MATLAB: Bsxfun in GPU in PCT v5.2

gpuParallel Computing Toolbox

I am trying to do bsxfun(@minus,b,A*x); It is giving error as: Undefined function 'bsxfun' for input arguments of type 'parallel.gpu.GPUArray'. It seems this function is not supported. I am using Parallel Computing Toolbox Version 5.2 (R2011b). But this function is listed in the built in functions that support gpuArray.

http://www.mathworks.in/help/toolbox/distcomp/bsic4fr-1.html#bsloua3-1

Here b = 5 X 1; A*x would be 5 X 10. and since b is a column vector, the final output will be 5 X 10, [b – A*x];

 I want to do matrix column-wise subtraction.

Best Answer

The link you are referring to is actually from the current release, R2012a. bsxfun support for GPUArray data is a new feature in R2012a as you can check in the release notes for Parallel Computing Toolbox Version 6.0 (R2012a).

http://www.mathworks.es/help/toolbox/distcomp/rn/bs8h9g9-1.html#btaf2mp-1

I recommend you update to the latest release to benefit from all the recent GPUArray support functionality.

Related Solutions

MATLAB: Run custom function on GPU with Parallel Computing Toolbox

Arrayfun on the GPU only supports elementwise operations, so it is not the command that you want to use in this case.

You need to use min and * (mtimes) directly, which are overloaded to run on the GPU. If at least one of A, b, or x are stored on the GPU (i.e. of type parallel.gpu.GPUArray), then your code snippet will run on the GPU.

MATLAB: Do I receive an “cuDriverGetVersion” was not found error when calling the gpuArray function in Parallel Computnig Toolbox 5.0 (R2010b)

The error indicates that the version of NVIDIA CUDA library installed on your system is out of date. To resolve this issue:

1. Verify that the NVIDIA graphics card you are using has CUDA Compute Capability of 1.3 or higher. A list of CUDA enabled graphics cards and their associated Compute Capability can be found at the following links:

http://www.nvidia.com/object/cuda_gpus.html

http://developer.download.nvidia.com/compute/cuda/3_1/toolkit/docs/NVIDIA_CUDA_C_ProgrammingGuide_3.1.pdf

If your card is not listed in the resources above, or does not have the Compute Capability of 1.3 or higher it is not supported for GPU Computation in Parallel Computing Toolbox 5.0.

2. If your card is listed in one of the above mentioned links and has Compute Capability of 1.3 or higher update the graphics drivers. The latest version of NVIDIA drivers can be found at:

http://www.nvidia.com/Download/index.aspx?lang=en-us

3. Install the CUDA Toolkit. The CUDA Toolkit can be downloaded from:

http://developer.nvidia.com/object/cuda_3_1_downloads.html

4. After updating the drivers and installing the CUDA Toolkit start MATLAB.

5. At the MATLAB Command Prompt type:

>>paralleldemo_gpu_devices

This will test whether MATLAB recognizes your GPU. If everything went well you should see output similar to:

                 DeviceName: 'Tesla C1060'
                DeviceIndex: 1
          ComputeCapability: 1.3000
       DeviceSupportsDouble: 1
              DriverVersion: 3
         MaxThreadsPerBlock: 512
           MaxShmemPerBlock: 16384
         MaxThreadBlockSize: [512 512 64]
                MaxGridSize: [65535 65535]
                  SIMDWidth: 32
                TotalMemory: 4.2948e+09
                 FreeMemory: 4.2573e+09
        MultiprocessorCount: 30
       GPUOverlapsTransfers: 1
     KernelExecutionTimeout: 0
            DeviceSupported: 1

Best Answer

Related Solutions

MATLAB: Run custom function on GPU with Parallel Computing Toolbox

MATLAB: Do I receive an “cuDriverGetVersion” was not found error when calling the gpuArray function in Parallel Computnig Toolbox 5.0 (R2010b)

Related Question