5

Possible Duplicate:
How do we get this magic performance-boosting 200 line patch?

Will the 26.38 natty kernel have this patch.

Community
  • 1
nidzo732
  • 395
  • 1
  • 10
  • It isn't magic. If you are foolish enough to run 64 cpu hogging tasks in a terminal window, it makes a dramatic difference, but nobody does that. This is a case of much ado about nothing. – psusi Mar 17 '11 at 13:27
  • @psusi System jobs are partitioned out. For users with heavy apt-postinst processes, servers or databases running in the background, they should notice that the system is much more responsive. I agree it isn't magic and it won't help everybody but for some users this *will* make a large difference and in my mind, it's well worth the hoo-haa. Progress is good. – Oli Mar 17 '11 at 14:24
  • 1
    Regarding the duplicates, we're dealing with a temporal issue. The two upstream threads of this both have value but [one](http://askubuntu.com/questions/13619) is largely speculative and answered justly and [the other](http://askubuntu.com/questions/13562) is instructional for older releases (well before the pull into 2.6.38). I know I have bias (I have an answer here) but I'd be happy leaving this open because it has a more definitive answer for 11.04. Still slightly speculative but I can't see a good reason for the situation to change now. – Oli Mar 17 '11 at 14:29
  • @Oli Yes, it will help if you are running a multithreaded/multiprocessing cpu bound server on your desktop, but I would question how common that is, and the sanity of anyone doing it ;) – psusi Mar 17 '11 at 15:00
  • @psusi I assume it also helps if you are running distributed computing projects (Boinc, SETI, Folding@home ...), if you are a doing hour-long background compiles, probably a handful of other use cases as well. Although a majority of users may never e.g. compile kernels or run Boinc, I wouldn't go so far as to call those that do so insane :) – j-g-faustus Mar 17 '11 at 15:15
  • @j-g-faustus only if they are multithreaded/multiprocessed. All this patch does is schedule all of the tasks in a terminal as a group, so if there are a dozen cpu hungry tasks in a terminal, then your foreground gui task is competing with 12 tasks for the cpu without this patch, but effectively only one with this patch. Then the dozen tasks fight with each other for the slice allocated to the group as a whole. If you are only talking about one task instead of a dozen, then this patch does nothing. Most people don't do background compiles with 64 threads ;) – psusi Mar 17 '11 at 15:19
  • @psusi Agree that most users probably won't notice a difference. (Although the grouping is apparently [no longer terminal-based](https://lwn.net/Articles/418884/).) Whether I will remains to be seen :) (I'm running Boinc and F@H, attempting to maximize CPU usage and desktop responsiveness simultaneously.) – j-g-faustus Mar 17 '11 at 16:18
  • @psusi I wouldn't say that people running processor intensive tasks in the terminal are foolish. Perhaps a lot of them, are in fact, working. – Zoe Mar 17 '11 at 22:26
  • @Zoe Gagnon if they are running one or two, sure. 64 on the other hand, is foolish. – psusi Mar 17 '11 at 22:51

0 Answers0