MATLAB: GPU gather command clarification and GPU memory usage

gpu memory gpuarray

Dear All,

1. If x is a gpuarray, and y its cpu correspondent, after the command

y=gather(x)

is x still on the GPU or is cleared (by being gathered on the cpu)?

2. Also, is there a way to see memory usage just on GPU at a certain time point, and not combined with the cpu memory usage?

3. Can x be saved on the hard drive without first being passed as y on the cpu?

Thank you,

Octavian.

Best Answer

In this case, x still exists on the GPU - in general, input arguments to functions are not modified in MATLAB (some exceptions exist where the objects are handle types).
Yes, look at the AvailableMemory field of the GPUDevice object returned by the function gpuDevice. See the documentation for more about this object
Yes.

Related Solutions

MATLAB: Is NVIDIA Quadro P500 good enough to use parallel computing toolbox

Yes, the Quadro P500 is only about a year old and should work well. (It is CUDA Capacity 6.1)

MATLAB: Parfor on GPU

Yes, you can do this. Whether you get much benefit depends on whether you have multiple GPUs in your system (under some circumstances, a single GPU might actually suffice if you have enough CPU work to keep things busy).

You might wish to do something like

spmd
  gpuDevice( 1 + mod( labindex - 1, gpuDeviceCount ) )
end

before you go any further (if you have multiple supported GPUs)

After that, gpuArrays can be passed into and out from PARFOR loops with no further modification. The following example shows this - but note that this is a proof of concept - it performs very badly because you're operating on scalar elements of the gpuArray.

g = gpuArray(1:10);
parfor ii=1:numel(g)
  x(ii) = 1/g(ii);
end

Best Answer

Related Solutions

MATLAB: Is NVIDIA Quadro P500 good enough to use parallel computing toolbox

MATLAB: Parfor on GPU

Related Question