[netperf-talk] CPU Usage

[chandramani]

Thanks for the info.

I am running netserver on windows XP and netperf on Soekris board with debian. 

By the way when I run netperf several times, throughput remains almost invariant but I get large variation in "service demand" and "utilization" values though I don't do anything with setup.

What could be the reason behind this? 
How to get the true measure of "service demand" and "utilization". 

regards
chandra
 



"Rick Jones"<rick.jones2 at hp.com> wrote:
chandramani wrote:
> Hi all
> 
> I am new to Netperf. Recently I installed netperf and ran it with following command
> 
> netperf -H "remotehostip" -l 20 -c -C 
> and got following result
> 
> Recv Send Send Utilization Service Demand
> Socket Socket Message Elapsed Send Recv Send Recv
> Size Size Size Time Throughput local remote local remote
> bytes bytes bytes secs. 10^6bits/s % T % L us/KB us/KB
> 
> 8192 16384 16384 20.01 23.03 42.67 17.45 151.773 62.056
> 
> On net I got little information regarding 'service demand'
> performence mesure. But I didn't get any thing regarding
> 'utilization'. Specifically what do symbols '%T' and '%L' mean?

IIRC they may be described in the manual.

http://www.netperf.org/svn/netperf2/trunk/doc .

Basically they describe the method used to measure the CPU utilization. 
  'L' means the "loopers" or CPU soaker processes.  'T' means that the 
"times" method was used, and that is not a good thing - what are your 
platforms?

> Is there any relationship between 'utilization' and 'service
> deamand'?

Yes. Service demand is the normalization (as it were) of the CPU 
utilization and the throughput.

rick jones

> 
> thanking in advance for any sort of help
> 
> regards
> chandra
> 
> _______________________________________________
> netperf-talk mailing list
> netperf-talk at netperf.org
> http://www.netperf.org/cgi-bin/mailman/listinfo/netperf-talk