On Tue, Dec 20, 2022 at 04:32:00PM +0000, Colton Lewis wrote: > Alexandru Elisei <alexandru.elisei@xxxxxxx> writes: > > > Though I'm not sure how you managed to get MAX_SMP to go down to 6 cores > > on > > a 12 core machine. MAX_SMP is initialized to $(getconf _NPROCESSORS_ONLN), > > so the body of the loop should never execute. I also tried it on a 6 core > > machine, and MAX_SMP was 6, not 3. > > > Am I missing something? > > To be clear, 12 cores was a simplified example I did not directly > verify. What happened to me was 152 cores being cut down to 4. I was > confused why one machine was running a test with 4 cores when my other > machines were running with 8 and traced it to that loop. In effect the > loop was doing MAX_SMP=floor(MAX_SMP / 2) until MAX_SMP <= 8. I printed > the iterations and MAX_SMP followed the sequence 152->76->38->19->9->4. Ah, I think I understand now. Were you running 32-bit arm tests? If so, it'd be good to point that out explicitly in the commit message (the 'arm:' prefix in the summary is ambiguous). Assuming the loop body was running because it needed to reduce MAX_SMP to 8 or lower for 32-bit arm tests, then we should be replacing the loop with something that caps MAX_SMP at 8 for 32-bit arm tests instead. Thanks, drew _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
