Cette page a été traduite par traduction automatique. Cliquez ici pour voir la dernière version en anglais.

Calcul GPU dans MATLAB

Accélérez votre code en exécutant les fonctions MATLAB^® sur un GPU

Si toutes les fonctions que vous souhaitez utiliser sont prises en charge sur le GPU, vous pouvez simplement utiliser gpuArray pour transférer les données d'entrée vers le GPU. Pour démarrer avec le calcul GPU, consultez Exécutez les fonctions MATLAB sur un GPU .

Pour le Deep Learning, MATLAB fournit une prise en charge parallèle automatique de plusieurs GPU. Voir Deep Learning with MATLAB on Multiple GPUs (Deep Learning Toolbox) .

Vous pouvez utiliser la fonction gpuDevice pour inspecter et sélectionner votre GPU et utiliser les fonctions gpuDeviceTable pour inspecter plusieurs GPU.

Fonctions

développer tout

Fonctions et classes clés

`gpuArray`	Tableau stocké sur GPU
`gather`	Transfer distributed array, `Composite` object, or `gpuArray` object to local workspace
`gpuDevice`	Interroger ou sélectionner un dispositif GPU
`GPUDeviceManager`	Manager for GPU Devices

Fonctions spécifiques au GPU

`gpuDeviceCount`	Number of GPU devices present
`gpuDeviceTable`	Table des propriétés des dispositifs GPU (depuis R2021a)
`wait (GPUDevice)`	Wait for GPU calculation to complete
`gputimeit`	Time required to run function on GPU
`existsOnGPU`	Determine if gpuArray or CUDAKernel is available on GPU
`isgpuarray`	Determine whether input is `gpuArray` (depuis R2020b)
`arrayfun`	Apply function to each element of array on GPU
`pagefun`	Apply function to each page of distributed or GPU array
`gpurng`	Control random number generation on the GPU
`reset`	Réinitialiser le dispositif GPU et effacer sa mémoire
`parallel.gpu.RandStream`	Random number stream on a GPU
`parallel.gpu.enableCUDAForwardCompatibility`	Query and set forward compatibility for GPU devices (depuis R2020b)

Rubriques

Exécutez le code MATLAB sur le GPU

Exécutez les fonctions MATLAB sur un GPU
Fournissez un argument gpuArray pour exécuter automatiquement des fonctions sur un GPU.
Identifier et sélectionner un dispositif GPU
Cet exemple montre comment utiliser gpuDevice pour identifier et sélectionner l'appareil que vous souhaitez utiliser.
Exigences de calcul GPU
Prise en charge des architectures GPU NVIDIA^®.
Establish Arrays on a GPU
Use gpuArray objects to store data on the GPU and perform calculation on the GPU.
Run MATLAB Functions on Multiple GPUs
This example shows how to run MATLAB® code on multiple GPUs in parallel, first on your local machine, then scaling up to a cluster.
Work with Remote GPUs
This example shows how to run MATLAB® code on multiple remote GPUs in a cluster. (depuis R2024a)
Run MATLAB using GPUs in the Cloud
Take advantage of powerful GPUs in the cloud to accelerate your code.
Train Network Using Automatic Multi-GPU Support (Deep Learning Toolbox)
This example shows how to use multiple GPUs on your local machine for deep learning training using automatic parallel support.
Work with Sparse Arrays on a GPU
Learn about the conditions for working with sparse arrays on a GPU.
Work with Complex Numbers on a GPU
Learn about the conditions for applying functions that might return complex values on a GPU.

Améliorer les performances sur le GPU

Measure and Improve GPU Performance
Time code running on your GPU and explore options for improving performance.
Improve Performance Using a GPU and Vectorized Calculations
This example shows you how to improve performance by running a function on the GPU instead of the CPU, and by vectorizing the calculations.
Improve Performance of Element-Wise MATLAB Functions on the GPU Using arrayfun
This example shows how to improve the performance of your code by running MATLAB® functions on the GPU using arrayfun.
Improve Performance of Small Matrix Problems on the GPU Using pagefun
This example shows how to use pagefun to improve the performance of independent operations applied to multiple matrices arranged in a multidimensional array.
Benchmarking A\b on the GPU
This example looks at how we can benchmark the solving of a linear system on the GPU.

Apprendre encore plus

Profile Your Code to Improve Performance
Use the Profiler to measure the time it takes to run your code and identify which lines of code consume the most time or which lines do not run.
Vectorization
Revise loop-based, scalar-oriented code to use MATLAB matrix and vector operations.
Random Number Streams on a GPU
Control the random number streams on a GPU to generate the same sequences of random numbers as on the CPU.
Generating Random Numbers on a GPU
This example shows how to switch between the different random number generators that are supported on the GPU.
Stencil Operations on a GPU
This example uses Conway's "Game of Life" to demonstrate how stencil operations can be performed using a GPU.
Compute the Mandelbrot Set Using GPU-Enabled Functions
This example shows how to use GPU-enabled MATLAB® functions to compute a well-known mathematical construction: the Mandelbrot set.

Informations connexes

Configuration système requise pour le GPU

Exemples présentés

Illustrating Three Approaches to GPU Computing: The Mandelbrot Set

Adapt your MATLAB® code to compute the Mandelbrot Set using a GPU.

Ouvrir le live script

Measure GPU Performance

Measure some of the key performance characteristics of a GPU.

Ouvrir le live script

Using GPU arrayfun for Monte-Carlo Simulations

Calculate prices for financial options on a GPU using Monte-Carlo methods.

Ouvrir le live script

$Use FFT2 on the GPU to Simulate Diffraction Patterns$

Use FFT2 on the GPU to Simulate Diffraction Patterns

Uses Parallel Computing Toolbox™ to perform a two-dimensional Fast Fourier Transform (FFT) on a GPU. The two-dimensional Fourier transform is used in optics to calculate far-field diffraction patterns. These diffraction patterns are observed when a monochromatic light source passes through a small aperture, such as in Young's double-slit experiment.

Ouvrir le live script

Affiner une image à l'aide du GPU

Cet exemple montre comment affiner une image à l'aide d'objets gpuArray et de fonctions compatibles GPU.

Ouvrir le live script