Even if one can declare 100's of CPU classes, isn't there a limitation on _effectively_ how many distinct levels of CPU share can be supported ? Whats the highest that has been tried ?
We tried 5-10 .. But that doesn't matter.
First the share settings truely accept the share/total_share calculation. So if you are willing to go up to a larger level of total_share then CPU scheduler can distinguish that ..
The main reason can be seen from include/linux/ckrm_sched.h in update_local_cvt.
Here we adjust the normalized (by the effective share of the class) the nanosecs a task ran at deschedule time.
nsecs are typically large while weights are < total_share.
So given the above I think we can distinguish CPU shares in the order of 100 classes.
------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement on who ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech
------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement on who ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech
