Server Admin Guide: A Tale of Two Servers: BIOS, GPU, and NUMA Tuning for XCP-ng: Preserving the valuable work done by Tobias Kreidl (@tjkreidl)

johnnezero

WHAT: This post is dedicated to preserving the valuable server administration guide content produced by @tjkreidl, originally titled "A Tale of Two Servers." These articles were first published in 2019 on the Citrix Blogs, which are unfortunately no longer available.

KUDOS: Special thanks to @john.c for performing some incredible web sleuthing to recover this content, and to @tjkreidl for creating these essential guides in the first place â€” as well as for "Breathing Life" back into them by highlighting their importance before they disappeared forever into the ferocious "Bit-Bucket."

Series Overview & Quick Reference

If you're looking for a high-level summary of the concepts covered in this series — including specific XCP-ng/XO commands for BIOS power management, GPU scheduling, and NUMA inspection — start here:
Quick Reference & Summary Guide

The Complete Series

=========================================

Part 1: How BIOS Settings Can Affect Your Apps and GPU Performance:

A deep dive into CPU Power Management (OS DBPM vs. System DBPM) and Turbo mode.
Read Article

Part 2: GPU Settings and Advanced BIOS Tuning

Exploring Uncore frequency, C1E states, and NVIDIA GPU Scheduler modes (Best Effort vs. Equal Share).
Read Article

Part 3: NUMA, CPUs, Sockets/Cores, and VM Performance

Understanding vNUMA, vCPU oversubscription, and the importance of VM startup order for optimal memory placement.
Read Article

Archived for the XCP-ng community. Questions or additional insights? Please discuss below!

poddingue

This is great to see, thank you for taking the time to rescue this; and thanks to @john.c for the recovery work and to @tjkreidl for writing it in the first place.
I went looking, and there is a small XCP-ng-specific piece on this in the official docs under NUMA affinity (https://docs.xcp-ng.org/compute#numa-affinity), but it's nothing like the depth of the Tale of Two Servers series, so having the originals archived is genuinely useful.
I won't pretend to judge how much of the 2019 BIOS and GPU-scheduler guidance still maps cleanly onto current hardware and XCP-ng versions; others here will know where it's aged and where it hasn't.
I'll make sure this is on our radar on the docs side, because it keeps coming up. Really appreciate you keeping this from disappearing.

johnnezero

@poddingue Thank you!
"Anything and Everything we can to to improve XCP-ng", is the "Name of the Game"