OpenCL Tegra WebCL WebGL
I would like to ask whether there are any plans to support OpenCL on Tegra processors. Right now the main burden are on its 4 CPUs, in the future Tegra may make its GPU more powerfull. People for instance have started talking about "smart" TVs. Well imagine a TV with a Tegra 4 processor, the OS is to the vendor and on top of it implementing HTML 5. Games, movies, multimedia content will become a reality. And here comes WebGL and WebCL...A company would surely want to create a game that would run almost anywhere...
I would like to ask whether there are any plans to support OpenCL on Tegra processors. Right now the main burden are on its 4 CPUs, in the future Tegra may make its GPU more powerfull. People for instance have started talking about "smart" TVs. Well imagine a TV with a Tegra 4 processor, the OS is to the vendor and on top of it implementing HTML 5. Games, movies, multimedia content will become a reality. And here comes WebGL and WebCL...A company would surely want to create a game that would run almost anywhere...

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#1
Posted 05/03/2012 09:07 PM   
Great question. I'd like to learn more about this too. Especially interesting is Neil Trevett's talk that [url="http://www.streamcomputing.eu/blog/2012-04-21/neil-trevett-on-opencl/"]StreamComputing posted[/url].

[quote]
At 45:10 he puts NVidia’s plans on HPC, before getting back to : “NVidia is going to support both [CUDA and OpenCL] in HPC. In Mobile it’s going to be all OpenCL“.

At 45:23 he repeats his statements: “In the mobile space I expect OpenCL to be the primary tool“.
[/quote]

Anyone at NVIDIA want to comment?
Great question. I'd like to learn more about this too. Especially interesting is Neil Trevett's talk that StreamComputing posted.





At 45:10 he puts NVidia’s plans on HPC, before getting back to : “NVidia is going to support both [CUDA and OpenCL] in HPC. In Mobile it’s going to be all OpenCL“.



At 45:23 he repeats his statements: “In the mobile space I expect OpenCL to be the primary tool“.





Anyone at NVIDIA want to comment?

John Melonakos ([email="john.melonakos@accelereyes.com"]john.melonakos@accelereyes.com[/email])

#2
Posted 05/03/2012 09:13 PM   
[quote name='AlexanderAgathos' date='04 May 2012 - 12:07 AM' timestamp='1336079260' post='1403950']
I would like to ask whether there are any plans to support OpenCL on Tegra processors. Right now the main burden are on its 4 CPUs, in the future Tegra may make its GPU more powerfull. People for instance have started talking about "smart" TVs. Well imagine a TV with a Tegra 4 processor, the OS is to the vendor and on top of it implementing HTML 5. Games, movies, multimedia content will become a reality. And here comes WebGL and WebCL...A company would surely want to create a game that would run almost anywhere...
[/quote]

One more question. The new OpenCL 1.2 offers Direct3D 11 interoperability with OpenCL not to say that someone can divide the device in subdevices which is very important. People want to create games/multimedia for both vendors Nvidia/AMD. When do you think you will have support OpenCL 1.2?

[quote name='AlexanderAgathos' date='04 May 2012 - 12:07 AM' timestamp='1336079260' post='1403950']

I would like to ask whether there are any plans to support OpenCL on Tegra processors. Right now the main burden are on its 4 CPUs, in the future Tegra may make its GPU more powerfull. People for instance have started talking about "smart" TVs. Well imagine a TV with a Tegra 4 processor, the OS is to the vendor and on top of it implementing HTML 5. Games, movies, multimedia content will become a reality. And here comes WebGL and WebCL...A company would surely want to create a game that would run almost anywhere...





One more question. The new OpenCL 1.2 offers Direct3D 11 interoperability with OpenCL not to say that someone can divide the device in subdevices which is very important. People want to create games/multimedia for both vendors Nvidia/AMD. When do you think you will have support OpenCL 1.2?


Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#3
Posted 05/03/2012 11:26 PM   
[quote name='AlexanderAgathos' date='04 May 2012 - 02:26 AM' timestamp='1336087600' post='1404007']
One more question. The new OpenCL 1.2 offers Direct3D 11 interoperability with OpenCL not to say that someone can divide the device in subdevices which is very important. People want to create games/multimedia for both vendors Nvidia/AMD. When do you think you will have support OpenCL 1.2?


