CUDALink`
CUDALink`

CUDAMatrix

CUDAMatrix[data]

yields a matrix of data which resides on a CUDA enabled GPU.

CUDAMatrix[data,type]

yields a matrix of the specified type.

Details and Options

  • The CUDALink application must be loaded using Needs["CUDALink`"].
  • The resources used by a CUDAMatrix are automatically reclaimed from the GPU when they are no longer used.
  • The data in a CUDAMatrix can be retrieved by Normal.
  • A number of functions accept and return CUDAMatrix objects allowing operations to work with data that stays on the GPU.
  • Possible types for CUDAVector are:
  • IntegerRealComplex
    "Integer32""Real32""ComplexReal32"
    "Integer64""Real64""ComplexReal64"

Examples

open allclose all

Basic Examples  (1)

First, load the CUDALink application:

Create a matrix of data that lives on the GPU:

Compute a dot product on the GPU (the result also resides on the GPU):

Retrieve the data from a CUDAMatrix:

Scope  (1)

Information  (1)

Information about the CUDA matrix:

Possible Issues  (1)

Normal of a CUDAMatrix is a NumericArray:

To get the underlying data, Normal is applied twice:

Wolfram Research (2022), CUDAMatrix, Wolfram Language function, https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html.

Text

Wolfram Research (2022), CUDAMatrix, Wolfram Language function, https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html.

CMS

Wolfram Language. 2022. "CUDAMatrix." Wolfram Language & System Documentation Center. Wolfram Research. https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html.

APA

Wolfram Language. (2022). CUDAMatrix. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html

BibTeX

@misc{reference.wolfram_2023_cudamatrix, author="Wolfram Research", title="{CUDAMatrix}", year="2022", howpublished="\url{https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html}", note=[Accessed: 27-April-2024 ]}

BibLaTeX

@online{reference.wolfram_2023_cudamatrix, organization={Wolfram Research}, title={CUDAMatrix}, year={2022}, url={https://reference.wolfram.com/language/CUDALink/ref/CUDAMatrix.html}, note=[Accessed: 27-April-2024 ]}