> Date: Sun, 17 Jan 2010 13:33:25 +0300 > From: "Raistmer" <[email protected]> > Subject: Re: [boinc_dev] anonymous ATI applications not getting work > To: "Travis Desell" <[email protected]>, "Paul D. Buck" > <[email protected]> > Cc: BOINC Dev Mailing List <[email protected]> > Message-ID: <18547e28abdf4c2db6b29ba0cfac3...@jane> > Content-Type: text/plain; format=flowed; charset="iso-8859-1"; > reply-type=original > > It's weird indeed cause app_info clearly states that app used for ATI GPU: > > <coproc> > <type>ATI</type> > ... > </coproc> > > ----- Original Message ----- > From: "Travis Desell" <[email protected]> > To: "Paul D. Buck" <[email protected]> > Cc: "BOINC Dev Mailing List" <[email protected]> > Sent: Sunday, January 17, 2010 12:14 PM > Subject: Re: [boinc_dev] anonymous ATI applications not getting work > > >> Seems... kind of counterintuitive. Is this because there's no way to >> tell if the anonymous application is for GPUs or not? >> >> >> On Jan 17, 2010, at 4:07 AM, Paul D. Buck wrote: >> >>> For the Anon platform they have to set CPU to yes ... the others to no >>> ... >>> >>> On Jan 17, 2010, at 12:21 AM, Travis Desell wrote: >>> >>>> We're having a problem where people using an ATI application on an >>>> anonymous platform aren't able to get any work. Specifically, if >>>> they've selected: >>>> >>>> Use CPU >>>> (enforced by 6.10+ clients) no >>>> Use NVIDIA GPU if present >>>> (enforced by 6.10+ clients) no >>>> Use ATI GPU if present >>>> (enforced by 6.10+ clients) yes >>>> >>>> Then they won't get any work. Is this something wrong with our >>>> scheduler settings, or have any other projects come across this problem? >>>> >>>> ---------------------------------------------------------------------------------------------------------- >>>> Travis Desell >>>> <deselt @ cs.rpi.edu> >>>> 1-518-867-1054 >>>> Worldwide Computing Laboratory ( http://wcl.cs.rpi.edu/ ) >>>> milky...@home ( http://milkyway.cs.rpi.edu/ ) >>>> Rensselaer Polytechnic Institute, 110 8th Street, Troy NY 12180, USA >>>> ---------------------------------------------------------------------------------------------------------- >>>> >>>> >>>> >>>> _______________________________________________ >>>> boinc_dev mailing list >>>> [email protected] >>>> http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev >>>> To unsubscribe, visit the above URL and >>>> (near bottom of page) enter your email address. >>> >> >> ---------------------------------------------------------------------------------------------------------- >> Travis Desell >> <deselt @ cs.rpi.edu> >> 1-518-867-1054 >> Worldwide Computing Laboratory ( http://wcl.cs.rpi.edu/ ) >> milky...@home ( http://milkyway.cs.rpi.edu/ ) >> Rensselaer Polytechnic Institute, 110 8th Street, Troy NY 12180, USA >> ---------------------------------------------------------------------------------------------------------- >> >> >> >> _______________________________________________ >> boinc_dev mailing list >> [email protected] >> http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev >> To unsubscribe, visit the above URL and >> (near bottom of page) enter your email address. >> > > > > ------------------------------ > > _______________________________________________ > boinc_dev mailing list > [email protected] > http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev > > > End of boinc_dev Digest, Vol 68, Issue 41 > ***************************************** >
It's not weird, it's BOINC. People (me included) have been reporting this issue on project message boards for more than a week now. In fact this started, with Travis' last server update, which was supposed to help with ATI work availability, however most, if not all power users there use app_info.xml, because it enables one to tweak app behavior by setting cmd line parameters. And to calm everyone down, Collatz has exactly the same issue. No work on GPU request if there is app_info.xml present; only on CPU work requests any work gets downloaded. But, who cares. It's more important to have "fair" credits and social-networked I-don't-know-what, than efficiently using volunteered CPU cycles, and *GOD forbid* doing it according to user specification. But then again, this is just me, maybe, if I get me one of those spacebook or myface accounts, then even I will not care anymore, how efficiently my computers do what I volunteered them for. I could send you miles upon miles of logs showing this going on at MW and Collatz, but fear that even if logs showed the obvious, I would still be asked to turn on certain log flags and produce light-years upon light-years of logs, so I won't. BR, Vid Vidmar _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
