This is the mail archive of the
ecos-patches@sourceware.org
mailing list for the eCos project.
Re: cpuload.cxx patch
- From: Andrew Lunn <andrew at lunn dot ch>
- To: Gerster Jochen-B01096 <B01096 at freescale dot com>
- Cc: ecos-patches at ecos dot sourceware dot org
- Date: Mon, 14 Aug 2006 21:30:28 +0200
- Subject: Re: cpuload.cxx patch
- References: <DF02D66C37C26B4C8018BF84F0E37FEC0D62CD01@zwg18exm01.ea.freescale.net>
On Mon, Jul 31, 2006 at 03:03:11PM +0200, Gerster Jochen-B01096 wrote:
> Hi
>
> here is the patch file for the cpuload.cxx so the average is no calc correctly (I hope)
> changes:
> - the reason was a bracket which wasn't set correct
> improved:
> - minimizes rounding errors
>
>
> So anyone can have a look at it and validate it!
>
> Attached the patch file :)
>
> Thanks to Jonathan Larmour for his support in using the diff tool!
>
>
> Jochen
Hi Jochen
I've at last got around to testing this patch. It works as advertised.
However the change to the averaging algorithm changes how fast the 1s
and 10s averages change. There is no real definition of what these two
figures mean, so the old version is just as correct as your version.
I'm not sure it is a good idea to change them, since users may have
become used to these meaningless figures meaning.
What is they advantage of your average algorithm?
Andrew