NVIDIA A40 48GB vs NVIDIA CMP 170HX 64GB (unlocked) for local LLMs
4 reports on the NVIDIA A40 48GB and 10 on the NVIDIA CMP 170HX 64GB (unlocked)
Which is faster for local LLMs?
Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 2.1x higher on the NVIDIA CMP 170HX 64GB (unlocked) (696 GB/s on the NVIDIA A40 48GB, 1,493 on the NVIDIA CMP 170HX 64GB (unlocked)). The estimates below compare them on the models people run most.
- Memory
- 48GB
- Memory bandwidth
- 696 GB/s
- FP16 compute
- 150 TFLOPS
- Reports
- 4
- Memory
- 64GB
- Memory bandwidth
- 1,493 GB/s
- FP16 compute
- not on record
- Reports
- 10
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.
| Model | NVIDIA A40 48GB | NVIDIA CMP 170HX 64GB (unlocked) |
|---|---|---|
| Qwen3.8 27B | 14.2 t/sQ8_0 | 30.4 t/sQ8_0 |
| Qwen3.8 125B · 6B active | 64.5 t/sQ2_K | 127.0 t/sQ3_K_S |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 35B · 3B active | 44.7 t/sQ8_0 | 109.5 t/sQ8_0 |
| Qwen3.6 27B | 14.2 t/sQ8_0 | 30.4 t/sQ8_0 |
| Gemma 4 26B · 4B active | 38.3 t/sQ8_0 | 82.1 t/sQ8_0 |
Change the context, RAM or model in the calculator: NVIDIA A40 48GB · NVIDIA CMP 170HX 64GB (unlocked)
Every report on each card, with its source: NVIDIA A40 48GB · NVIDIA CMP 170HX 64GB (unlocked)