Zhihao Jia personal blog
Compiling LLMs into a MegaKernel: A path to low-latency inference
TL;DR: We developed a compiler that automatically transforms LLM inference into a single megakernel — a fused GPU kernel that performs…
- Points
- 314
- HN comments
- 76
- Submissions
- 1
- Cited by
- 0 commenters
AI & Machine LearningSystems & Performance2025
On Hacker News
| Submitted | Title on HN | Points | Comments |
|---|---|---|---|
| Jun 19, 2025 | Compiling LLMs into a MegaKernel: A path to low-latency inference by matt_d | 314 | 76 |
