Megakernels help when launch profile is the bottleneck, says _xjdr
_xjdr · x · 2026-11-20
xjdr argues that "we don't do megakernels" is a naive take. Megakernels help when your launch profile is the major bottleneck: if you know the model you're running and the batch size, they let you omit redundant loads and pointless shape conversions, improving inference efficiency.
More from Infra
- Quantum Computing Takes Aim at AI Workloads as IBM and QuEra Map Fault-Tolerant Roadmaps — mikeflache · 2026-11-20
- Tencent Earnings Call: Latest Chip Restrictions Will Accelerate China's Domestic Semiconductor Supply Chain — basedjensen · 2026-11-20
- Chorus paper: shared n-gram tables cut memory 87% while beating DeepSeek's Engram — brown2green · 2026-11-20
- px0: A 20MB browser-based IDE for AI-generated code, built by chopping layers with first-principles thinking — arpit_bhayani · 2026-11-20
- Tsinghua-spinoff Qingyi launches Ascend 950DT 10,000-GPU token factory, lifting utilization from 40% to 80% — 量子位 · 2026-11-20
- SenseNova-Loop 9B runs on 8GB VRAM at 22 t/s, uses half the tokens of Qwen3.8 — R_Duncan · 2026-11-20