Do I need MatLab's parallel computing toolbox to run MatLAb on NVIDIA?

[tamfatkh]

Hi
I need to know the answer for these question
Do I need MatLab’s parallel computing toolbox to run MatLAb on NVIDIA?
Is the accelerator enough and efficient?

[jack]

Hi

I need to know the answer for these question

Do I need MatLab’s parallel computing toolbox to run MatLAb on NVIDIA?

Is the accelerator enough and efficient?

As far as I know, the parallel computing toolbox is for multi-core processors and distributed computing across a network, not to build stuff for CUDA. If you want to write .mex files for CUDA, there is a special compiler included in the toolkit (nvmex). Or, you can try the Accelereyes’ Jacket project, which handles the GPU computations for you (so you just write normal matlab code, and it’s translated behind the scenes).

[Sarnath]

As far as I know, the parallel computing toolbox is for multi-core processors and distributed computing across a network, not to build stuff for CUDA. If you want to write .mex files for CUDA, there is a special compiler included in the toolkit (nvmex). Or, you can try the Accelereyes’ Jacket project, which handles the GPU computations for you (so you just write normal matlab code, and it’s translated behind the scenes).

www.accelereyes.com

People from accelereyes do participate in this forum (jmelonkas? or some1). May b, they should be able to give you a better picture on this.