Technologies›Graphics, Imaging & Compositing

MetalPerformanceShaders

iOSmacOStvOSwatchOSvisionOS

MetalPerformanceShaders provides GPU-accelerated Metal kernels for image processing, linear algebra, and neural-network primitives, optimized per device. Image work centers on MPSImage and MPSImageDescriptor, with types such as MPSImageEdgeMode, MPSImageFeatureChannelFormat, MPSRegion, and MPSScaleTransform describing pixel data and transforms, while MPSMatrix, MPSVector, and their descriptors support linear-algebra operations. You build kernels on MPSKernel, configure them with values like MPSKernelOptions and MPSDataType, and encode them through MPSCommandBuffer, using MPSState and MPSTemporaryImage to manage transient GPU resources. Protocols including MPSDeviceProvider, MPSImageAllocator, and MPSHeapProvider let you supply the device and allocation strategy these kernels run against on iOS, macOS, tvOS, and visionOS.

Core Kernels and Resources 8

Foundational kernel base classes and the GPU resources, state objects, and command buffer used to run every shader.

  • Cl
    MPSKerneliOS 9.0+
    @class MPSKernel
  • Cl
    MPSCommandBufferiOS 13.0+
    @class MPSCommandBuffer
  • Cl
    MPSStateiOS 11.0+
    @class MPSState
  • Cl
    MPSStateResourceListiOS 11.3+
    An object that collects the Metal resources backing an MPSState instance.
  • Cl
    MPSPredicateiOS 13.0+
    @class MPSPredicate
  • Cl
    MPSKeyedUnarchiveriOS 11.3+
    @class MPSKeyedUnarchiver
  • Cl
    MPSUnaryImageKerneliOS 9.0+
    @class MPSUnaryImageKernel
  • Cl
    MPSBinaryImageKerneliOS 9.0+
    @class MPSBinaryImageKernel

Images 3

Image objects and descriptors that hold pixel data on the GPU, including transient resources.

  • Cl
    MPSImageiOS 10.0+
    @class MPSImage
  • Cl
    MPSImageDescriptoriOS 10.0+
    @class MPSImageDescriptor
  • Cl
    MPSTemporaryImageiOS 10.0+
    @class MPSTemporaryImage

Matrices, Vectors, and Arrays 9

Linear-algebra data types and their descriptors for matrices, vectors, and multidimensional arrays.

  • Cl
    MPSMatrixiOS 10.0+
    @class MPSMatrix
  • Cl
    MPSMatrixDescriptoriOS 10.0+
    @class MPSMatrixDescriptor
  • Cl
    MPSTemporaryMatrixiOS 11.0+
    @abstract A MPSMatrix allocated on GPU private memory.
  • Cl
    MPSVectoriOS 11.0+
    @class MPSVector
  • Cl
    MPSVectorDescriptoriOS 11.0+
    @class MPSVectorDescriptor
  • Cl
    MPSTemporaryVectoriOS 11.0+
    @abstract A MPSVector allocated on GPU private memory.
  • Cl
    MPSNDArrayiOS 13.0+
    @class MPSNDArray
  • Cl
    MPSNDArrayDescriptoriOS 13.0+
    @class MPSNDArrayDescriptor
  • Cl
    MPSTemporaryNDArrayiOS 13.0+
    @class MPSTemporaryNDArray

Image Filters and Convolutions 17

Convolution, blur, edge-detection, and morphology kernels that filter image data.

  • Cl
    MPSImageConversioniOS 10.0+
    @class MPSImageConversion
  • Cl
    MPSImageConvolutioniOS 9.0+
    @class MPSImageConvolution
  • Cl
    MPSImageLaplacianiOS 10.0+
    @class MPSImageLaplacian
  • Cl
    MPSImageBoxiOS 9.0+
    @class MPSImageBox
  • Cl
    MPSImageTentiOS 9.0+
    @class MPSImageTent
  • Cl
    MPSImageGaussianBluriOS 9.0+
    @class MPSImageGaussianBlur
  • Cl
    MPSImageSobeliOS 9.0+
    @class MPSImageSobel
  • Cl
    MPSImageCannyiOS 14.0+
    @class MPSImageCanny
  • Cl
    MPSImageMedianiOS 9.0+
    @class MPSImageMedian
  • Cl
    MPSImageAreaMaxiOS 9.0+
    @class MPSImageAreaMax
  • Cl
    MPSImageAreaMiniOS 9.0+
    @class MPSImageAreaMin
  • Cl
    MPSImageDilateiOS 9.0+
    @class MPSImageDilate
  • Cl
    MPSImageErodeiOS 9.0+
    @class MPSImageErode
  • Cl
    MPSImageGuidedFilteriOS 11.3+
    @class MPSImageGuidedFilter
  • Cl
    MPSImageEDLinesiOS 13.4+
    @class MPSImageEDLines
  • Cl
    MPSImageEuclideanDistanceTransformiOS 11.3+
    @class MPSImageEuclideanDistanceTransform
  • Cl
    MPSImageFindKeypointsiOS 11.0+
    @class MPSImageFindKeypoints

Image Pyramids 5

Multi-resolution pyramid kernels for building and combining image scale levels.

  • Cl
    MPSImagePyramidiOS 10.0+
    @class MPSImagePyramid
  • Cl
    MPSImageGaussianPyramidiOS 10.0+
    @class MPSImageGaussianPyramid
  • Cl
    MPSImageLaplacianPyramidiOS 10.0+
    @class MPSImageLaplacianPyramid
  • Cl
    MPSImageLaplacianPyramidSubtractiOS 10.0+
    A kernel that forms Laplacian pyramid levels by subtracting upsampled scales.
  • Cl
    MPSImageLaplacianPyramidAddiOS 10.0+
    A kernel that reconstructs an image by adding Laplacian pyramid levels.

Image Arithmetic and Scaling 9

Per-pixel arithmetic, transpose, and resampling kernels that combine or resize images.

  • Cl
    MPSImageArithmeticiOS 11.0+
    @class MPSImageArithmetic
  • Cl
    MPSImageAddiOS 11.0+
    @class MPSImageAdd
  • Cl
    MPSImageSubtractiOS 11.0+
    @class MPSImageSubtract
  • Cl
    MPSImageMultiplyiOS 11.0+
    @class MPSImageMultiply
  • Cl
    MPSImageDivideiOS 11.0+
    @class MPSImageDivide
  • Cl
    MPSImageScaleiOS 11.0+
    @class MPSImageScale
  • Cl
    MPSImageLanczosScaleiOS 9.0+
    @class MPSImageLanczosScale
  • Cl
    MPSImageBilinearScaleiOS 11.0+
    @class MPSImageBilinearScale
  • Cl
    MPSImageTransposeiOS 9.0+
    @class MPSImageTranspose

Image Histograms and Statistics 9

Kernels that compute histograms, integrals, and statistical measures over image regions.

