Submissions from fergusfinn.com

		Speculative KV coding: losslessly compressing KV cache by up to ~4× (fergusfinn.com)
		5 points by kkm 2 days ago \| past \| discuss
		70x faster cold(ish) starts for SGLang (fergusfinn.com)
		1 point by kkm 3 days ago \| past \| discuss
		Bringing Up DeepSeek-V4-Flash on AMD MI300X (fergusfinn.com)
		120 points by kkm 4 days ago \| past \| 23 comments
		Pushing memory bound CUDA kernels past the speed of light with data compression (fergusfinn.com)
		2 points by somnial 9 days ago \| past \| discuss
		Speculative KV coding: ~4× losslessly compressed KV cache using a small model (fergusfinn.com)
		2 points by somnial 25 days ago \| past
		In search of wasted bits: how much information do LLM weights carry? (fergusfinn.com)
		1 point by gmays 27 days ago \| past
		Redundant Information in LLM Weights (fergusfinn.com)
		5 points by mezark 32 days ago \| past
		Tans: Precomputing RANS (fergusfinn.com)
		3 points by mezark 37 days ago \| past
		Also-RANS: Asymmetric Numeral Systems for Entropy Coding (fergusfinn.com)
		25 points by mezark 37 days ago \| past
		70x faster cold(ish) starts for SGLang (fergusfinn.com)
		1 point by somnial 40 days ago \| past
		70x faster cold(ish) starts for SGLang (fergusfinn.com)
		4 points by mezark 43 days ago \| past
		Parallel Primitives for Multi-Agent Workflows (fergusfinn.com)
		1 point by mezark 4 months ago \| past
		LLM powered data structures: A lock-free binary search tree (fergusfinn.com)
		1 point by somnial 4 months ago \| past
		Parallel Primitives for Multi-Agent Workflows (fergusfinn.com)
		1 point by somnial 5 months ago \| past
		Scheduling in LLM Inference (fergusfinn.com)
		1 point by somnial 6 months ago \| past
		How fast can an LLM go? (fergusfinn.com)
		2 points by kkm 6 months ago \| past
		How fast can an LLM go? (fergusfinn.com)
		2 points by gmays 7 months ago \| past
		How fast can an LLM go? (fergusfinn.com)
		2 points by somnial 7 months ago \| past

HN For You