NVIDIA GTX 1080 Ti vs Intel Arc B580 12GB for local LLMs
3 reports on the NVIDIA GTX 1080 Ti and 10 on the Intel Arc B580 12GB
Which is faster for local LLMs?
On Llama 2 7B at 4-bit, the NVIDIA GTX 1080 Ti has 62.0 generation tokens/s (one run); prompt processing 1084.0 tokens/s (one run); the Intel Arc B580 12GB has generation speed not reported; prompt processing 4454.0 tokens/s (one run). Memory bandwidth, which caps how fast a card can write, is 6% higher on the NVIDIA GTX 1080 Ti (484 GB/s on the NVIDIA GTX 1080 Ti, 456 on the Intel Arc B580 12GB).
- Memory
- 11GB
- Memory bandwidth
- 484 GB/s
- FP16 compute
- 11.3 TFLOPS
- Reports
- 3
- Memory
- 12GB
- Memory bandwidth
- 456 GB/s
- FP16 compute
- 117 TFLOPS
- Reports
- 10
Measured on both
Generation and prompt processing speeds from plain runs (one device, one request, no speculative decoding, the whole model in the device's memory) of the same model size at the same bit class. Each side shows the engines and reported context behind its median. Context can mean a configured window or prompt depth, so these are community comparisons with differing settings. How to read these
| Model | NVIDIA GTX 1080 Ti | Intel Arc B580 12GB |
|---|---|---|
| Llama 2 7B4-bit | 62.0one runPP 1084.0 (one run)llama.cppcontext not stated | -0 runsPP 4454.0 (one run)llama.cppcontext not stated |
Estimated by the calculator
The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape. Where a model was also measured above, the measurement wins.
| Model | NVIDIA GTX 1080 Ti | Intel Arc B580 12GB |
|---|---|---|
| Llama 2 7B | 101.4 t/sQ3_K_S | 71.7 t/sQ4_K_S |
| Qwen3.8 27B | 1.6 t/sQ8_0, layers on CPU | 1.6 t/sQ8_0, layers on CPU |
| Qwen3.8 125B · 6B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 35B · 3B active | 13.8 t/sQ6_K, experts in RAM | 13.7 t/sQ6_K, experts in RAM |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 27B | 1.6 t/sQ8_0, layers on CPU | 1.6 t/sQ8_0, layers on CPU |
| Gemma 4 26B · 4B active | 8.4 t/sQ8_0, experts in RAM | 74.5 t/sQ2_K |
Change the context, RAM or model in the calculator: NVIDIA GTX 1080 Ti · Intel Arc B580 12GB
Every report on each card, with its source: NVIDIA GTX 1080 Ti · Intel Arc B580 12GB