  • Cl
    MPSImageHistogramiOS 9.0+
    @class MPSImageHistogram
  • Cl
    MPSImageNormalizedHistogramiOS 9.0+
    @class MPSImageNormalizedHistogram
  • Cl
    MPSImageHistogramEqualizationiOS 9.0+
    @class MPSImageHistogramEqualization
  • Cl
    MPSImageHistogramSpecificationiOS 9.0+
    @class MPSImageHistogramSpecification
  • Cl
    MPSImageIntegraliOS 9.0+
    @class MPSImageIntegral
  • Cl
    MPSImageIntegralOfSquaresiOS 9.0+
    @class MPSImageIntegralOfSquares
  • Cl
    MPSImageStatisticsMinAndMaxiOS 11.0+
    @class MPSImageStatisticsMinAndMax
  • Cl
    MPSImageStatisticsMeanAndVarianceiOS 11.0+
    @class MPSImageStatisticsMeanAndVariance
  • Cl
    MPSImageStatisticsMeaniOS 11.0+
    @class MPSImageStatisticsMean

Image Reductions and Thresholds 14

Row and column reductions plus thresholding kernels that condense or binarize image data.

  • Cl
    MPSImageReduceUnaryiOS 11.3+
    @class MPSImageReduceUnary
  • Cl
    MPSImageReduceRowMiniOS 11.3+
    @class MPSImageReduceRowMin
  • Cl
    MPSImageReduceColumnMiniOS 11.3+
    @class MPSImageReduceColumnMin
  • Cl
    MPSImageReduceRowMaxiOS 11.3+
    @class MPSImageReduceRowMax
  • Cl
    MPSImageReduceColumnMaxiOS 11.3+
    @class MPSImageReduceColumnMax
  • Cl
    MPSImageReduceRowMeaniOS 11.3+
    @class MPSImageReduceRowMean
  • Cl
    MPSImageReduceColumnMeaniOS 11.3+
    @class MPSImageReduceColumnMean
  • Cl
    MPSImageReduceRowSumiOS 11.3+
    @class MPSImageReduceRowSum
  • Cl
    MPSImageReduceColumnSumiOS 11.3+
    @class MPSImageReduceColumnSum
  • Cl
    MPSImageThresholdBinaryiOS 9.0+
    @class MPSImageThresholdBinary
  • Cl
    MPSImageThresholdBinaryInverseiOS 9.0+
    @class MPSImageThresholdBinaryInverse
  • Cl
    MPSImageThresholdTruncateiOS 9.0+
    @class MPSImageThresholdTruncate
  • Cl
    MPSImageThresholdToZeroiOS 9.0+
    @class MPSImageThresholdToZero
  • Cl
    MPSImageThresholdToZeroInverseiOS 9.0+
    @class MPSImageThresholdToZeroInverse

Matrix Operations 18

Kernels for matrix and vector products, solvers, decompositions, softmax, and copies.

  • Cl
    MPSMatrixUnaryKerneliOS 11.0+
    @class MPSMatrixUnaryKernel
  • Cl
    MPSMatrixBinaryKerneliOS 11.0+
    @class MPSMatrixBinaryKernel
  • Cl
    MPSMatrixMultiplicationiOS 10.0+
    @class MPSMatrixMultiplication
  • Cl
    MPSMatrixVectorMultiplicationiOS 11.0+
    @class MPSMatrixVectorMultiplication
  • Cl
    MPSMatrixSolveTriangulariOS 11.0+
    @class MPSMatrixSolveTriangular
  • Cl
    MPSMatrixSolveLUiOS 11.0+
    @class MPSMatrixSolveLU
  • Cl
    MPSMatrixSolveCholeskyiOS 11.0+
    @class MPSMatrixSolveCholesky
  • Cl
    MPSMatrixDecompositionLUiOS 11.0+
    @class MPSMatrixDecompositionLU
  • Cl
    MPSMatrixDecompositionCholeskyiOS 11.0+
    @class MPSMatrixDecompositionCholesky
  • Cl
    MPSMatrixCopyiOS 11.0+
    A kernel that copies values between matrices, with optional transpose and offset.
  • Cl
    MPSMatrixCopyDescriptoriOS 11.0+
    @abstract A list of copy operations
  • Cl
    MPSMatrixSoftMaxiOS 11.0+
    @class MPSMatrixSoftMax
  • Cl
    MPSMatrixLogSoftMaxiOS 11.0+
    @class MPSMatrixLogSoftMax
  • Cl
    MPSMatrixSoftMaxGradientiOS 12.0+
    @class MPSMatrixSoftMaxGradient
  • Cl
    MPSMatrixLogSoftMaxGradientiOS 12.0+
    @class MPSMatrixLogSoftMaxGradient
  • Cl
    MPSMatrixFindTopKiOS 11.3+
    @class MPSMatrixFindTopK
  • Cl
    MPSImageCopyToMatrixiOS 11.0+
    @class MPSImageCopyToMatrix
  • Cl
    MPSMatrixCopyToImageiOS 12.0+
    @class MPSMatrixCopyToImage

Random Number Generation 4

Kernels and descriptors that generate random matrices and vectors on the GPU.

  • Cl
    MPSMatrixRandomiOS 13+
    @class MPSMatrixRandom
  • Cl
    MPSMatrixRandomMTGP32iOS 13+
    @class MPSMatrixRandomMTGP32
  • Cl
    MPSMatrixRandomPhiloxiOS 13+
    @class MPSMatrixRandomPhilox
  • Cl
    MPSMatrixRandomDistributionDescriptoriOS 13+
    @class MPSMatrixRandomDistributionDescriptor

Neural Network Foundations 9

Base kernel classes, gradient state, padding, and recurrent descriptors that underpin the neural-network graph.

  • Cl
    MPSCNNKerneliOS 10.0+
    @class MPSCNNKernel
  • Cl
    MPSCNNBinaryKerneliOS 11.0+
    @class MPSCNNBinaryKernel
  • Cl
    MPSCNNGradientKerneliOS 11.3+
    @class MPSCNNGradientKernel
  • Cl
    MPSCNNMultiaryKerneliOS 13.0+
    @class MPSCNNMultiaryKernel
  • Cl
    MPSNNDefaultPaddingiOS 11.0+
    @abstract This class provides some pre-rolled padding policies for common tasks
  • Cl
    MPSNNGradientStateiOS 11.3+
    @class A state created to record a MPSCNNKernel properties
  • Cl
    MPSNNBinaryGradientStateiOS 11.3+
    @class A state created to record MPSCNNBinaryKernel properties
  • Cl
    MPSNNMultiaryGradientStateiOS 13.0+
    An object that carries gradient state for a multi-input neural-network layer.
  • Cl
    MPSRNNDescriptoriOS 11.0+
    @class MPSRNNDescriptor

Convolution Layers 14

Convolution, transpose-convolution, and fully connected layers with their descriptors and weight state.

