<p dir="ltr">Sorry for my language, I mean exactly that. So where of the way problem to archieve more linear improve when we increase the number of cores. I suppose that maybe is because the scheme of memory needed to render the model. Maybe I s not possible process all using the cache memery of each kernell. </p>
<div class="gmail_quote">On 29 Mar 2014 00:09, "Matthew Heimlich" <<a href="mailto:matt.heimlich@gmail.com">matt.heimlich@gmail.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
What do you mean by "process the model"? Currently models are<br>
processed before exporting to the GPU, and they are then copied<br>
completely into vRAM. All the GPU does is trace rays and calculate<br>
shader ops, then display the result.<br>
<br>
On Fri, Mar 28, 2014 at 6:50 PM, GeKo <<a href="mailto:geko.pua@gmail.com">geko.pua@gmail.com</a>> wrote:<br>
> But i really mean about the way that cycles process de model. I dont have<br>
> any knowledge about raytracing render, but i am shocked about how the new<br>
> hardware know improve so much rhe rendering, even when they have 5 time more<br>
> cores.<br>
><br>
> On 28 Mar 2014 23:11, "Brecht Van Lommel" <<a href="mailto:brechtvanlommel@pandora.be">brechtvanlommel@pandora.be</a>><br>
> wrote:<br>
>><br>
>> There are various possibilities:<br>
>><br>
>> * Memory and texture memory read/write bottlenecks<br>
>> * Different Mhz for cores<br>
>> * Different GPU architecture<br>
>> * Not all cores occupied due to divergence or mismatched batch sizes<br>
>> * CPU Overhead<br>
>><br>
>> There's too many interacting parts to be sure without closely<br>
>> investigating.<br>
>><br>
>> Brecht.<br>
>><br>
>> On Fri, Mar 28, 2014 at 9:30 PM, GeKo <<a href="mailto:geko.pua@gmail.com">geko.pua@gmail.com</a>> wrote:<br>
>> > Short question, Why the execution time don't decrease "linearly"<br>
>> > according<br>
>> > to the number of cuda cores?<br>
>> ><br>
>> > _______________________________________________<br>
>> > Bf-cycles mailing list<br>
>> > <a href="mailto:Bf-cycles@blender.org">Bf-cycles@blender.org</a><br>
>> > <a href="http://lists.blender.org/mailman/listinfo/bf-cycles" target="_blank">http://lists.blender.org/mailman/listinfo/bf-cycles</a><br>
>> ><br>
>> _______________________________________________<br>
>> Bf-cycles mailing list<br>
>> <a href="mailto:Bf-cycles@blender.org">Bf-cycles@blender.org</a><br>
>> <a href="http://lists.blender.org/mailman/listinfo/bf-cycles" target="_blank">http://lists.blender.org/mailman/listinfo/bf-cycles</a><br>
><br>
><br>
> _______________________________________________<br>
> Bf-cycles mailing list<br>
> <a href="mailto:Bf-cycles@blender.org">Bf-cycles@blender.org</a><br>
> <a href="http://lists.blender.org/mailman/listinfo/bf-cycles" target="_blank">http://lists.blender.org/mailman/listinfo/bf-cycles</a><br>
><br>
_______________________________________________<br>
Bf-cycles mailing list<br>
<a href="mailto:Bf-cycles@blender.org">Bf-cycles@blender.org</a><br>
<a href="http://lists.blender.org/mailman/listinfo/bf-cycles" target="_blank">http://lists.blender.org/mailman/listinfo/bf-cycles</a><br>
</blockquote></div>