Porting Ninfer to CMP 170HX doubles Qwen performance with technical tweaks
ubrtnk · reddit · 2026-08-23
Details the porting of Ninfer (a high-performance inference engine) to the CMP 170HX data center GPU, achieving significant speedups:
- Context: Leveraged architectural similarities between RTX 3090 and CMP 170HX, using AI agents to assist in the porting process.
- Technical Fixes: Resolved cudaErrorCooperativeLaunchTooLarge caused by SM count differences (70 vs 82) by dynamically adjusting split-K schedules. Removed Blackwell-only kernels and CUDA forward-compat libs.
- Results: Achieved a 2x performance increase on Qwen3.6-35B, supporting 262K context with efficient VRAM usage (Int8 KV Cache).
Includes detailed Docker run configs and llama-swap integration settings.
More from Infra
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16
- Cerebras pushes disaggregated inference: heterogeneous prefill pools feeding wafer-scale decode — ai · 2026-10-16