[cod] Linux vs. Windows

Steven Hartland killing at multiplay.co.uk
Sun Nov 21 08:36:37 EST 2004


Basically the mistake people make on windows is to read off
the straight % where as the real % is that number multiplied by the
number of CPU's show. e.g.
1. A Single P4 with HT enabled or a Dual without HT enabled
showing 25% for a server is actually 25% * 2 = 50%
2. A Dual P4 with HT enable showing 25% for a server is
actually using 25% * 2 = 100% ( bad :P )

This mistake can also be made on linux if you are using the totals.
Also 2.4 based kernels have an issue where they incorrectly report
CPU usage there is no way around this apart from upgrading to
a 2.6 kernel or highering the HZ to 1000 ( unconfirmed ).

    Steve / K
----- Original Message ----- 
From: "Christopher Choo" <chrischoo at fragnetics.com>

> Not being very technically proficient myself I find this statement very
> interesting. Yes, it does seem unusual to me at a 32-player server can be taking
> about 15% on a Windows machine. However, is there a more accurate way in which
> we should read our CPU usage? Similarly for Linux, if top sometimes misbehaves,
> what program would give more accurate results? Would inaccuracies in measuring
> CPU usage using Dual-Processors or HT in Windows also show up for Linux?
> When you mentioned that actual usage is about 4x as reported in task manager,
> what is the reason behind this and where can I find out more?




================================================
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