On Tue, Jul 10, 2018 at 12:43 PM Jeroen Demeyer <j.deme...@ugent.be> wrote: > > On 2018-07-10 12:37, Erik Bray wrote: > > If I want to limit > > parallelization on a machine that I'm concurrently using for other > > work, that's kind of a decision a human has to make. > > One could argue the other way around: if I want Sage to use all > resources on the machine that it's running on, that's kind of a decision > a human has to make. In other words: guessing is hard. > > The default of 1 was certainly meant as feature: the idea being that > parallellism is something that should be enabled explicitly. A bit like > GNU make, which uses a single process by default but optionally supports > multi-processing.
I think this is good argument, but it's also a bit of a problem in that @parallel is supposed to be kind of "out-of-the-box" parallelization, and if its default is not provide any parallelization I envision users saying "I put @parallel on my code and it didn't run any faster" (if anything it's slower). So I don't think 1 is a good low-end default for that alone. Maybe instead min(2, os.cpu_count()) if you're going to argue from that end. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