  • Cl
    MPSCNNConvolutionDescriptoriOS 10.0+
    @class MPSCNNConvolutionDescriptor
  • Cl
    MPSCNNSubPixelConvolutionDescriptoriOS 11.0+
    @abstract MPSCNNSubPixelConvolutionDescriptor can be used to create MPSCNNConvolution object that does sub pixel upsamling
  • Cl
    MPSCNNDepthWiseConvolutionDescriptoriOS 11.0+
    @abstract MPSCNNDepthWiseConvolutionDescriptor can be used to create MPSCNNConvolution object that does depthwise convolution
  • Cl
    MPSCNNConvolutionGradientStateiOS 11.3+
    @class MPSCNNConvolutionGradientState
  • Cl
    MPSCNNConvolutionTransposeGradientStateiOS 13.0+
    @class MPSCNNConvolutionTransposeGradientState
  • Cl
    MPSCNNConvolutionWeightsAndBiasesStateiOS 11.3+
    @class MPSCNNConvolutionWeightsAndBiasesState
  • Cl
    MPSCNNConvolutioniOS 10.0+
    @class MPSCNNConvolution
  • Cl
    MPSCNNConvolutionGradientiOS 11.3+
    @class MPSCNNConvolutionGradient
  • Cl
    MPSCNNFullyConnectediOS 10.0+
    @class MPSCNNFullyConnected
  • Cl
    MPSCNNFullyConnectedGradientiOS 11.3+
    @class MPSCNNFullyConnectedGradient
  • Cl
    MPSCNNConvolutionTransposeiOS 11.0+
    @class MPSCNNConvolutionTranspose
  • Cl
    MPSCNNConvolutionTransposeGradientiOS 13.0+
    @class MPSCNNConvolutionTransposeGradient
  • Cl
    MPSCNNBinaryConvolutioniOS 11.0+
    @class MPSCNNBinaryConvolution
  • Cl
    MPSCNNBinaryFullyConnectediOS 11.0+
    @class MPSCNNBinaryFullyConnected

Activation Functions 17

Neuron descriptors and activation-function kernels and their gradients.

  • Cl
    MPSNNNeuronDescriptoriOS 11.3+
    @class MPSNNNeuronDescriptor
  • Cl
    MPSCNNNeuroniOS 10.0+
    @class MPSCNNNeuron
  • Cl
    MPSCNNNeuronGradientiOS 11.3+
    @class MPSCNNNeuronGradient
  • Cl
    MPSCNNNeuronLineariOS 10.0+
    @class MPSCNNNeuronLinear
  • Cl
    MPSCNNNeuronReLUiOS 10.0+
    @class MPSCNNNeuronReLU
  • Cl
    MPSCNNNeuronPReLUiOS 11.0+
    @class MPSCNNNeuronPReLU
  • Cl
    MPSCNNNeuronSigmoidiOS 10.0+
    @class MPSCNNNeuronSigmoid
  • Cl
    MPSCNNNeuronHardSigmoidiOS 11.0+
    @class MPSCNNNeuronHardSigmoid
  • Cl
    MPSCNNNeuronTanHiOS 10.0+
    @class MPSCNNNeuronTanH
  • Cl
    MPSCNNNeuronAbsoluteiOS 10.0+
    @class MPSCNNNeuronAbsolute
  • Cl
    MPSCNNNeuronSoftPlusiOS 11.0+
    @class MPSCNNNeuronSoftPlus
  • Cl
    MPSCNNNeuronSoftSigniOS 11.0+
    @class MPSCNNNeuronSoftSign
  • Cl
    MPSCNNNeuronELUiOS 11.0+
    @class MPSCNNNeuronELU
  • Cl
    MPSCNNNeuronReLUNiOS 11.0+
    @class MPSCNNNeuronReLUN
  • Cl
    MPSCNNNeuronPoweriOS 11.3+
    @class MPSCNNNeuronPower
  • Cl
    MPSCNNNeuronExponentialiOS 11.3+
    @class MPSCNNNeuronExponential
  • Cl
    MPSCNNNeuronLogarithmiOS 11.3+
    @class MPSCNNNeuronLogarithm

Pooling Layers 10

Pooling kernels that downsample feature maps and their gradient counterparts.

  • Cl
    MPSCNNPoolingiOS 10.0+
    @class MPSCNNPooling
  • Cl
    MPSCNNPoolingMaxiOS 10.0+
    @class MPSCNNPoolingMax
  • Cl
    MPSCNNPoolingAverageiOS 10.0+
    @class MPSCNNPoolingAverage
  • Cl
    MPSCNNPoolingL2NormiOS 11.0+
    @class MPSCNNPoolingL2Norm
  • Cl
    MPSCNNDilatedPoolingMaxiOS 11.0+
    @class MPSCNNDilatedPoolingMax
  • Cl
    MPSCNNPoolingGradientiOS 11.3+
    @class MPSCNNPoolingGradient
  • Cl
    MPSCNNPoolingAverageGradientiOS 11.3+
    @class MPSCNNPoolingAverageGradient
  • Cl
    MPSCNNPoolingMaxGradientiOS 11.3+
    @class MPSCNNPoolingMaxGradient
  • Cl
    MPSCNNPoolingL2NormGradientiOS 11.3+
    @class MPSCNNPoolingL2NormGradient
  • Cl
    MPSCNNDilatedPoolingMaxGradientiOS 11.3+
    @class MPSCNNDilatedPoolingMaxGradient

Normalization Layers 19

Spatial, channel, batch, instance, and group normalization kernels with their state and gradients.

  • Cl
    MPSCNNSpatialNormalizationiOS 10.0+
    @class MPSCNNSpatialNormalization
  • Cl
    MPSCNNSpatialNormalizationGradientiOS 11.3+
    @class MPSCNNSpatialNormalizationGradient
  • Cl
    MPSCNNLocalContrastNormalizationiOS 10.0+
    @class MPSCNNLocalContrastNormalization
  • Cl
    MPSCNNLocalContrastNormalizationGradientiOS 11.3+
    @class MPSCNNLocalContrastNormalizationGradient
  • Cl
    MPSCNNCrossChannelNormalizationiOS 10.0+
    @class MPSCNNCrossChannelNormalization
  • Cl
    MPSCNNCrossChannelNormalizationGradientiOS 11.3+
    @class MPSCNNCrossChannelNormalizationGradient
  • Cl
    MPSCNNNormalizationGammaAndBetaStateiOS 11.3+
    @class MPSCNNNormalizationGammaAndBetaState
  • Cl
    MPSCNNNormalizationMeanAndVarianceStateiOS 12.0+
    @class MPSCNNNormalizationMeanAndVarianceState
  • Cl
    MPSCNNBatchNormalizationStateiOS 11.3+
    @class MPSCNNBatchNormalizationState
  • Cl
    MPSCNNBatchNormalizationiOS 11.3+
    @class MPSCNNBatchNormalization
  • Cl
    MPSCNNBatchNormalizationStatisticsiOS 11.3+
    @class MPSCNNBatchNormalizationStatistics
  • Cl
    MPSCNNBatchNormalizationGradientiOS 11.3+
    @class MPSCNNBatchNormalizationGradient
  • Cl
    MPSCNNBatchNormalizationStatisticsGradientiOS 11.3+
    @class MPSCNNBatchNormalizationStatisticsGradient
  • Cl
    MPSCNNInstanceNormalizationGradientStateiOS 11.3+
    @class MPSCNNInstanceNormalizationGradientState
  • Cl
    MPSCNNInstanceNormalizationiOS 11.3+
    @class MPSCNNInstanceNormalization
  • Cl
    MPSCNNInstanceNormalizationGradientiOS 11.3+
    @class MPSCNNInstanceNormalizationGradient
  • Cl
    MPSCNNGroupNormalizationGradientStateiOS 13.0+
    @class MPSCNNGroupNormalizationGradientState
  • Cl
    MPSCNNGroupNormalizationiOS 13.0+
    @class MPSCNNGroupNormalization
  • Cl
    MPSCNNGroupNormalizationGradientiOS 13.0+
    @class MPSCNNGroupNormalizationGradient

Loss and Softmax 13

Loss layers, softmax kernels, and gradient operations used in training.

