Archive

Dive into a collection of past posts and rediscover timeless content.

How AMD MI300X Hardware Shapes Inference Optimization

Oct 6, 2026

•

4 min read

How AMD MI300X Hardware Shapes Inference Optimization

Three Levels of Inference Profiling on the AMD MI300X

Oct 2, 2026

•

1 min read

Three Levels of Inference Profiling on the AMD MI300X

DeepSeek on vLLM V1: The Bottleneck Moved from KV Cache to Burst Admission

Jun 7, 2026

•

6 min read

DeepSeek on vLLM V1: The Bottleneck Moved from KV Cache to Burst Admission

Why your GPU is idle but your requests are queued

May 30, 2026

•

3 min read

Why your GPU is idle but your requests are queued

Why LLM Inference Needs Two Different GPUs

Mar 13, 2026

•

5 min read

Why LLM Inference Needs Two Different GPUs