Anyone using CUDA or CTM?

[CBONE]

Which would be better for Bioinformatics calculations? Are Multi-GPU setups automagically supported or do they need to be explicitly coded for?

Any help is appreciated!

[cadaveca]

Both have NDA, and are not available to public, yet.

[CBONE]

I've seen academic presentations where CUDA has been used and mentioned by name with an 8800. Is an academic release different than a public release?

[cadaveca]

Do YOU have the code? Can YOU go get it, right now, and provide a link?

Acemdemia is those who need such functionality. Stanford has a hand in developing CTM, from what I understand...however someone who needs these sort of tools will be the ones developing such, as only they know exactly what sort of functionality is needed.

CUDA has lots of sample code HERE (http://developer.download.nvidia.com/compute/cuda/sdk/website/samples.html), however, you'll find it's all pretty basic, as those that are developing software that uses this functionality look to charge mega dollars to corporations for such work.


The CUDA Toolkit is a complete software development solution for programming CUDA-enabled GPUs. The Toolkit includes standard FFT and BLAS libraries, a C-compiler for the NVIDIA GPU and a runtime driver. The CUDA runtime driver is a separate standalone driver that interoperates with OpenGL and Microsoft® DirectX® drivers from NVIDIA.

You'll find the SDK, no problem, but what you are asking for is like asking a game developer for thier code, when gaming first began...like asking Valve to give away the Source engine code before it was released.

CTM i have not seen public code for yet. Pretty sure this is still NDA-tight.



you need gpgpu.org, man.

[CBONE]

Do YOU have the code? Can YOU go get it, right now, and provide a link?

Acemdemia is those who need such functionality. Stanford has a hand in developing CTM, from what I understand...however someone who needs these sort of tools will be the ones developing such, as only they know exactly what sort of functionality is needed.

No, I can't. I was wondering if there are different NDAs out there. I figured that these professors wouldn't be able to blab about it in front of crowds of grad students and other professors if it was under NDA still. I'm wanting to find out if it would be suitable for DNA/Protein alignments and searches, but I don't know where to look for more info regarding that aspect. Everything out there is so general or graphics oriented that I can't tell.


CUDA has lots of sample code HERE (http://developer.download.nvidia.com/compute/cuda/sdk/website/samples.html), however, you'll find it's all pretty basic, as those that are developing software that uses this functionality look to charge mega dollars to corporations for such work.

You'll find the SDK, no problem, but what you are asking for is like asking a game developer for thier code, when gaming first began...like asking Valve to give away the Source engine code before it was released.

How do you figure? I'm just wondering which is better for a particular application, that's all. Same as if I asked what GPU is better for a game.



CTM i have not seen public code for yet. Pretty sure this is still NDA-tight.



you need gpgpu.org, man.

Thanks for the pointer. :toast:

[cadaveca]

The problem @ hand is that the apps themselves are still under development...remermber that ATi's card's featuring the greater extent of this functionality are not released yet, so hence the lack of info.

I don't see why you do not understand the monetary motovation for keeping such apps under cover...analysing the stocks, for example, and being able to predict trends, can make someone ALOT of money. ;) Do you expect them ot just throw that away?

Even FAH is still just using shaders...but you can bet they(Stanford) are working hard at making proper use of CTM and CUDA.


Also note that the driver for CUDA is not the same as 3D driver for G80...

[CBONE]

I don't see why you do not understand the monetary motovation for keeping such apps under cover...analysing the stocks, for example, and being able to predict trends, can make someone ALOT of money. ;) Do you expect them ot just throw that away?

Maybe we are getting stuck on wording; obviously the applications that end-users write wouldn't be available for free. I know that those apps could be worth money or be valuable for research. That's why I'm needing to find out which would be better suited for my purposes. (Research now, possibly money later :D ) I can't split resources or budget for two systems and take the time to write in two languages to see which is better. I was just seeing if anyone knew before I just pick one (probably CUDA) and put in orders.

[cadaveca]

Well, do you see anyone else posting?:stick: No mixing words here...