If you need an overview of what credentials Jenkins uses to clone repositories, you'll find that the Jenkins API is a bit convoluted. This is what I did instead:

1. From a Windows machine, access Jenkins as a UNC share and dump its contents:

Get-ChildItem -Path \\my-jenkins-instance\*\config.xml | ForEach-Object { $targetPath = "C:\Users\me\Desktop\jenkins\" + $_.Directory.Name + "_" + $_.Name;  $_ | Copy-Item -Destination $targetPath -Force }

The output will be a zip file with XML files containing job configurations. The filename contains the project name (as that information is not otherwise contained in the XML file).

You can use cgroups to cap the CPU usage of processes. Thus, you can make the shader compilation process use less CPU and produce less heat, in exchange for longer compilation times.

Requires: Linux, a somewhat recent kernel, possibly a package for cgcreate.

sudo cgcreate -g cpu:/slow # Create a CPU cgroup
sudo cgclassify -g cpu:slow $(pidof fossilize_replay) # Move all shader compilation processes to the cgroup
sudo cgset -r cpu.max='2000000 1000000' slow # Cap the CPU usage (in this case: to 2 CPUs)

To build support for ALPHA on gem5 you essentially need to compile v19 in the correct environment, which can be difficult due to Python incompatibilities. Here are the steps I took:

git clone https://github.com/ArturKlauser/gem5-dev

Apply this patch to gem5-dev/docker/gem5-dev.sh: