Mirpri's Blog

Why Decoding is memory-bound for LLMs and how to optimize it

Research May 14, 2026

Breaking down the bottlenecks in LLM decoding and how speculative decoding can help optimize performance.

[Read Full Story...]

Catagories

Research Algorithm Course Uncategorized Development

More

The Knapsack Problem: From 0/1 to Advanced

Algorithm Feb 25, 2026
P5960 【模板】差分约束

Algorithm Feb 23, 2026
P2803 学校选址 II

Algorithm Feb 22, 2026
Steam DLC

Uncategorized Feb 22, 2026
P1020 导弹拦截

Algorithm Feb 21, 2026
The New std::print in C++23: A Modern Replacement for cout and printf

Development Feb 20, 2026
Set Up a WordPress Website

Development Feb 19, 2026
Mastering React and Vue, Why I Still Give WordPress a Try

Development Feb 19, 2026
Markdown Style Guide

Uncategorized Jun 19, 2024
Using MDX

Uncategorized Jun 1, 2024