Rik Kisnah - Blog
Posts
About
Publications
Resume
Teach
MCP
github
linkedin
#Performance
Measure GPU Utilisation and MFU
Nov 14, 2023
Measure GPU Utilisation and MFU
Diagnose a Slow NCCL Job
Mar 14, 2023
Diagnose a Slow NCCL Job
Number Formats: FP32 to FP8
Feb 15, 2022
Number Formats: FP32 to FP8
The GPU Memory Hierarchy
Nov 12, 2019
The GPU Memory Hierarchy
Design a Distributed Cache
Jul 10, 2018
Design a Distributed Cache