Login

0 Latency CUDA Still Coming?

Tips & tricks, working results, technical support

0 Latency CUDA Still Coming?

Postby Puranon » Fri May 11, 2012 11:00 pm

Has there been any progress on a 0 latency CUDA powered Nebula Reverb?
Puranon
User Level X
User Level X
 
Posts: 118
Joined: Sun Apr 04, 2010 3:20 pm

Re: 0 Latency CUDA Still Coming?

Postby enriquesilveti » Sat May 12, 2012 12:37 pm

0 latency in a CPU > PCIe > CPU bridge is impossible. The best situation should be 1 sample latency at 192 kHz. You are talking about to run vectorial engine under CUDA. That is still under develop.
Enrique Silveti.
Acustica Audio customer and technical support.

MBP 11.5 (i7-4870 | 16 GB | 512 SDD)
SP4 (i5-6300 | 8 GB | 256 SDD)
UFX | Lyra2 | USBPre2
VM U15 | VM Win10 | VM OSX 10.12
N4/NAT4 | SPX3 | RX5 | LN2C | Smaart8 | R5 | PT12 | PX10 | NIK5
User avatar
enriquesilveti
Expert
Expert
 
Posts: 2663
Joined: Sun Mar 28, 2010 9:00 pm
Location: Lodi | Madrid | Buenos Aires

Re: 0 Latency CUDA Still Coming?

Postby Martinez » Sat May 12, 2012 11:44 pm

Cuda would be cool if it could be easy to set up.
At the moment it looks like a bit of a head ache to try.
Running a Nebula computer with a couple of tesla cards would be very cool though, can't imagine running out of processing power in a hurry.

Would even be cool to do some Aqura plug in eq's or comps to port over to UAD.
Martinez
Member
Member
 
Posts: 254
Joined: Mon Aug 09, 2010 2:03 pm

Re: 0 Latency CUDA Still Coming?

Postby enriquesilveti » Sun May 13, 2012 12:21 pm

martinez wrote:Cuda would be cool if it could be easy to set up. At the moment it looks like a bit of a head ache to try.

1. Install CUDA toolkit. EXACT VERSION!
2. Install AA CUDA bride.
Just to setups, no so difficult!

Running a Nebula computer with a couple of tesla cards would be very cool though, can't imagine running out of processing power in a hurry.

Giancarlo is running this setup from a lot of years.

Would even be cool to do some Aqura plug in eq's or comps to port over to UAD.


You can run FREQD engine under CUDA easily, just edit Acqua FX MAST page like in Nebula. About UAD AcquaFX/iDESK could be ported to UAD but we should solve the N2V issue, were we are going to load the vectors, does UAD has RAM? We should use a host RAM bridge, does UAD3 has RAM slots? PCIe DSP + RAM is not the same than use a PC in a GigaLAN?
Enrique Silveti.
Acustica Audio customer and technical support.

MBP 11.5 (i7-4870 | 16 GB | 512 SDD)
SP4 (i5-6300 | 8 GB | 256 SDD)
UFX | Lyra2 | USBPre2
VM U15 | VM Win10 | VM OSX 10.12
N4/NAT4 | SPX3 | RX5 | LN2C | Smaart8 | R5 | PT12 | PX10 | NIK5
User avatar
enriquesilveti
Expert
Expert
 
Posts: 2663
Joined: Sun Mar 28, 2010 9:00 pm
Location: Lodi | Madrid | Buenos Aires

Re: 0 Latency CUDA Still Coming?

Postby the19thbear » Mon May 14, 2012 9:12 pm

You should keep an eye out for soniccore. The hearth of their soundcards is tigersharc dsp's - lots of em. And ram
-With scope 6 (the OS for the cards) will come shortly and the platform will be completely open to develop on.
Running nebula on dsp's would be a great thing!
http://www.soniccore.com
the19thbear
User Level V
User Level V
 
Posts: 50
Joined: Thu Jul 01, 2010 12:26 am

Re: 0 Latency CUDA Still Coming?

Postby Puranon » Fri May 18, 2012 12:28 am

Cuda runs on Nvidia GPUs that are able to render incredibly powerful 3D graphics (polygons) with user feedback and only a split second delay.

I'm asking when will we see a Cuda with similar undetectable latency instead of these massive delay penalties we get now for using Nebula Reverb over the standard version.
Puranon
User Level X
User Level X
 
Posts: 118
Joined: Sun Apr 04, 2010 3:20 pm

Re: 0 Latency CUDA Still Coming?

Postby enriquesilveti » Wed May 23, 2012 9:41 am

the19thbear wrote:You should keep an eye out for soniccore. The hearth of their soundcards is tigersharc dsp's - lots of em. And ram
-With scope 6 (the OS for the cards) will come shortly and the platform will be completely open to develop on.
Running nebula on dsp's would be a great thing!
http://www.soniccore.com


Seems to UAD and SC goes to open platform the problem here is how to load N2V files plus how to translate the C++ code to DSP. We could try with an Acqua FX or iDESK before.
Enrique Silveti.
Acustica Audio customer and technical support.

MBP 11.5 (i7-4870 | 16 GB | 512 SDD)
SP4 (i5-6300 | 8 GB | 256 SDD)
UFX | Lyra2 | USBPre2
VM U15 | VM Win10 | VM OSX 10.12
N4/NAT4 | SPX3 | RX5 | LN2C | Smaart8 | R5 | PT12 | PX10 | NIK5
User avatar
enriquesilveti
Expert
Expert
 
Posts: 2663
Joined: Sun Mar 28, 2010 9:00 pm
Location: Lodi | Madrid | Buenos Aires

Re: 0 Latency CUDA Still Coming?

Postby enriquesilveti » Wed May 23, 2012 9:43 am

Puranon wrote:Cuda runs on Nvidia GPUs that are able to render incredibly powerful 3D graphics (polygons) with user feedback and only a split second delay.

I'm asking when will we see a Cuda with similar undetectable latency instead of these massive delay penalties we get now for using Nebula Reverb over the standard version.


Yes that could be done, we called Nebula pure CUDA. but still under develop.
Enrique Silveti.
Acustica Audio customer and technical support.

MBP 11.5 (i7-4870 | 16 GB | 512 SDD)
SP4 (i5-6300 | 8 GB | 256 SDD)
UFX | Lyra2 | USBPre2
VM U15 | VM Win10 | VM OSX 10.12
N4/NAT4 | SPX3 | RX5 | LN2C | Smaart8 | R5 | PT12 | PX10 | NIK5
User avatar
enriquesilveti
Expert
Expert
 
Posts: 2663
Joined: Sun Mar 28, 2010 9:00 pm
Location: Lodi | Madrid | Buenos Aires

Re: 0 Latency CUDA Still Coming?

Postby dwagrimm » Thu Jun 07, 2012 9:32 pm

Will be there a roadmap for Nebula pure CUDA?
dwagrimm
User Level IX
User Level IX
 
Posts: 90
Joined: Fri Jun 10, 2011 1:36 pm


Return to Working with Nebula

Who is online

Users browsing this forum: No registered users and 8 guests

cron