  • Cl
    MPSCNNLossDataDescriptoriOS 11.3+
    @class MPSCNNLossDataDescriptor
  • Cl
    MPSCNNLossLabelsiOS 11.3+
    @class MPSCNNLossLabels
  • Cl
    MPSCNNLossDescriptoriOS 11.3+
    @class MPSCNNLossDescriptor
  • Cl
    MPSCNNLossiOS 11.3+
    @class MPSCNNLoss
  • Cl
    MPSCNNYOLOLossDescriptoriOS 12+
    @class MPSCNNYOLOLossDescriptor
  • Cl
    MPSCNNYOLOLossiOS 12.0+
    A kernel that computes the YOLO object-detection loss.
  • Cl
    MPSNNForwardLossiOS 13.0+
    @class MPSNNForwardLoss
  • Cl
    MPSNNLossGradientiOS 13.0+
    @class MPSNNLossGradient
  • Cl
    MPSNNInitialGradientiOS 13.0+
    @class MPSNNInitialGradient
  • Cl
    MPSCNNSoftMaxiOS 10.0+
    @class MPSCNNSoftMax
  • Cl
    MPSCNNSoftMaxGradientiOS 11.3+
    @class MPSCNNSoftMaxGradient
  • Cl
    MPSCNNLogSoftMaxiOS 10.0+
    @class MPSCNNLogSoftMax
  • Cl
    MPSCNNLogSoftMaxGradientiOS 11.3+
    @class MPSCNNLogSoftMaxGradient

Element-wise Operations 13

Arithmetic, comparison, and Gram-matrix kernels that combine feature channels and their gradients.

  • Cl
    MPSCNNArithmeticGradientStateiOS 11.3+
    @class MPSCNNArithmeticGradientState
  • Cl
    MPSCNNArithmeticiOS 11.3+
    @class MPSCNNArithmetic
  • Cl
    MPSCNNAddiOS 11.3+
    @class MPSCNNAdd
  • Cl
    MPSCNNSubtractiOS 11.3+
    @class MPSCNNSubtract
  • Cl
    MPSCNNMultiplyiOS 11.3+
    @class MPSCNNMultiply
  • Cl
    MPSCNNDivideiOS 11.3+
    @class MPSCNNDivide
  • Cl
    MPSNNCompareiOS 12.1+
    @class MPSNNCompare
  • Cl
    MPSCNNArithmeticGradientiOS 11.3+
    @class MPSCNNArithmeticGradient
  • Cl
    MPSCNNAddGradientiOS 11.3+
    @class MPSCNNAddGradient
  • Cl
    MPSCNNSubtractGradientiOS 11.3+
    @class MPSCNNSubtractGradient
  • Cl
    MPSCNNMultiplyGradientiOS 11.3+
    @class MPSCNNMultiplyGradient
  • Cl
    MPSNNGramMatrixCalculationiOS 13.0+
    @class MPSNNGramMatrixCalculation
  • Cl
    MPSNNGramMatrixCalculationGradientiOS 13.0+
    @class MPSNNGramMatrixCalculationGradient

Upsampling and Dropout 9

Upsampling and dropout kernels with their gradient and state objects.

  • Cl
    MPSCNNUpsamplingiOS 11.0+
    @class MPSCNNUpsampling
  • Cl
    MPSCNNUpsamplingNearestiOS 11.0+
    @class MPSCNNUpsamplingNearest
  • Cl
    MPSCNNUpsamplingBilineariOS 11.0+
    @class MPSCNNUpsamplingBilinear
  • Cl
    MPSCNNUpsamplingGradientiOS 11.3+
    @class MPSCNNUpsamplingGradient
  • Cl
    MPSCNNUpsamplingNearestGradientiOS 11.3+
    @class MPSCNNUpsamplingNearestGradient
  • Cl
    MPSCNNUpsamplingBilinearGradientiOS 11.3+
    @class MPSCNNUpsamplingBilinearGradient
  • Cl
    MPSCNNDropoutGradientStateiOS 11.3+
    @class MPSCNNDropoutGradientState
  • Cl
    MPSCNNDropoutiOS 11.3+
    @class MPSCNNDropout
  • Cl
    MPSCNNDropoutGradientiOS 11.3+
    @class MPSCNNDropoutGradient

