mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Software

Reply
 
Thread Tools
Old 2016-09-08, 01:27   #1
Madpoo
Serpentine Vermin Jar
 
Madpoo's Avatar
 
Jul 2014

29·113 Posts
Default "Master" and "helper" threads

I ran across a funky situation today. I was doing some updates on a few machines and when swapping out memory, a good heavy run with Prime95 is always a good idea.

Well, one machine had an issue (the lights out controller on an HP) that caused, for whatever reason, HEAVY interrupts on the first core of the first CPU. Weird, but whatever, I figured I'd get to that later.

Prime95 was setup to use all 4 cores on that CPU in one worker, and the affinity was set to start with that very first core...the one getting all the system interrupts.

That resulted in that worker barely doing anything. It saw the heavy CPU usage on that one core and basically paused itself. The other 3 cores assigned to that worker weren't busy but Prime95 didn't pay attention to those.

The interesting thing was that I changed the first core on that worker to core #2 instead of core #1. Then that worker could run okay, although that busy core didn't really contribute anything.

This whole experience brought me back to when we were discussing this very thing... using the affinity mask to rearrange which core was first so that exactly this scenario (interrupt handling) would be mitigated.

So, just throwing this out there for posterity... it would indeed be a pretty good idea to change your affinity mask to make the first worker core something besides cpu #0. That first core seems to be the one it looks at (mostly?) to see if the system is busy enough that it should throttle down or whatever? Not too sure on the specifics.
Madpoo is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Aouessare-El Haddouchi-Essaaidi "test": "if Mp has no factor, it is prime!" wildrabbitt Miscellaneous Math 11 2015-03-06 08:17
Prime95 minor bug, "TortureThreads" not checked for maximum threads Explorer09 Software 1 2014-08-19 04:13
"no threads newer" counter cheesehead Forum Feedback 0 2008-09-17 05:14
Would Minimizing "iterations between results file" may reveal "is not prime" earlier? nitai1999 Software 7 2004-08-26 18:12

All times are UTC. The time now is 19:39.

Tue Oct 20 19:39:56 UTC 2020 up 40 days, 16:50, 1 user, load averages: 1.65, 1.76, 1.79

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.