|
|
Log in / Subscribe / Register

Toward better CPU idle-time predictions

Toward better CPU idle-time predictions

Posted Nov 7, 2014 5:53 UTC (Fri) by Russ.Dill@gmail.com (guest, #52805)
In reply to: Toward better CPU idle-time predictions by marcH
Parent article: Toward better CPU idle-time predictions

Keyboard events are a poor example. If you delay the processing of a keyboard event by 50ms, there won't be much of an effect on the system. But for other sporadic events, like IO, or network, delaying the processing can significantly effect the throughput of the system.

to post comments

Toward better CPU idle-time predictions

Posted Nov 7, 2014 17:09 UTC (Fri) by Darkmere (subscriber, #53695) [Link]

If you delay the processing of a keyboard event with 50 ms you suddenly have a 3 frame latency at 60 fps, and a 6 frame latency at the desired 120 fps that is needed for things like the Occulus rift.

And trust me, you notice it when you get a 3 frame higher input latency than before. It's a very noticeable thing when you're in a game.

Toward better CPU idle-time predictions

Posted Nov 8, 2014 17:00 UTC (Sat) by marcH (subscriber, #57642) [Link] (1 responses)

> If you delay the processing of a keyboard event by 50ms,

Are you suggesting this to save power?

I don't think the article was delaying anything, was it?

Toward better CPU idle-time predictions

Posted Nov 13, 2014 5:32 UTC (Thu) by Funcan (guest, #44209) [Link]

Mis-guessing that there is no keyboard even due, and going into a deep sleep state, causes delays


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds