<div class="gmail_quote">On Tue, Jun 9, 2009 at 1:40 PM, Kai-Uwe Behrmann <span dir="ltr"><<a href="mailto:ku.b@gmx.de">ku.b@gmx.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> Date: Mon, 8 Jun 2009 22:46:32 +0200<br>
> From: Sven Langkamp <<a href="mailto:sven.langkamp@gmail.com">sven.langkamp@gmail.com</a>><br>
<div class="im"><br>
> I have commited a first version of the iterator benchmark we dicussed<br>
> yesterday. It basically iterates over a paint device without doing memcpys.<br>
> The results are as we expected:<br>
><br>
> -HLine- and VLineIterator perform quite similar.<br>
> -the rect iterator about 10% faster than hline<br>
> -the random accesor is about 40% slower than hline<br>
<br>
</div>Could you give some numbers of iterator access/modification in pixel per<br>
seconds for a non Krita developer.</blockquote><div><br>The benchmark is pretty simple and you should be careful with the actual numbers. The values change a bit with different system load.<br>It's just measure how long it takes an iterator to iterate over each pixel paint device without doing a memcpy. Colorspace is RGB 8.<br>
<br>Running it on a Core Duo 2GHz I get:<br><br>HLineIterator 12,5 MPixel/sec<br>RectIterator 15,3 MPixel/sec<br>RandomIterator 8 MPixel/sec<br><br>With doing a memcpy I get:<br><br>
HLineIterator 5.8 MPixel/sec<br><br><br>I curious how that compares to other implementations.<br> </div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
thanks in advance,<br>
Kai-Uwe Behrmann<br>
<font color="#888888">--<br>
developing for colour management<br>
<a href="http://www.behrmann.name" target="_blank">www.behrmann.name</a> + <a href="http://www.oyranos.org" target="_blank">www.oyranos.org</a></font><br></blockquote></div><br>