LLM Server Setup Part 1 – Base OS

This post is Part 1 in a series on how to configure a system for LLM deployments and development usage. The configuration will be suitable for multi-user deployments and also useful for smaller development systems. Part 1 is about the base Linux server setup.

How To Create A Docker Container For AMD AOCCv4 Compiler Plus Spack Build Tools

AMD has recently released version 4.0 of their AOCC compiler which includes support for AVX512 on the Zen4 architecture. This post details building a Docker image containing the Spack package manager/build system together with AMD AOCCv4.0.0 compilers. This will be used as the build image for multi-stage Dockerfiles that will be used to compile scientific applications and benchmarks with targeted Zen3/4 optimizations. It is the first step in that process.

UPDATE v0.2 NVIDIA GPU Powerlimit Setup

This is just a short post to announce a more usable version of the NVIDIA GPU powerlimit setup script that I released a few months ago. This update to version 0.2 uses an interactive mode to set GPU powerlimits and optionally setup a systemd unit file to set these limits on subsequent reboots.

LTSP Configuration for Benchmark Platform of Diskless Workstations

In this post we look at using a testing Lab of Windows systems as a benchmarking platform for Linux scientific application using network boot with nfsroot and home mounts. Linux is boot on the systems “diskless” leaving the Windows installs untouched. LTSP turned out to be a great time saver for setting up the configuration.

NVIDIA GPU Power Limit vs Performance

This post presents testing data showing that power-limit reduction on NVIDIA GPUs have give significant benefits for both high wattage and lower wattage GPUs. Power-limit vs Performance data is presented for 1-4 A5000 and 1-4 RTX3090 GPUs.