Classes 198

  • Cl
    MPSRNNSingleGateDescriptoriOS 11.0+
    @class MPSRNNSingleGateDescriptor
  • Cl
    MPSGRUDescriptoriOS 11.0+
    @class MPSGRUDescriptor
  • Cl
    MPSLSTMDescriptoriOS 11.0+
    @class MPSLSTMDescriptor
  • Cl
    MPSRNNRecurrentImageStateiOS 11.0+
    @class MPSRNNRecurrentImageState
  • Cl
    MPSRNNImageInferenceLayeriOS 11.0+
    @class MPSRNNImageInferenceLayer
  • Cl
    MPSRNNRecurrentMatrixStateiOS 11.0+
    @class MPSRNNRecurrentMatrixState
  • Cl
    MPSRNNMatrixInferenceLayeriOS 11.0+
    @class MPSRNNMatrixInferenceLayer
  • Cl
    MPSRNNMatrixTrainingStateiOS 12.0+
    @class MPSRNNMatrixTrainingState
  • Cl
    MPSRNNMatrixTrainingLayeriOS 12.0+
    @class MPSRNNMatrixTrainingLayer
  • Cl
    MPSMatrixNeuroniOS 11.0+
    @class MPSMatrixNeuron
  • Cl
    MPSMatrixNeuronGradientiOS 12.0+
    @class MPSMatrixNeuronGradient
  • Cl
    MPSMatrixFullyConnectediOS 11.0+
    @class MPSMatrixFullyConnected
  • Cl
    MPSMatrixFullyConnectedGradientiOS 12.0+
    @class MPSMatrixFullyConnectedGradient
  • Cl
    MPSMatrixSumiOS 11.0+
    @class MPSMatrixSum
  • Cl
    MPSMatrixBatchNormalizationiOS 12.0+
    @class MPSMatrixBatchNormalization
  • Cl
    MPSMatrixBatchNormalizationGradientiOS 12.0+
    @class MPSMatrixBatchNormalizationGradient
Show 182 more
  • Cl
    MPSNNGridSampleiOS 13.0+
  • Cl
    MPSNNOptimizerDescriptoriOS 12.0+
    @class MPSNNOptimizerDescriptor
  • Cl
    MPSNNOptimizeriOS 12.0+
    @class MPSNNOptimizer
  • Cl
    MPSNNOptimizerStochasticGradientDescentiOS 12.0+
    @class MPSNNOptimizerStochasticGradientDescent
  • Cl
    MPSNNOptimizerRMSPropiOS 12.0+
    @class MPSNNOptimizerRMSProp
  • Cl
    MPSNNOptimizerAdamiOS 12.0+
    @class MPSNNOptimizerAdam
  • Cl
    MPSNNReduceUnaryiOS 11.3+
    @class MPSNNReduceUnary
  • Cl
    MPSNNReduceRowMiniOS 11.3+
    @class MPSNNReduceRowMin
  • Cl
    MPSNNReduceColumnMiniOS 11.3+
    @class MPSNNReduceColumnMin
  • Cl
    MPSNNReduceFeatureChannelsMiniOS 11.3+
    @class MPSNNReduceFeatureChannelsMin
  • Cl
    MPSNNReduceFeatureChannelsArgumentMiniOS 12.0+
    @class MPSNNReduceFeatureChannelsArgumentMin
  • Cl
    MPSNNReduceRowMaxiOS 11.3+
    @class MPSNNReduceRowMax
  • Cl
    MPSNNReduceColumnMaxiOS 11.3+
    @class MPSNNReduceColumnMax
  • Cl
    MPSNNReduceFeatureChannelsMaxiOS 11.3+
    @class MPSNNReduceFeatureChannelsMax
  • Cl
    MPSNNReduceFeatureChannelsArgumentMaxiOS 12.0+
    @class MPSNNReduceFeatureChannelsArgumentMax
  • Cl
    MPSNNReduceRowMeaniOS 11.3+
    @class MPSNNReduceRowMean
  • Cl
    MPSNNReduceColumnMeaniOS 11.3+
    @class MPSNNReduceColumnMean
  • Cl
    MPSNNReduceFeatureChannelsMeaniOS 11.3+
    @class MPSNNReduceFeatureChannelsMean
  • Cl
    MPSNNReduceRowSumiOS 11.3+
    @class MPSNNReduceRowSum
  • Cl
    MPSNNReduceColumnSumiOS 11.3+
    @class MPSNNReduceColumnSum
  • Cl
    MPSNNReduceFeatureChannelsSumiOS 11.3+
    @class MPSNNReduceFeatureChannelsSum
  • Cl
    MPSNNReduceBinaryiOS 11.3+
    @class MPSNNReduceBinary
  • Cl
    MPSNNReduceFeatureChannelsAndWeightsMeaniOS 11.3+
  • Cl
    MPSNNReduceFeatureChannelsAndWeightsSumiOS 11.3+
  • Cl
    MPSNNLocalCorrelationiOS 13.0+
    @class MPSNNLocalCorrelation
  • Cl
    MPSNNReshapeiOS 11.3+
  • Cl
    MPSNNReshapeGradientiOS 12.1+
    @class MPSNNReshapeGradient
  • Cl
    MPSNNPadiOS 12.1+
  • Cl
    MPSNNPadGradientiOS 12.1+
    @class MPSNNPadGradient
  • Cl
    MPSNNResizeBilineariOS 12.0+
    @class MPSNNResizeBilinear
  • Cl
    MPSNNCropAndResizeBilineariOS 12.0+
    @class MPSNNCropAndResizeBilinear
  • Cl
    MPSNNSliceiOS 11.3+
  • Cl
    MPSNNImageNodeiOS 11.0+
    @class MPSNNImageNode
  • Cl
    MPSNNStateNodeiOS 11.0+
    @class MPSNNStateNode
  • Cl
    MPSNNGradientStateNodeiOS 11.3+
    @class MPSNNGradientStateNode
  • Cl
    MPSCNNConvolutionGradientStateNodeiOS 11.3+
  • Cl
    MPSCNNConvolutionTransposeGradientStateNodeiOS 13.0+
  • Cl
    MPSNNBinaryGradientStateNodeiOS 11.3+
    @class MPSNNBinaryGradientStateNode
  • Cl
    MPSNNMultiaryGradientStateNodeiOS 13.0+
  • Cl
    MPSNNArithmeticGradientStateNodeiOS 11.3+
  • Cl
    MPSNNFilterNodeiOS 11.0+
    @class MPSNNFilterNode
  • Cl
    MPSNNGradientFilterNodeiOS 11.3+
    @class MPSNNGradientFilterNode
  • Cl
    MPSCNNConvolutionNodeiOS 11.0+
    @abstract A MPSNNFilterNode representing a MPSCNNConvolution kernel
  • Cl
    MPSCNNFullyConnectedNodeiOS 11.0+
    @abstract A MPSNNFilterNode representing a MPSCNNFullyConnected kernel
  • Cl
    MPSCNNBinaryConvolutionNodeiOS 11.0+
    @abstract A MPSNNFilterNode representing a MPSCNNBinaryConvolution kernel
  • Cl
    MPSCNNBinaryFullyConnectedNodeiOS 11.0+
    @abstract A MPSNNFilterNode representing a MPSCNNBinaryFullyConnected kernel
  • Cl
    MPSCNNConvolutionTransposeNodeiOS 11.0+
    @abstract A MPSNNFilterNode representing a MPSCNNConvolutionTranspose kernel
  • Cl
    MPSCNNConvolutionGradientNodeiOS 11.3+
  • Cl
    MPSCNNFullyConnectedGradientNodeiOS 13.0+
  • Cl
    MPSCNNConvolutionTransposeGradientNodeiOS 13.0+
  • Cl
    MPSCNNNeuronNodeiOS 11.0+
    @abstract virtual base class for MPSCNNNeuron nodes
  • Cl
    MPSCNNNeuronAbsoluteNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronAbsolute kernel
  • Cl
    MPSCNNNeuronELUNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronELU kernel
  • Cl
    MPSCNNNeuronReLUNNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronReLUN kernel
  • Cl
    MPSCNNNeuronLinearNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronLinear kernel
