Things like from-scratch test (including install) on each of the primary supported operating systems. Also, probably a test on Mathias's system as that would be quite good to have in the mix of checks. Plus whatever else we can think of. As many of these tests as possible should be on SDSC instances and ideally we can gradually automate a spin up of a range of target OS-based instances that are then automatically tested on.
Amongst the open questions: can we spin up Windows instances on SDSC that are connected to one of the GPUs we have there?
Things like from-scratch test (including install) on each of the primary supported operating systems. Also, probably a test on Mathias's system as that would be quite good to have in the mix of checks. Plus whatever else we can think of. As many of these tests as possible should be on SDSC instances and ideally we can gradually automate a spin up of a range of target OS-based instances that are then automatically tested on.
Amongst the open questions: can we spin up Windows instances on SDSC that are connected to one of the GPUs we have there?