Re: Performance Adjustments: Silly Frames and poll() timeout

On Tue, 6 Apr 2004, Lei Zhang wrote:
> So you have 768 /3 = 256 sessions in a single process?  Can you make 
> your application host smaller number of sessions?  I played a little 
> with the timeout value of poll() a little a while ago, my purpose then 
> was to see if that helps reduce the CPU usage of my beepcore-c based 
> application - and noticed the problem you described in this thread.

I have done some more real testing, and find that subtle changes in the 
vicinity of this poll() call have a rather large impact on performance 
(latency in particular).  It looks like for my current setup about a 10 
msec timeout to poll() is optimal; reducing it further does not help and 
seems to slightly increase CPU consumption.  But going up to even 20 msec 
makes things noticeably slower.

Who wrote this code?  It is obviously rather intricate and in many ways 
hard to improve upon, but it seems to me some sort of change may be called 
for (if nothing else, to always set the timeout to 1/HZ).

-------------------------------------------------------
-- David C. Niemi           Adeptech Systems, Inc.   --
-- Reston, Virginia, USA    http://www.adeptech.com/ --
-------------------------------------------------------



-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click