  • Cl
    MPSCNNNeuronReLUNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronReLU kernel
  • Cl
    MPSCNNNeuronSigmoidNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronSigmoid kernel
  • Cl
    MPSCNNNeuronHardSigmoidNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronHardSigmoid kernel
  • Cl
    MPSCNNNeuronSoftPlusNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronSoftPlus kernel
  • Cl
    MPSCNNNeuronSoftSignNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronSoftSign kernel
  • Cl
    MPSCNNNeuronTanHNodeiOS 11.0+
    @abstract A node representing a MPSCNNNeuronTanH kernel
  • Cl
    MPSCNNNeuronPReLUNodeiOS 11.0+
    @abstract A ReLU node with parameter a provided independently for each feature channel
  • Cl
    MPSCNNNeuronPowerNodeiOS 11.3+
    @abstract A node representing a MPSCNNNeuronPower kernel
  • Cl
    MPSCNNNeuronExponentialNodeiOS 11.3+
    @abstract A node representing a MPSCNNNeuronExponential kernel
  • Cl
    MPSCNNNeuronLogarithmNodeiOS 11.3+
    @abstract A node representing a MPSCNNNeuronLogarithm kernel
  • Cl
    MPSCNNNeuronGeLUNodeiOS 13.0+
    @abstract A node representing a MPSCNNNeuronGeLU kernel
  • Cl
    MPSCNNNeuronGradientNodeiOS 11.3+
    @abstract A node representing a MPSCNNNeuronGradient
  • Cl
    MPSNNUnaryReductionNodeiOS 12.1+
    @abstract A node for a unary MPSNNReduce node.
  • Cl
    MPSNNReductionRowMinNodeiOS 12.1+
  • Cl
    MPSNNReductionColumnMinNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsMinNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsArgumentMinNodeiOS 12.1+
  • Cl
    MPSNNReductionRowMaxNodeiOS 12.1+
  • Cl
    MPSNNReductionColumnMaxNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsMaxNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsArgumentMaxNodeiOS 12.1+
  • Cl
    MPSNNReductionRowMeanNodeiOS 12.1+
  • Cl
    MPSNNReductionColumnMeanNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsMeanNodeiOS 12.1+
  • Cl
    MPSNNReductionSpatialMeanNodeiOS 12.1+
  • Cl
    MPSNNReductionRowSumNodeiOS 12.1+
  • Cl
    MPSNNReductionColumnSumNodeiOS 12.1+
  • Cl
    MPSNNReductionFeatureChannelsSumNodeiOS 12.1+
  • Cl
    MPSCNNPoolingNodeiOS 11.0+
    @abstract A node for a MPSCNNPooling kernel
  • Cl
    MPSCNNPoolingAverageNodeiOS 11.0+
    @abstract A node representing a MPSCNNPoolingAverage kernel
  • Cl
    MPSCNNPoolingL2NormNodeiOS 11.0+
    @abstract A node representing a MPSCNNPoolingL2Norm kernel
  • Cl
    MPSCNNPoolingMaxNodeiOS 11.0+
    @abstract A node representing a MPSCNNPoolingMax kernel
  • Cl
    MPSCNNDilatedPoolingMaxNodeiOS 11.0+
    @abstract A node for a MPSCNNDilatedPooling kernel
  • Cl
    MPSCNNPoolingGradientNodeiOS 11.3+
  • Cl
    MPSCNNPoolingMaxGradientNodeiOS 11.3+
  • Cl
    MPSCNNPoolingAverageGradientNodeiOS 11.3+
  • Cl
    MPSCNNPoolingL2NormGradientNodeiOS 11.3+
  • Cl
    MPSCNNDilatedPoolingMaxGradientNodeiOS 11.3+
  • Cl
    MPSCNNNormalizationNodeiOS 11.0+
    @abstract virtual base class for CNN normalization nodes
  • Cl
    MPSCNNSpatialNormalizationNodeiOS 11.0+
    @abstract Node representing MPSCNNSpatialNormalization
  • Cl
    MPSCNNSpatialNormalizationGradientNodeiOS 11.3+
  • Cl
    MPSCNNLocalContrastNormalizationNodeiOS 11.0+
    @abstract Node representing MPSCNNLocalContrastNormalization
  • Cl
    MPSCNNLocalContrastNormalizationGradientNodeiOS 11.3+
  • Cl
    MPSCNNCrossChannelNormalizationNodeiOS 11.0+
    @abstract Node representing MPSCNNCrossChannelNormalization
  • Cl
    MPSCNNCrossChannelNormalizationGradientNodeiOS 11.3+
  • Cl
    MPSCNNInstanceNormalizationNodeiOS 11.3+
  • Cl
    MPSCNNInstanceNormalizationGradientNodeiOS 11.3+
  • Cl
    MPSCNNGroupNormalizationNodeiOS 13.0+
  • Cl
    MPSCNNGroupNormalizationGradientNodeiOS 13.0+
  • Cl
    MPSCNNBatchNormalizationNodeiOS 11.3+
    @class MPSCNNBatchNormalizationNode
  • Cl
    MPSCNNBatchNormalizationGradientNodeiOS 11.3+
    @class MPSCNNBatchNormalizationGradientNode
  • Cl
    MPSNNScaleNodeiOS 11.0+
    @abstract Abstract Node representing a image resampling operation
  • Cl
    MPSNNBilinearScaleNodeiOS 11.0+
    @abstract A MPSNNScale object that uses bilinear interpolation for resampling
  • Cl
    MPSNNLanczosScaleNodeiOS 11.0+
    @abstract A MPSNNScale object that uses the Lanczos resampling filter
  • Cl
    MPSNNBinaryArithmeticNodeiOS 11.0+
    @abstract virtual base class for basic arithmetic nodes
  • Cl
    MPSNNAdditionNodeiOS 11.0+
    @abstract returns elementwise sum of left + right
  • Cl
    MPSNNSubtractionNodeiOS 11.0+
    @abstract returns elementwise difference of left - right
  • Cl
    MPSNNMultiplicationNodeiOS 11.0+
    @abstract returns elementwise product of left * right
  • Cl
    MPSNNDivisionNodeiOS 11.0+
    @abstract returns elementwise quotient of left / right
  • Cl
    MPSNNComparisonNodeiOS 12.1+
    @abstract returns elementwise comparison of left and right
  • Cl
    MPSNNArithmeticGradientNodeiOS 11.3+
  • Cl
    MPSNNAdditionGradientNodeiOS 11.3+
    @abstract returns gradient for either primary or secondary source image from the inference pass.
  • Cl
    MPSNNSubtractionGradientNodeiOS 11.3+
    @abstract returns gradient for either primary or secondary source image from the inference pass.
  • Cl
    MPSNNMultiplicationGradientNodeiOS 11.3+
    @abstract returns gradient for either primary or secondary source image from the inference pass.
  • Cl
    MPSCNNDropoutNodeiOS 11.3+
