Thanks for that, though that makes sense to me if the multicore application is using all cores. What if it's using less than the maximum amount, like in this case? And it only happens when a GPU app is used in combination with the MT app here.
Although it **is** a nice green feature, as the system is less loaded, uses less electricity, generates less heat. :-) I'll go wait for Milkyway to return from their server switchover. Then try to get some of their N-Body multithreaded tasks. Those use all cores. Then see what that does when a GPU app is running, as if I read it correctly, as long as the GPU app is running --and using a minute amount of CPU-- the multithreaded app shouldn't start, right? On Thu, Jan 5, 2012 at 8:26 PM, David Anderson <[email protected]> wrote: > The current policy is: > If a multicore app is running, don't overcommit the CPUs > (i.e. don't schedule 4.1 threads on 4 cores). > > This is because multicore apps may run inefficiently > if the CPUs are even slightly overcommitted > (at least, that was the case with AQUA). > > We can reconsider this if there's evidence that the > above assumption doesn't hold in general. > > -- David > -- -- Jord. _______________________________________________ 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.
