It is currently Mon Feb 17, 2020 11:46 am



Reply to topic  [ 7 posts ] 
 Interposing Open GL and sending that across the wire 
Author Message
Certif-Eyed!
User avatar

Joined: Sat Sep 01, 2012 11:53 pm
Posts: 541
Reply with quote
You can do native OpenGL on android.

Why not just communicate all the opengl commands across the wire and re-render that on the android device. You might even be able to get away with that in wireless.

_________________
Gear VR: Maybe OVR isn't so evil after all!


Tue Nov 26, 2013 4:26 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
That would require significantly higher bandwidth, not to mention the poorer performance of a mobile GPU vs a proper gaming GPU on a PC.


Tue Nov 26, 2013 6:30 pm
Profile
3D Angel Eyes (Moderator)
User avatar

Joined: Sat Apr 12, 2008 8:18 pm
Posts: 11394
Reply with quote
Yeah, you would not gain much from doing this.

The OpenGL API calls themselves would be quick to send wirelessly, but then what about loading resources (like textures).

You could say the textures would be stored on the mobile device, but then what is the point of running the game on a PC?

_________________
check my blog - cybereality.com


Tue Nov 26, 2013 11:27 pm
Profile
Cross Eyed!

Joined: Sun Jun 02, 2013 4:21 pm
Posts: 175
Reply with quote
Give it a couple of years. The mobile GPUs of today are very capable. I think it's only a matter of time before being capable of rendering 2008-2010 era games natively.

And PC rendering will get obsolete anyway, once the bandwith issues are resolved. Nvidia GRID server farms doing all the work then.


Wed Nov 27, 2013 2:18 am
Profile
One Eyed Hopeful

Joined: Wed Mar 27, 2013 5:03 am
Posts: 39
Location: Germany
Reply with quote
Bandwidth is not the main Problem, latency is.


Thu Nov 28, 2013 9:21 am
Profile
Cross Eyed!

Joined: Sun Jun 02, 2013 4:21 pm
Posts: 175
Reply with quote
the_wretched wrote:
Bandwidth is not the main Problem, latency is.


Latency time = screen capture (memory copy) + RGB(A)-YUV conversion + encoding + packetization + transfer + decoding + rendering.

If you have solid GPUs on both side, the only thing remaining is the transfer part - bandwith. Which is why we are still using cables to deliver the video feed.


Thu Nov 28, 2013 11:19 am
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
the_wretched wrote:
Bandwidth is not the main Problem, latency is.


Well, if you're trying to shuffle large amounts of data around, then a decrease in bandwidth is an increase in latency.

Consider this:

A two channel Thunderbolt interface (AFAIK the fastest consumer-level serial port) has a bandwidth of 20 Gigabits/sec

On the othe hand, the PCIe 16X 2.0 interface used by most GPUs has a bandwidth of 8 Gigabytes/sec, or 64Gigabits/sec
A GPU using a PCIe 3.0 port has nearly double the amount of bandwidth as one attached to a 2.0 port. Either way, both are substantially higher than Thunderbolt, so I would say bandwidth is still very much a big problem.


Thu Nov 28, 2013 4:49 pm
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 7 posts ] 

Who is online

Users browsing this forum: Tbot and 5 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group
Designed by STSoftware.