CMP170HX Mining Cards Tested: 4x64GB for LLM Inference, Performance and Value Analysis
m94301 · reddit · 2026-08-12
User m94301 tests CMP170HX mining cards (4x8GB modded to 64GB each) for LLM inference. Benchmarks show single-card runs 20B models, multi-card runs up to 120B with 1M context. Provides token generation speeds, prefill speeds, and discusses PCIe bandwidth, power. Author says at $200 it's a steal, but performance is Ampere-class.
Related event: Benchmarking CMP170HX Mining GPUs for Efficient LLM Inference(2 posts)→
More from Infra
- AI-Driven Timing-Aware Power Optimization Uses Learned Ranking With Baseline-Preserving Repair — Amit Sharma · 2026-10-16
- Nscale AI datacenter plagued by coolant leaks, transformer fire, 10 safety violations — saranormous · 2026-10-16
- WSL 3.0 ships with native containers and Linux 6.18 kernel, syscall overhead down ~16% — solyarisoftware · 2026-10-16
- Are 22 GB modded RTX 2080 Ti cards a viable budget local LLM option? — Dave86ch · 2026-10-16
- Intel's neural-compressor: SOTA low-bit LLM quantization with FP4/NVFP4 support — HaihaoShen · 2026-10-16
- Tencent squeezes 770B-parameter Hunyuan Hy4 Preview into 214 GiB at ~2.38 bits per weight — coolmum18 · 2026-10-16