BitNet runs on CPU, while Qwen3.6-35B hits 100 tok/s on 2× B60 XPU
QuixiAI · x · 2026-07-22
- The author says BitNet can run on CPU and calls it a big deal.
- In a reply, they add that Qwen3.6-35B-A3B NVFP4 is running at 100 tok/s on 2× B60 XPU using their SYCL kernels.
- The post is essentially a performance/inference claim about pushing low-bit models onto commodity or near-commodity hardware.
Related event: QuixiAI Enables Cross-Backend Execution and Boosts CPU Inference(3 posts)→
More from Infra
- Gigatoken claims 500–1000× speedups over Hugging Face tokenizers — ZainHasan6 · 2026-07-22
- UK datacenter plans run into a water shortage dispute over national infrastructure status — Sure_Ad_9884 · 2026-07-22
- NVIDIA’s Rubin adds tile-level kernel dependency triggers for finer overlap — Moh1tAgarwal · 2026-07-22
- A billboard claims one company can deliver an AI data center in nine months — BenBajarin · 2026-07-22
- India’s chip fabs still depend on imported ultra-pure chemicals — santoshpanda · 2026-07-22
- Sakana AI and NVIDIA show a GPU-friendly way to cut LLM data waste — SakanaAILabs · 2026-07-22