<div dir="ltr">I am pretty sure that you can run open GL over TCPIP</div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Jul 15, 2015 at 5:04 PM, John 'Warthog9' Hawley <span dir="ltr"><<a href="mailto:warthog9@eaglescrag.net" target="_blank">warthog9@eaglescrag.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On 07/15/2015 08:42 AM, Ashot Arshakyan wrote:<br>
> Hello. I have a project where I need to render thousands of OpenGL<br>
> frames each second by using a x86 based PC and a typical high end NVidia<br>
> GPU.<br>
> The problem is I have a pretty limited space for the electronics of the<br>
> device I'm working on so I was hoping I could use a single board PC<br>
> instead of using a mini ATX.<br>
> SInce the main rendering job is done on the GPU, I thought the<br>
> Minnowboard CPU would be fast enough for the job. What are your thoughts?<br>
> Now I just need to know how I can connect a standard GPU to Minnowboard.<br>
> I'm thinking of powering the GPU with a flat 12V 30A power supply<br>
> instead of ATX to again save some space.<br>
<br>
</div></div>So there's a couple of things you should be aware of while pondering<br>
this. It *MIGHT* work, but you'll need to consider that modern graphics<br>
cards use a LOT of bandwidth for moving data in and out of the graphics<br>
card. They usually employee 16 lanes of PCI-e, which at it's oldest and<br>
slowest is about 4GByte/s. So there's two potential issues here:<br>
1) The MinnowBoard only has a single lane of PCI-e v2<br>
available for use through the high speed expansion<br>
header. This gives you a maximum throughput of<br>
about 500MByte/s<br>
<br>
2) *MANY* graphics cards won't negotiate their PCI-e<br>
lanes down to the point where it will work in a x1<br>
connector. This is mostly an assumption from the<br>
manufacturers that the card would never be used in<br>
such a slot. Case in point when we tried to get a,<br>
admittedly rather high end, graphics card working on<br>
the MAX on April 1st this year, it didn't work as<br>
the card couldn't negotiate down to a single PCI-e<br>
lane<br>
<br>
The only way to overcome 2, that I can think of, is to put a PCI-e<br>
switch on the bus that can provide 16 lanes, but only upstream via 1.<br>
This still limits your maximum bandwidth to 500MByte/s though, but it<br>
should get you past the lane negotiation issue. That's a substantial<br>
amount of work though, and likely a need for a custom lure to be built<br>
to accommodate it. Just some thoughts.<br>
<span class="HOEnZb"><font color="#888888"><br>
- John 'Warthog9' Hawley<br>
_______________________________________________<br>
elinux-MinnowBoard mailing list<br>
<a href="mailto:elinux-MinnowBoard@lists.elinux.org">elinux-MinnowBoard@lists.elinux.org</a><br>
<a href="http://lists.elinux.org/mailman/listinfo/elinux-minnowboard" rel="noreferrer" target="_blank">http://lists.elinux.org/mailman/listinfo/elinux-minnowboard</a><br>
</font></span></blockquote></div><br></div>