WOLFRAM

Wolfram Language & System Documentation Center

CUDASparseMatrix[sparse,type]

yields a sparse matrix of given type which resides on a CUDA enabled GPU.

Details and Options
Details and Options Details and Options
Examples  
Basic Examples  
Scope  
Information  
Possible Issues  
See Also
Tech Notes
Related Guides
Cite this Page
CUDALink`
CUDALink`

CUDASparseMatrix

CUDASparseMatrix[sparse,type]

yields a sparse matrix of given type which resides on a CUDA enabled GPU.

Details and Options

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

Examples

open all close all

Basic Examples  (1)

First, load the CUDALink application:

Wolfram Language code: Needs["CUDALink`"]

Create a sparse vector of data that lives on the GPU:

Wolfram Language code: mat = CUDASparseMatrix[ SparseArray[Table[{i, i} -> i, {i, 4}]], "Real32"]

Retrieve the data from a CUDASparseMatrix:

Wolfram Language code: Normal[mat]

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

Wolfram Language code: CUDADot[mat, mat]

Scope  (1)

Information  (1)

Information about the CUDA sparse matrix:

Wolfram Language code: Information[CUDASparseMatrix[SparseArray[Table[{i, i} -> 1., {i, 4}]]]]

Possible Issues  (1)

Normal of a CUDASparseMatrix is a SparseArray:

Wolfram Language code: cuspmat = CUDASparseMatrix[SparseArray[Table[{i, i} -> 1., {i, 4}]]]; Normal[cuspmat]

To get the underlying data, Normal is applied twice:

Wolfram Language code: Normal[Normal[cuspmat]]
Wolfram Research (2022), CUDASparseMatrix, Wolfram Language function, https://reference.wolfram.com/language/CUDALink/ref/CUDASparseMatrix.html.

Text

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

CMS

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

APA

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

BibTeX

@misc{reference.wolfram_2026_cudasparsematrix, author="Wolfram Research", title="{CUDASparseMatrix}", year="2022", howpublished="\url{https://reference.wolfram.com/language/CUDALink/ref/CUDASparseMatrix.html}", note=[Accessed: 15-June-2026]}

BibLaTeX

@online{reference.wolfram_2026_cudasparsematrix, organization={Wolfram Research}, title={CUDASparseMatrix}, year={2022}, url={https://reference.wolfram.com/language/CUDALink/ref/CUDASparseMatrix.html}, note=[Accessed: 15-June-2026]}