Im not talking about turning HT off Im talking about setting Nuke to 16 from 8 on dual core. I tested it on 3 boxes with different comps and on none was it faster to set nuke to use more the the max real cores. but it was faster to run 2 instances set to 8 each to render. I don't recall this being the case on z800's running linux.
Randy S. Little http://reel.rslittle.com http://imdb.com/name/nm2325729/ <http://www.linkedin.com/in/rslittle> On Fri, Feb 3, 2012 at 20:46, Diogo Girondi <[email protected]> wrote: > >From what I've seen these new procs run faster with HT on. The old dual > core Xeons (circa 2007) on the other hand use to be faster with the HT > turned off. At least with Nuke from the tests I did at the time. > > On Sat, Feb 4, 2012 at 12:47 AM, Deke Kincaid <[email protected]>wrote: > >> I just said that someone on Ars-technica did a test and found that >> using hyperthreading was faster then not using it. Not that I tested >> it. >> >> -deke >> >> On Fri, Feb 3, 2012 at 16:59, Randy Little <[email protected]> wrote: >> > Peter and Deke both said thats not the case anymore and that using the >> hyper >> > threads since westmer i7 came out would be just as fast. To which I >> said >> > what you just said. I just wanted to verify it. In fact even in comp >> in >> > the GUI 8 is faster then 16. So either Peter and Deke are wrong or Mac >> > threading isn't on par with the other platforms. >> > >> > Randy S. Little >> > http://reel.rslittle.com >> > http://imdb.com/name/nm2325729/ >> > >> > >> > >> > >> > On Fri, Feb 3, 2012 at 16:42, Nathan Rusch <[email protected]> >> wrote: >> >> >> >> I would actually *expect* the two separate instances of 8 threads >> apiece >> >> to outperform a single instance using 16, for a couple reasons: >> >> >> >> 1) Not everything in Nuke is multithreaded. Throwing more threads at >> some >> >> things won’t get you anywhere, but throwing (effectively) two machines >> at >> >> them will. >> >> >> >> 2) There have been test results both from within The Foundry (albeit >> some >> >> time ago) and from end users reporting performance degradation when >> trying >> >> to get Nuke to make full use of virtual CPUs. Single-threaded operators >> >> notwithstanding, a performance increase using two 8-thread processes >> (even >> >> though they’re technically utilizing the virtual cores) could be the >> result >> >> of letting the OS kernel handle the CPU scheduling for the virtual >> cores >> >> between the two processes. >> >> >> >> -Nathan >> >> >> >> >> >> From: Randy Little >> >> Sent: Friday, February 03, 2012 4:07 PM >> >> To: Nuke user discussion >> >> Subject: Re: [Nuke-users] Threads >> >> >> >> dual quad xeon 8 cores real 8 virtual. (westmers?) running 2 instances >> >> with 8 shouldn't be faster then a single 16 especially if you take all >> the >> >> extra memory and i/o thats happening since its doubled. (this just >> goes back >> >> to that previous thread where they where asking to change the default >> >> threads to include the virtual cores. >> >> >> >> Randy S. Little >> >> http://reel.rslittle.com >> >> http://imdb.com/name/nm2325729/ >> >> >> >> >> >> >> >> >> >> On Fri, Feb 3, 2012 at 15:54, Nathan Rusch <[email protected]> >> >> wrote: >> >>> >> >>> How many physical cores do you have? >> >>> >> >>> -Nathan >> >>> >> >>> >> >>> From: Randy Little >> >>> Sent: Friday, February 03, 2012 3:47 PM >> >>> To: Nuke user discussion >> >>> Subject: [Nuke-users] Threads >> >>> >> >>> ok after the previous conversation I did some test here. on current >> gen >> >>> xeon MACS. literally I can render 2 of the same comp with interleaved >> >>> frames significantly faster then the same comp set to 16 threads. >> Its not >> >>> even close. LIke 30% faster to render 2 instances of nuke on the >> same comp >> >>> with 8 threads vs one version on 16. Are we still having threading >> issues >> >>> on the Mac side. >> >>> >> >>> Randy S. Little >> >>> http://www.rslittle.com >> >>> >> >>> >> >>> ________________________________ >> >>> _______________________________________________ >> >>> Nuke-users mailing list >> >>> [email protected], http://forums.thefoundry.co.uk/ >> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >>> >> >>> _______________________________________________ >> >>> Nuke-users mailing list >> >>> [email protected], http://forums.thefoundry.co.uk/ >> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> >> >> >> >> >> >> ________________________________ >> >> _______________________________________________ >> >> Nuke-users mailing list >> >> [email protected], http://forums.thefoundry.co.uk/ >> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> >> >> >> >> _______________________________________________ >> >> Nuke-users mailing list >> >> [email protected], http://forums.thefoundry.co.uk/ >> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> > >> > >> > >> > _______________________________________________ >> > Nuke-users mailing list >> > [email protected], http://forums.thefoundry.co.uk/ >> > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> _______________________________________________ >> Nuke-users mailing list >> [email protected], http://forums.thefoundry.co.uk/ >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> > > > _______________________________________________ > Nuke-users mailing list > [email protected], http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >
_______________________________________________ Nuke-users mailing list [email protected], http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
