That's what the theory says, yes, and I think the design document too. But when I was testing multithreaded apps at Milkway, I saw the same as YoYo is reporting now.
>________________________________ > From: "McLeod, John" <[email protected]> >To: yoyo <[email protected]>; BOINC Developers Mailing List ><[email protected]>; BOINC Projects <[email protected]> >Sent: Monday, 30 June 2014, 19:36 >Subject: Re: [boinc_dev] Creditnew for mt apps > > >For 2 tasks from the same WU, two computers should be getting the same credit >whether they are running 1 core or 8 cores. The single core computer will >just take longer to get it done. > >If both computers have 8 cores, and one is running 1 8 core task at a time, >and one is running 8 1 core task at a time, and they are otherwise similar, I >would expect that the credit per hour is similar. > >-----Original Message----- >From: yoyo [mailto:[email protected]] >Sent: Monday, June 30, 2014 2:33 PM >To: McLeod, John; BOINC Developers Mailing List; BOINC Projects >Subject: Re: [boinc_dev] Creditnew for mt apps > >But the users claims, that the systems gets the same credits per hour >even if they use different amount of cores. > >McLeod, John schrieb: >> Yes, it is intentional. They should be using more cores, but for >> proportionally less time. (8 CPU cores for 1/8th of the time that a single >> CPU core can do the task.) >> >> -----Original Message----- >> From: boinc_dev [mailto:[email protected]] On Behalf Of yoyo >> Sent: Monday, June 30, 2014 2:11 PM >> To: BOINC Developers Mailing List; BOINC Projects >> Subject: [boinc_dev] Creditnew for mt apps >> >> Hello, >> >> there is a discussion in YAFU >> (http://yafu.myfirewall.org/yafu/forum_thread.php?id=160) where users >> complains, that computer which uses more cores to finish a work unit >> gets the same credits as computer which are using less cores. >> Is this the intention of creditnew or a bug? >> >> YAFU runs the stock Boinc server- and wrapper code together with the >> sample_bitwise_validator, sample_assimilator and quorum=1. >> >> kind regards, >> yoyo >> _______________________________________________ >> 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 >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 To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
