Tinkering with Mining Cards: 50% Speed Boost for llama.cpp on CMP 170HX
fragment_me · reddit · 2026-08-14
A developer shared performance tuning experiences running llama.cpp on two unlocked CMP 170HX mining cards (65GB).
By forcing GGMLCUDAFORCECUBLAS ON, the Prompt Processing (PP) speed for the Qwen 3.6 27B model increased from 1k to 1.5k tokens per second, marking a roughly 50% improvement. However, the author notes that even with this boost, the overall speed remains slower than standard cards like the RTX 3090.
Related event: Benchmarking CMP170HX Mining GPUs for Efficient LLM Inference(2 posts)→
More from Infra
- New hardware day: local LLM deployment the moment it arrives — lxfater · 2026-10-16
- Together AI's score centering fixes RL training collapse under quantized inference — nrehiew_ · 2026-10-16
- Inference is splitting prefill and decode across specialized chips, not just H100s — AccBalanced · 2026-10-16
- Rebasely launches in beta: Git hosting built for humans and coding agents — handotdev · 2026-10-16
- Drama3 TTS lands in audio.cpp: up to 5x real-time on RTX 5090, <8GB VRAM mode too — Acceptable-Cycle4645 · 2026-10-16
- Three CoreWeave co-founders have sold $3.3B in shares since IPO amid credit worries — SumitGup · 2026-10-16