Show HN: Proxima serves 4x more requests with no hardware change on vLLMgithub.com·3 pts·pythongiant·1
Show HN: KVBoost – chunk-level KV cache reuse for HuggingFace, 5–48x faster TTFTpythongiant.github.io·20 pts·pythongiant·18
CUDA Programming: From Zero to GPU Kernels – A Beginner's Guidepythongiant.github.io·1 pts·pythongiant·1
Show HN: I built GPT from scratch to understand how it workspythongiant.github.io·7 pts·pythongiant·0