lkml.org 
[lkml]   [1999]   [Jul]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Scheduling latencies news: less RAM = less latency
On Sat, Jul 31, 1999 at 05:44:30PM -0700, Linus Torvalds wrote:
> > I'm easily willing to trade 1-5% of the CPU in exchange of a responsive <5ms
> > latency system.
>
> It shouldn't be necessary. In most cases bad latency should be traceable
> to bad performance - you can improve latency by _improving_ performance
> rather than making it worse.

I'm _really_ in favor of low latency soft rt tasks in Linux, but adding
calls to the scheduler has a cost and it may not be easy to measure
until it is too late.

while size !=0
get a buffer
copy from user space to buffer
send buffer to device queue
decrement size

may be a whole lot faster than

while size !=0
get a buffer
copy from user space to buffer
send buffer to device queue
if(needs_resched)schedule();
decrement size

simply because of cache locality + schedule cost + process switch time.
So which is more important?


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:53    [W:0.075 / U:0.920 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site