GPU Computing with Apple
GPU computing has become ubiquitous in many areas, ranging from scientific computing and machine learning to games and many more. GPUs provide a parallel processing model that makes it possible to accelerate computations significantly and to handle large datasets efficiently, leading to faster and more powerful processing capabilities. With its GPU-aware array framework, high-level functions and powerful compiler, the Wolfram Language offers state-of-the-art functionality designed to leverage the capabilities of Apple GPUs. Directly manipulate data with GPUs and boost productivity using high-level linear algebra, statistics and mathematical functions without sacrificing performance.
GPUArray — array accessible to GPUs
Normal — convert to an ordinary array
Arrays
GPUArrayQ — test whether an object is a valid GPU array
ArrayQ ▪ VectorQ ▪ MatrixQ ▪ ArrayDepth ▪ Dimensions ▪ Length ▪ Part ▪ First ▪ Last ▪ ByteCount
Discrete Fourier Analysis »
Fourier — find the discrete Fourier transform
InverseFourier — find the discrete inverse Fourier transform
Random Number Generation »
SeedRandom — use "GPU" method for pseudorandom number generation
RandomInteger ▪ RandomReal ▪ RandomComplex
Statistics »
Mean — descriptive statistics
Variance ▪ StandardDeviation ▪ Kurtosis
Sort — order statistics
Mathematical Functions »
Plus, Times — arithmetic functions
Subtract ▪ Divide ▪ Total ▪ Mean
Sin, Cos — trigonometric functions
Tan ▪ ArcSin ▪ ArcCos ▪ ArcTan
Exp, Log — transcendental functions
Log2 ▪ Log10 ▪ Power ▪ Sqrt ▪ CubeRoot
Sinh, Cosh — hyperbolic functions
Tanh ▪ ArcSinh ▪ ArcCosh ▪ ArcTanh
Abs, Sign — absolute values and signs
Visualization »
ArrayPlot, MatrixPlot — plot values of an array
ListPlot — data visualization
ListLinePlot ▪ ListDensityPlot ▪ ListContourPlot ▪ ListLogPlot ▪ ListPolarPlot ▪ ParallelAxisPlot ▪ RadialAxisPlot
Code Compilation & Generation »
FunctionCompileExport — export code suitable for GPU architecture