QLink: Compiling a GPU into Qwen Weights to Execute CUDA

Inspired by QwenGPU, QLink compiles a small GPU into Qwen3.5-2B's weights so it executes CUDA kernels by decoding, with Claude-written mini-NCCL "qccl" running Ring AllReduce across four model-GPUs.

2026-11-03 ~ 2026-11-03 · 3 related posts

• QLink compiles a GPU into Qwen weights and runs CUDA kernels via token decoding — zeeshanp_ · 2026-11-03
• QLink insight: one model's store is already a valid memory write for another — zeeshanp_ · 2026-11-03
• qccl: a tiny NCCL running Ring AllReduce across 4 QwenGPU model-GPUs — zeeshanp_ · 2026-11-03