  • Cl
    MPSCNNDropoutGradientNodeiOS 11.3+
  • Cl
    MPSNNLabelsNodeiOS 11.3+
    @class MPSNNLabelsNode
  • Cl
    MPSCNNLossNodeiOS 11.3+
    @class MPSCNNLossNode
  • Cl
    MPSCNNYOLOLossNodeiOS 11.3+
    @class MPSCNNYOLOLossNode
  • Cl
    MPSNNConcatenationNodeiOS 11.0+
    Node representing a the concatenation (in the feature channel dimension) of the results from one or more kernels
  • Cl
    MPSNNConcatenationGradientNodeiOS 11.3+
    @class MPSNNConcatenationGradientNode
  • Cl
    MPSNNReshapeNodeiOS 12.1+
    @abstract A node for a MPSNNReshape kernel
  • Cl
    MPSNNReshapeGradientNodeiOS 12.1+
  • Cl
    MPSNNReductionSpatialMeanGradientNodeiOS 12.1+
  • Cl
    MPSNNPadNodeiOS 12.1+
    @class MPSNNPadNode
  • Cl
    MPSNNPadGradientNodeiOS 12.1+
  • Cl
    MPSCNNSoftMaxNodeiOS 11.0+
    Node representing a MPSCNNSoftMax kernel
  • Cl
    MPSCNNSoftMaxGradientNodeiOS 11.3+
    Node representing a MPSCNNSoftMaxGradient kernel
  • Cl
    MPSCNNLogSoftMaxNodeiOS 11.0+
    Node representing a MPSCNNLogSoftMax kernel
  • Cl
    MPSCNNLogSoftMaxGradientNodeiOS 11.3+
    Node representing a MPSCNNLogSoftMaxGradient kernel
  • Cl
    MPSCNNUpsamplingNearestNodeiOS 11.0+
    Node representing a MPSCNNUpsamplingNearest kernel
  • Cl
    MPSCNNUpsamplingBilinearNodeiOS 11.0+
    Node representing a MPSCNNUpsamplingBilinear kernel
  • Cl
    MPSCNNUpsamplingNearestGradientNodeiOS 11.3+
    Node representing a MPSCNNUpsamplingNearest kernel
  • Cl
    MPSCNNUpsamplingBilinearGradientNodeiOS 11.3+
    Node representing a MPSCNNUpsamplingBilinear kernel
  • Cl
    MPSNNGramMatrixCalculationNodeiOS 13.0+
    Node representing a @ref MPSNNGramMatrixCalculation kernel
  • Cl
    MPSNNGramMatrixCalculationGradientNodeiOS 13.0+
    Node representing a @ref MPSNNGramMatrixCalculationGradient kernel
  • Cl
    MPSNNForwardLossNodeiOS 13.0+
    Node representing a @ref MPSNNForwardLoss kernel
  • Cl
    MPSNNLossGradientNodeiOS 13.0+
    Node representing a @ref MPSNNLossGradient kernel
  • Cl
    MPSNNInitialGradientNodeiOS 13.0+
    @class MPSNNInitialGradientNode
  • Cl
    MPSNNGraphiOS 11.0+
    @class MPSNNGraph
  • Cl
    MPSNDArrayMultiaryBaseiOS 13.0+
  • Cl
    MPSNDArrayMultiaryKerneliOS 13.0+
  • Cl
    MPSNDArrayMultiaryGradientKerneliOS 13.0+
  • Cl
    MPSNDArrayUnaryKerneliOS 13.0+
  • Cl
    MPSNDArrayUnaryGradientKerneliOS 13.0+
  • Cl
    MPSNDArrayBinaryKerneliOS 13.0+
  • Cl
    MPSNDArrayBinaryPrimaryGradientKerneliOS 13+
    @class MPSNDArrayDivisionPrimaryGradient
  • Cl
    MPSNDArrayBinarySecondaryGradientKerneliOS 13+
    @class MPSNDArrayDivisionSecondaryGradient
  • Cl
    MPSNDArrayGradientStateiOS 13.0+
    @class A state created to record a MPSCNNKernel properties
  • Cl
    MPSNDArrayMatrixMultiplicationiOS 13+
    @class MPSNDArrayMatrixMultiplication
  • Cl
    MPSNDArrayStridedSliceiOS 13+
    @class MPSNDStridedSlice
  • Cl
    MPSNDArrayStridedSliceGradientiOS 13+
    @class MPSNDStridedSliceGradient
  • Cl
    MPSNDArrayGatherGradientStateiOS 13.0+
    @class A state created to record a MPSNDArrayGather kernel properties
  • Cl
    MPSNDArrayGatheriOS 13+
    @class MPSNDArrayGather
  • Cl
    MPSNDArrayGatherGradientiOS 13+
    @class MPSNDArrayGatherGradient
  • Cl
    MPSNDArrayIdentityiOS 18.0+
    @class MPSNDArrayIdentityKernel
  • Cl
    MPSNDArrayQuantizationDescriptoriOS 18.0+
    @class MPSNDArrayQuantizationDescriptor
  • Cl
    MPSNDArrayAffineQuantizationDescriptoriOS 18.0+
    @class MPSNDArrayAffineQuantizationDescriptor
  • Cl
    MPSNDArrayLUTQuantizationDescriptoriOS 18.0+
    @class MPSNDArrayLUTQuantizationDescriptor
  • Cl
    MPSNDArrayQuantizedMatrixMultiplicationiOS 18.0+
    @class MPSNDArrayQuantizedMatrixMultiplication
  • Cl
    MPSNDArrayLUTDequantizeiOS 18.0+
    @class MPSNDArrayLUTDequantize
  • Cl
    MPSNDArrayVectorLUTDequantizeiOS 18.0+
    @class MPSNDArrayVectorLUTDequantize
  • Cl
    MPSNDArrayAffineInt4DequantizeiOS 18.0+
    @class MPSNDArrayAffineInt4Dequantize
  • Cl
    MPSFunctioniOS 27.0+
  • Cl
    MPSFColorConversioniOS 27.0+
  • Cl
    MPSAccelerationStructureGroup
    @brief A group of acceleration structures which may be used together in an instance acceleration
  • Cl
    MPSAccelerationStructure
    @brief A data structure built over geometry used to accelerate ray tracing
  • Cl
    MPSPolygonBuffer
    @brief A vertex buffer and optional index and mask buffer for a set of polygons
  • Cl
    MPSPolygonAccelerationStructure
    @brief An acceleration structure built over polygonal shapes
  • Cl
    MPSTriangleAccelerationStructure
    @brief An acceleration structure built over triangles
  • Cl
    MPSQuadrilateralAccelerationStructure
    @brief An acceleration structure built over quadrilaterals
  • Cl
    MPSInstanceAccelerationStructure
    @brief An acceleration structure built over instances of other acceleration structures
  • Cl
    MPSSVGFiOS 13.0+
    @brief Reduces noise in images rendered with Monte Carlo ray tracing methods
  • Cl
    MPSSVGFDefaultTextureAllocatoriOS 13.0+
    @brief A default implementation of the MPSSVGFTextureAllocator protocol. Maintains a cache of
  • Cl
    MPSSVGFDenoiseriOS 13.0+
    @brief A convenience object which uses an MPSSVGF object to manage the denoising process
  • Cl
    MPSTemporalAAiOS 13.0+
    @brief Reduces aliasing in an image by accumulating samples over multiple frames
  • Cl
    MPSRayIntersector
    @class MPSRayIntersector

