somnial

Recent Submissions

1. 1. ▲ What happens when a GPU writes memory (blog.doubleword.ai) 5 points · 1 month ago · 0 comments
2. 2. ▲ What happens when a GPU reads memory? (blog.doubleword.ai) 15 points · 2 months ago · 0 comments
3. 3. ▲ The case for disaggregated LLM serving (blog.doubleword.ai) 4 points · 2 months ago · 0 comments
4. 4. ▲ On-the-fly snapshot compression for elastic inference at scale (blog.doubleword.ai) 6 points · 2 months ago · 1 comment
5. 5. ▲ NVLink, NVSwitch, and All That (blog.doubleword.ai) 5 points · 3 months ago · 2 comments
6. 6. ▲ The Anatomy of an Instruction Pipeline Hazard (hiraditya.github.io) 9 points · 3 months ago · 0 comments
7. 7. ▲ Width vs. Depth: Speculating on the Margin (blog.doubleword.ai) 17 points · 3 months ago · 1 comment
8. 8. ▲ Pushing memory bound CUDA kernels past the speed of light with data compression (fergusfinn.com) 2 points · 4 months ago · 0 comments
9. 9. ▲ Speculative KV coding: ~4× losslessly compressed KV cache using a small model (fergusfinn.com) 2 points · 5 months ago · 0 comments
10. 10. ▲ 70x faster cold(ish) starts for SGLang (fergusfinn.com) 1 point · 5 months ago · 0 comments