AMD MI50 32GB vs NVIDIA CMP 170HX 64GB (unlocked) for local LLMs
12 reports on the AMD MI50 32GB and 10 on the NVIDIA CMP 170HX 64GB (unlocked)
Which is faster for local LLMs?
On Qwen3.8 27B at 8-bit, the AMD MI50 32GB runs at 22.8 tokens per second (one run) and the NVIDIA CMP 170HX 64GB (unlocked) at 50.0 (one run). Those are community reports whose engines and settings differ, so the gap is a rough sign, not a controlled measure of the two cards. Memory bandwidth, which caps how fast a card can write, is 46% higher on the NVIDIA CMP 170HX 64GB (unlocked) (1,024 GB/s on the AMD MI50 32GB, 1,493 on the NVIDIA CMP 170HX 64GB (unlocked)).
- Memory
- 32GB
- Memory bandwidth
- 1,024 GB/s
- FP16 compute
- 26.5 TFLOPS
- Reports
- 12
- Memory
- 64GB
- Memory bandwidth
- 1,493 GB/s
- FP16 compute
- not on record
- Reports
- 10
Measured on both
Median tokens per second 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 quant level. Engines and context lengths can still differ between the runs. How to read these
| Model | AMD MI50 32GB | NVIDIA CMP 170HX 64GB (unlocked) |
|---|---|---|
| Qwen3.8 27B8-bit | 22.8one run | 50.0one run |
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 | AMD MI50 32GB | NVIDIA CMP 170HX 64GB (unlocked) |
|---|---|---|
| Qwen3.8 27B | 25.7 t/sQ6_K | 30.4 t/sQ8_0 |
| Qwen3.8 125B · 6B active | 19.6 t/sQ2_K, experts in RAM | 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 | 93.1 t/sQ5_K_M | 109.5 t/sQ8_0 |
| Qwen3.6 27B | 25.7 t/sQ6_K | 30.4 t/sQ8_0 |
| Gemma 4 26B · 4B active | 65.4 t/sQ6_K | 82.1 t/sQ8_0 |
| DeepSeek V4.1 Flash 552B · 16B active | doesn't fit at these settings | doesn't fit at these settings |
Change the context, RAM or model in the calculator: AMD MI50 32GB · NVIDIA CMP 170HX 64GB (unlocked)
Every report on each card, with its source: AMD MI50 32GB · NVIDIA CMP 170HX 64GB (unlocked)