Structures 54

  • St
    MPSKernelOptions
  • St
    MPSAliasingStrategy
  • St
    MPSOffset
    @struct MPSOffset
  • St
    MPSOrigin
    @struct MPSOrigin
  • St
    MPSSize
    @struct MPSSize
  • St
    MPSDimensionSlice
    @struct MPSDimensionSize
  • St
    MPSRegion
    @struct MPSRegion
  • St
    MPSScaleTransform
    @struct MPSScaleTransform
  • St
    MPSImageCoordinate
    @struct MPSImageCoordinate
  • St
    MPSImageRegion
    @struct MPSImageRegion
  • St
    MPSImageReadWriteParams
    @struct MPSImageReadWriteParams
  • St
    MPSStateTextureInfo
  • St
    MPSDeviceCapsValues
  • St
    MPSCustomKernelIndex
  • St
    MPSMatrixOffset
    @struct MPSMatrixOffset
  • St
    MPSIntegerDivisionParams
Show 38 more
  • St
    MPSCustomKernelSourceInfo
  • St
    MPSCustomKernelInfo
  • St
    MPSImageType
  • St
    MPSCustomKernelArgumentCount
  • St
    MPSMatrixCopyOffsets
    @struct MPSMatrixCopyOffsets
  • St
    MPSMatrixRandomDistribution
  • St
    MPSImageKeypointRangeInfo
    @brief Specifies information to find the keypoints in an image.
  • St
    MPSImageKeypointData
    @brief Specifies keypoint information.
  • St
    MPSImageHistogramInfo
    @brief Specifies information to compute the histogram for channels of an image.
  • St
    MPSNNConvolutionAccumulatorPrecisionOption
  • St
    MPSNNTrainingStyle
  • St
    MPSCNNBatchNormalizationFlags
  • St
    MPSNNPaddingMethod
  • St
    MPSCNNConvolutionGradientOption
  • St
    MPSNNComparisonType
  • St
    MPSNDArrayOffsets
  • St
    MPSNDArraySizes
  • St
    MPSNDArrayQuantizationScheme
  • St
    MPSFunctions_AABB
    @abstract Axis aligned bounding box for color gamut
  • St
    MPSFColorConversionOptions
  • St
    _MPSPackedFloat3
  • St
    _MPSAxisAlignedBoundingBox
    @brief An axis aligned bounding box with a min and max point
  • St
    MPSRayOriginDirection
    @brief Represents a 3D ray with an origin and a direction
  • St
    MPSRayPackedOriginDirection
    @brief Represents a 3D ray with an origin and a direction
  • St
    MPSRayOriginMinDistanceDirectionMaxDistance
    @brief Represents a 3D ray with an origin, a direction, and an intersection
  • St
    MPSRayOriginMaskDirectionMaxDistance
    @brief Represents a 3D ray with an origin, a direction, and a mask to filter out intersections
  • St
    MPSIntersectionDistance
    @brief Returned intersection result which contains the distance from the ray origin to the
  • St
    MPSIntersectionDistancePrimitiveIndex
    @brief Intersection result which contains the distance from the ray origin to the
  • St
    MPSIntersectionDistancePrimitiveIndexBufferIndex
    @brief Intersection result which contains the distance from the ray origin to the
  • St
    MPSIntersectionDistancePrimitiveIndexCoordinates
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSIntersectionDistancePrimitiveIndexBufferIndexCoordinates
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSIntersectionDistancePrimitiveIndexInstanceIndex
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSIntersectionDistancePrimitiveIndexBufferIndexInstanceIndex
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSIntersectionDistancePrimitiveIndexInstanceIndexCoordinates
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSIntersectionDistancePrimitiveIndexBufferIndexInstanceIndexCoordinates
    @brief Intersection result which contains the distance from the ray origin to the intersection
  • St
    MPSAccelerationStructureUsage
    @brief Options describing how an acceleration structure will be used
  • St
    MPSRayMaskOptions
    @brief Options for the MPSRayIntersector ray mask options property
  • St
    MPSDeviceOptions

Enumerations 32

  • En
    MPSImageEdgeMode
  • En
    MPSImageFeatureChannelFormat
  • En
    MPSFloatDataTypeBit
  • En
    MPSFloatDataTypeShift
  • En
    MPSDataType
  • En
    MPSPurgeableStateiOS 8.0+
  • En
    MPSDataLayoutiOS 11.0+
  • En
    MPSStateResourceType
  • En
    MPSAlphaType
  • En
    MPSMatrixDecompositionStatus
  • En
    MPSCNNConvolutionFlags
  • En
    MPSCNNBinaryConvolutionFlags
  • En
    MPSCNNBinaryConvolutionType
  • En
    MPSCNNNeuronType
  • En
    MPSCNNConvolutionWeightsLayout
  • En
    MPSCNNWeightsQuantizationType
Show 16 more
  • En
    MPSCNNLossType
  • En
    MPSCNNReductionType
  • En
    MPSRNNSequenceDirection
  • En
    MPSRNNBidirectionalCombineMode
  • En
    MPSRNNMatrixId
  • En
    MPSNNRegularizationTypeiOS 12.0+
  • En
    MPSAccelerationStructureStatus
    @brief Possible values of the acceleration structure status property
  • En
    MPSPolygonType
  • En
    MPSTransformTypeiOS 12.0+
    @brief Instance transformation type options
  • En
    MPSTemporalWeightingiOS 13.0+
    @brief Controls how samples are weighted over time
  • En
    MPSIntersectionTypeiOS 12.0+
    @brief Options for the MPSRayIntersector intersection type property
  • En
    MPSTriangleIntersectionTestType
    @brief Options for the MPSRayIntersector triangle intersection test type property
  • En
    MPSBoundingBoxIntersectionTestType
    @brief Options for the MPSRayIntersector bounding box intersection test type property
  • En
    MPSRayDataTypeiOS 12.0+
    @brief Options for the MPSRayIntersector ray data type property
  • En
    MPSIntersectionDataTypeiOS 12.0+
    @brief Intersection data type options
  • En
    MPSRayMaskOperator
    @brief Options for the MPSRayIntersector ray mask operator property

Protocols 16

  • Pr
    MPSDeviceProvider
    @abstract A way of extending a NSCoder to enable the setting of MTLDevice for unarchived objects
  • Pr
    MPSImageAllocator
    @abstract A class that allocates new MPSImage or MPSTemporaryImage
  • Pr
    MPSHeapProvider
  • Pr
    MPSNDArrayAllocator
  • Pr
    MPSNNPadding
    @protocol MPSNNPadding
  • Pr
    MPSImageSizeEncodingState
    @abstract MPSStates conforming to this protocol contain information about a image size elsewhere in the graph
  • Pr
    MPSCNNConvolutionDataSource
    @protocol MPSCNNConvolutionDataSource
  • Pr
    MPSCNNBatchNormalizationDataSourceiOS 11.3+
    @protocol MPSCNNBatchNormalizationDataSource
  • Pr
    MPSCNNInstanceNormalizationDataSourceiOS 11.3+
    @protocol MPSCNNInstanceNormalizationDataSource
  • Pr
    MPSCNNGroupNormalizationDataSourceiOS 13.0+
    @protocol MPSCNNGroupNormalizationDataSource
  • Pr
    MPSHandle
    @protocol MPSHandle
  • Pr
    MPSNNTrainableNode
  • Pr
    MPSImageTransformProvider
  • Pr
    MPSNNGramMatrixCallback
    @protocol MPSNNGramMatrixCallback
  • Pr
    MPSNNLossCallback
    @protocol MPSNNLossCallback
  • Pr
    MPSSVGFTextureAllocatoriOS 13.0+
    @brief Protocol dictating how texture allocator objects should operate so that they can be used

Type Aliases 9

  • Ty
    MPSDeviceCaps
  • Ty
    MPSFunctionConstant
  • Ty
    MPSFunctionConstantInMetal
  • Ty
    MPSCopyAllocator
  • Ty
    MPSGradientNodeBlock
    @abstract Block callback for customizing gradient nodes as they are constructed
  • Ty
    MPSNNGraphCompletionHandler
    @abstract A notification when computeAsyncWithSourceImages:completionHandler: has finished
  • Ty
    MPSPackedFloat3
  • Ty
    MPSAxisAlignedBoundingBox
    @brief An axis aligned bounding box with a min and max point
  • Ty
    MPSAccelerationStructureCompletionHandler
    @brief A block of code invoked when an operation on an MPSAccelerationStructure is completed
← Graphics, Imaging & Compositing