| |||||||||||||
The next stepThe next stepPosted Dec 17, 2004 3:00 UTC (Fri) by bronson (subscriber, #4806)Parent article: Boot-time clock frequency selection
Once this is finished, the next step is obvious: make it fully dynamic! A daemon could watch the system shift between I/O and interactive use and adjust HZ to fit (similar to speedfreq et al).
Scary...
(Log in to post comments)
The real next step Posted Dec 17, 2004 12:15 UTC (Fri) by xav (guest, #18536) [Link] The Real Next Step would be to totally get rid of the timer interrupt, and thus have the processor really sleeping when the system is idle. There have already been some patches on lkml a while ago, I don't remember why they were not merged but I liked the idea very much.
|
|||||||||||||