[/quote]

I understand that what I propose is one version of the technology branch. If I was a director or a proposal tech in a technology branch in a company I would like a reassurance. Of course I know that in major companies the R&D branch should be able to proceed with Research concerning theses issues (like Samsung is already doing). So really I am not upset receiving no answer I was just hoping I would.
[quote name='AlexanderAgathos' date='04 May 2012 - 02:26 AM' timestamp='1336087600' post='1404007']

One more question. The new OpenCL 1.2 offers Direct3D 11 interoperability with OpenCL not to say that someone can divide the device in subdevices which is very important. People want to create games/multimedia for both vendors Nvidia/AMD. When do you think you will have support OpenCL 1.2?









I understand that what I propose is one version of the technology branch. If I was a director or a proposal tech in a technology branch in a company I would like a reassurance. Of course I know that in major companies the R&D branch should be able to proceed with Research concerning theses issues (like Samsung is already doing). So really I am not upset receiving no answer I was just hoping I would.

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#4
Posted 05/04/2012 10:05 AM   
I would like to ignite also imagination since the new WOA (Windows on Arm) is approaching. I do not link companies often but this is a must link from ASUS

[media]http://www.youtube.com/watch?feature=player_embedded&v=jIK51LRLud4[/media]


So OpenCL could be very useful here to code a program targeted for a desktop and a tablet. Now that Tegra 4 is coming out it will be a lot of fuss. The device on the Desktop is the GPU mostly and parts on the CPU but on a Tegra the devices will be the 4-core CPU. So you can imagine how usefull will be OpenCL for Tegra...
Microsoft is saying that there will be a showcase for future ARM PCs I am showing you this video to understand what the future will be...

Cheers,
Alexander.
I would like to ignite also imagination since the new WOA (Windows on Arm) is approaching. I do not link companies often but this is a must link from ASUS



[media]http://www.youtube.com/watch?feature=player_embedded&v=jIK51LRLud4[/media]





So OpenCL could be very useful here to code a program targeted for a desktop and a tablet. Now that Tegra 4 is coming out it will be a lot of fuss. The device on the Desktop is the GPU mostly and parts on the CPU but on a Tegra the devices will be the 4-core CPU. So you can imagine how usefull will be OpenCL for Tegra...

Microsoft is saying that there will be a showcase for future ARM PCs I am showing you this video to understand what the future will be...



Cheers,

Alexander.

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#5
Posted 05/11/2012 01:57 AM   
I don't think you can get any information here because no staff from Nvidia cares about this forum. Just look at the [url="http://developer.nvidia.com/opencl"]OpenCL page of Nvidia' website[/url]. Can you believe they have removed all the old contents include some important programming guide documents (can still be found in GPU Computing SDK) and only some meaningless introduction texts left. Lets pray they won't abandon the OpenCL support in the future.
I don't think you can get any information here because no staff from Nvidia cares about this forum. Just look at the OpenCL page of Nvidia' website. Can you believe they have removed all the old contents include some important programming guide documents (can still be found in GPU Computing SDK) and only some meaningless introduction texts left. Lets pray they won't abandon the OpenCL support in the future.

#6
Posted 05/11/2012 05:04 AM   
I am certain that they will start thinking about OpenCL in the portable solutions...it is a matter of time...Tegra-4 will come right on time with WOA. Anyway I will have to wait until 2013. Plenty of other things to look till then. :-)




I am certain that they will start thinking about OpenCL in the portable solutions...it is a matter of time...Tegra-4 will come right on time with WOA. Anyway I will have to wait until 2013. Plenty of other things to look till then. :-)








Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#7
Posted 05/11/2012 06:03 AM   
NVIDIA is trying to postpone proper OpenCL support for as long as possible. Luckily, all the competition is pushing real hard, and more and more softwares come built on top of it. Even in my institute people starting to learn GPU programming tend to shift towards OpenCL, rather than CUDA.

Maybe sometime NV will realise that going against the wind is counterproductive.
NVIDIA is trying to postpone proper OpenCL support for as long as possible. Luckily, all the competition is pushing real hard, and more and more softwares come built on top of it. Even in my institute people starting to learn GPU programming tend to shift towards OpenCL, rather than CUDA.



Maybe sometime NV will realise that going against the wind is counterproductive.

#8
Posted 05/11/2012 06:41 AM   
[quote name='Meteorhead' date='11 May 2012 - 09:41 AM' timestamp='1336718488' post='1406852']
NVIDIA is trying to postpone proper OpenCL support for as long as possible. Luckily, all the competition is pushing real hard, and more and more softwares come built on top of it. Even in my institute people starting to learn GPU programming tend to shift towards OpenCL, rather than CUDA.

