |
| Llama 3.2 1B Instruct | Q4_K_M LocalScore | 194 llamafile | 120 llamafile | RTX 4080 SUPER +62% | LocalScore RTX 4080 SUPER · LocalScore Tesla P100 |
| Qwen2.5 14B Instruct | Q4_K_M LocalScore | 44.9 llamafile | 19.8 llamafile | RTX 4080 SUPER +126% | LocalScore RTX 4080 SUPER · LocalScore Tesla P100 |
Generation speed in tokens/s, best measured backend for each chip (shown under the number). Backends can differ between the two chips; check them before drawing conclusions. Measured results only, nothing estimated.
Specs
| NVIDIA RTX 4080 SUPER | NVIDIA Tesla P100 | |
|---|---|---|
| Memory | 16 GB | 16 GB |
| Memory bandwidth | 736 GB/s | 732 GB/s |
All results for the NVIDIA RTX 4080 SUPER · All results for the NVIDIA Tesla P100 · More comparisons