[cod] Some feedback on 1.41d

Joe Hunkeler wolve at dodfrontline.com
Sat Sep 25 17:10:38 EDT 2004


I get what you're saying.  Alright, from now on I'll just report the 
real CPU usage.  Although it would make sense for the total usage to 
report the true overall usage of the 4 CPUs.  It's not a 2 CPU 2 Virtual 
system though, the specs actually read quad.

Steven Hartland wrote:

> Ok so in this case your saying that cod is using 23%
> ( if I follow the thread correctly ). If this is the case its
> as I suspected we aren't comparing like for like.
> You are stating figures from total virtual CPU. Where
> as I'm stating figures as single real CPU. This is always
> how CPU reading should be done.
>
> Taking into account virtual CPU's is pointless, for one
> they will gain hardly anything in a game server environment.
> I wish they did but they don't in fact in all tests we performed
> they only resulted in lag in high load situations and hence
> your should really disable HT.
>
> Now comparing like for like you are using ~90% CPU
> of a single real CPU in comparison to my quoted 70%
> CPU. As I have shown, in that graph posted earlier,
> load varies wildly on a map / gametype basis. So unless
> we happen to be running the same gametype and map
> we cant accurately compare results even now.
>
> No on was calling you stupid, we all have things we
> can learn and the point I was trying to make is that
> a 2.4 based kernel CAN actually record incorrect CPU
> loads. An example of this is I could right an app which
> uses ~99% of the available CPU and still record 0%
> in top and ps. This is was no reflection on you reading
> your results but since you have posted then it does
> show you are reading them wrong. I will say again
> always quote the % from a single CPU so at least we
> are always in the same ball park.
>
>    Steve / K
> ----- Original Message ----- From: "Joe Hunkeler" 
> <wolve at dodfrontline.com>
>
>> I wasn't aware that I was stupid :).
>>
>> Linux pub 2.4.21-9.0.3.ELsmp #1 SMP Tue Apr 20 19:49:13 EDT 2004 i686 
>> i686 i386 GNU/Linux
>>
>> CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
>>           total   23.2%    0.0%    0.1%   0.0%     0.3%    0.0%   76.1%
>>           cpu00    0.2%    0.0%    0.0%   0.2%     0.6%    0.0%   99.0%
>>           cpu01    0.2%    0.0%    0.0%   0.0%     0.0%    0.0%   99.8%
>>           cpu02    4.7%    0.0%    0.5%   0.0%     0.0%    0.0%   94.6%
>>           cpu03   87.8%    0.0%    0.0%   0.0%     1.0%    0.0%   11.2%
>> Mem:  1028456k av,  972912k used,   55544k free,       0k shrd,  
>> 104108k buff
>>                    611440k actv,  100884k in_d,   11280k in_c
>> Swap: 2040244k av,   33568k used, 2006676k free                  
>> 426336k cached
>>
>> 4557 pubuo     25   0 95412  93M  2772 R    18.6  9.2 477:13   2 
>> coduo_lnxded
>>
>> Looks accurate enough for me...
>
>
>
>
> ================================================
> This e.mail is private and confidential between Multiplay (UK) Ltd. 
> and the person or entity to whom it is addressed. In the event of 
> misdirection, the recipient is prohibited from using, copying, 
> printing or otherwise disseminating it or any information contained in 
> it.
> In the event of misdirection, illegible or incomplete transmission 
> please telephone (023) 8024 3137
> or return the E.mail to postmaster at multiplay.co.uk.
>
>
>
>




More information about the Cod mailing list