Maybe sometime NV will realise that going against the wind is counterproductive.
[/quote]

Maybe when they see 3DStudio and Maya Rendering in OpenCL with several frames per second they will get the message...
[quote name='Meteorhead' date='11 May 2012 - 09:41 AM' timestamp='1336718488' post='1406852']

NVIDIA is trying to postpone proper OpenCL support for as long as possible. Luckily, all the competition is pushing real hard, and more and more softwares come built on top of it. Even in my institute people starting to learn GPU programming tend to shift towards OpenCL, rather than CUDA.



Maybe sometime NV will realise that going against the wind is counterproductive.





Maybe when they see 3DStudio and Maya Rendering in OpenCL with several frames per second they will get the message...

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#9
Posted 05/11/2012 12:06 PM   
[quote name='AlexanderAgathos' date='11 May 2012 - 08:06 PM' timestamp='1336737992' post='1406938']
Maybe when they see 3DStudio and Maya Rendering in OpenCL with several frames per second they will get the message...
[/quote]

Another problem I concern is Nvidia's product strategy. They seem to focus on the game performance instead of computation on the Kepler GeForce cards. It is terrible if they want to force us to buy the expensive Quadro or Tesla cards for good CUDA or OpenCL performance.
[quote name='AlexanderAgathos' date='11 May 2012 - 08:06 PM' timestamp='1336737992' post='1406938']

Maybe when they see 3DStudio and Maya Rendering in OpenCL with several frames per second they will get the message...





Another problem I concern is Nvidia's product strategy. They seem to focus on the game performance instead of computation on the Kepler GeForce cards. It is terrible if they want to force us to buy the expensive Quadro or Tesla cards for good CUDA or OpenCL performance.

#10
Posted 05/12/2012 02:47 AM   
[quote name='ZHAO Peng' date='12 May 2012 - 05:47 AM' timestamp='1336790832' post='1407283']
Another problem I concern is Nvidia's product strategy. They seem to focus on the game performance instead of computation on the Kepler GeForce cards. It is terrible if they want to force us to buy the expensive Quadro or Tesla cards for good CUDA or OpenCL performance.
[/quote]

Yes I have seen in tests that AMD's graphics card outperform significantly NVIDIA's consumer graphics cards in OpenCL...you are right. They have to look into this. I will update my version of Apple's OpenCL examples for AMD's GPU. It is not fair having only NVIDIA's. It is very easy to do this.
[quote name='ZHAO Peng' date='12 May 2012 - 05:47 AM' timestamp='1336790832' post='1407283']

Another problem I concern is Nvidia's product strategy. They seem to focus on the game performance instead of computation on the Kepler GeForce cards. It is terrible if they want to force us to buy the expensive Quadro or Tesla cards for good CUDA or OpenCL performance.





Yes I have seen in tests that AMD's graphics card outperform significantly NVIDIA's consumer graphics cards in OpenCL...you are right. They have to look into this. I will update my version of Apple's OpenCL examples for AMD's GPU. It is not fair having only NVIDIA's. It is very easy to do this.

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#11
Posted 05/12/2012 03:09 AM   
Quoting... It was Freescale than won the new-new BMW deal for next-gen hardware because of one small thing - Vivante's GPU featureset. Vivante is not a new player in the field, but the low profile firm has a GPU which supports OpenCL, which was listed on BMW's corporate "RFQ" (Request For Quote). According to our contacts close to heart of the matter, OpenCL was the reason why Nvidia "is out". ---------------------- I hate it when I am right sometimes. Especially this time.
Quoting...
It was Freescale than won the new-new BMW deal for next-gen hardware because of one small thing - Vivante's GPU featureset.

Vivante is not a new player in the field, but the low profile firm has a GPU which supports OpenCL, which was listed on BMW's corporate "RFQ" (Request For Quote). According to our contacts close to heart of the matter, OpenCL was the reason why Nvidia "is out".
----------------------
I hate it when I am right sometimes. Especially this time.

Dr. Alexander Agathos Degree in Maths, PhD in Computer Science (specialization in Computer Graphics)

Researcher in HPC, Department of Mathematics and Informatics, West University of Timisoara, Romania.

[email="agalex@iit.demokritos.gr"]Email[/email]

Website

--------------------
A famous researcher in parallelism said : Parallelism is a great problem. The basic question is though are you part of the solution or part of the problem?

The challenges of parallel computing are simple to enumerate : expose lots of parallelism and minimize communication (said another way : maximize locality)

The more complicated you present a subject the less you have really mastered it...Albert Einstein.

#12
Posted 01/28/2013 05:13 PM   
Scroll To Top