Gerwin Klein

For this specific case: the platform doesn't have a timer yet, so it doesn't really make sense to build sel4bench for it. There is probably a mechanism for that already...

I'm temporarily switching off QUARTZ64 until this issues is resolved, so that we can get sel4bench results in the meantime.

This needs more work. If we change something, we need to change it both in the deployment workflow and the PR workflow. We don't want them to go out of...

seL4/seL4#523 is now merged 🙂

That does sound reasonable. @kent-mcleod any concerns on adding a build and run with clang for all benchmarks? It'll take more machine time, but it could be informative.

Then again, the x86 runs each take 50 min, and if we want comparable numbers they need to run on the same machine, so we would be blowing the time...

If we want gcc/clang to be comparable, should it not be the exact same machine? In that case it doesn't matter if we buy more. We do have more x86...

You're right, if the tutorial as described is broken, this should definitely remain open.

Something seems funky with the commits, the PR contains some that are already in master. Can you try rebasing your branch over master?

> Think that's fixed now, something weird happened when I was trying to fix an overly long commit message. Thanks, that is better. The commits now unfortunately all have two...