View Single Post
Old 2020-02-06, 22:12   #70
"Patrick Poitras"
Oct 2019

2×33 Posts

That was one of my initial considered plans. Most of them were a no-go because they required either a network connection or at least some malleable database to generate work units on the fly, which the cluster is not in a position to handle. That would be the next step, if the current go-to-80 mining works well. The mining expedition is less of a "we need it" and more of a "these processor cycles would be wasted on benchmarks, but aliquot fulfills some of the requirements we need from a test, so might as well make them count"

I think we might very well run a bunch of the low-hanging fruit factors, but that would be to test the distribution-end of the cluster, which we aren't touching right now.

I've also considered running ECM on a bunch of >C140 sequences, but decided against it, as there are no clumsy output files that we need to recuperate. (The thing we're testing the cluster for, essentially)

As for BOINC, I'm already running it on my computers at home. It's currently set to only take small composites, but it uses the processor cycles that are not being used by something else. It's currently working on one of the sequences with a downdriver, exciting stuff!
PFPoitras is offline   Reply With Quote