• Packages that use GPUContext 
    Package Description
    org.apache.sysml.runtime.instructions.gpu.context  
    org.apache.sysml.runtime.matrix.data  

    • Uses of GPUContext in org.apache.sysml.runtime.instructions.gpu.context

      Methods in org.apache.sysml.runtime.instructions.gpu.context that return types with arguments of type GPUContext 
      Modifier and Type Method and Description
      static List<GPUContext> GPUContextPool.reserveAllGPUContexts()
      Reserves and gets an initialized list of GPUContexts
      Methods in org.apache.sysml.runtime.instructions.gpu.context with parameters of type GPUContext 
      Modifier and Type Method and Description
      static CSRPointer CSRPointer.allocateEmpty(GPUContext gCtx, long nnz2, long rows)
      Factory method to allocate an empty CSR Sparse matrix on the GPU
      static CSRPointer CSRPointer.allocateForDgeam(GPUContext gCtx, jcuda.jcusparse.cusparseHandle handle, CSRPointer A, CSRPointer B, int m, int n)
      Estimates the number of non zero elements from the results of a sparse cusparseDgeam operation C = a op(A) + b op(B)
      static CSRPointer CSRPointer.allocateForMatrixMultiply(GPUContext gCtx, jcuda.jcusparse.cusparseHandle handle, CSRPointer A, int transA, CSRPointer B, int transB, int m, int n, int k)
      Estimates the number of non-zero elements from the result of a sparse matrix multiplication C = A * B and returns the CSRPointer to C with the appropriate GPU memory.
      static CSRPointer GPUObject.columnMajorDenseToRowMajorSparse(GPUContext gCtx, jcuda.jcusparse.cusparseHandle cusparseHandle, jcuda.Pointer densePtr, int rows, int cols)
      Convenience method to convert a CSR matrix to a dense matrix on the GPU Since the allocated matrix is temporary, bookkeeping is not updated.
      static jcuda.Pointer GPUObject.transpose(GPUContext gCtx, jcuda.Pointer densePtr, int m, int n, int lda, int ldc)
      Transposes a dense matrix on the GPU by calling the cublasDgeam operation

    • Uses of GPUContext in org.apache.sysml.runtime.matrix.data

      Methods in org.apache.sysml.runtime.matrix.data with parameters of type GPUContext 
      Modifier and Type Method and Description
      static void LibMatrixCUDA.abs(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "abs" operation on a matrix on the GPU
      static void LibMatrixCUDA.acos(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "acos" operation on a matrix on the GPU
      static void LibMatrixCUDA.asin(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "asin" operation on a matrix on the GPU
      static void LibMatrixCUDA.atan(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "atan" operation on a matrix on the GPU
      static void LibMatrixCUDA.axpy(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName, double constant)
      Performs daxpy operation
      static void LibMatrixCUDA.batchNormalizationBackward(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject dout, org.apache.sysml.runtime.controlprogram.caching.MatrixObject scale, org.apache.sysml.runtime.controlprogram.caching.MatrixObject ret, org.apache.sysml.runtime.controlprogram.caching.MatrixObject retScale, org.apache.sysml.runtime.controlprogram.caching.MatrixObject retBias, double epsilon)
      This method computes the backpropagation errors for image, scale and bias of batch normalization layer
      static void LibMatrixCUDA.batchNormalizationForwardInference(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject scale, org.apache.sysml.runtime.controlprogram.caching.MatrixObject bias, org.apache.sysml.runtime.controlprogram.caching.MatrixObject runningMean, org.apache.sysml.runtime.controlprogram.caching.MatrixObject runningVar, org.apache.sysml.runtime.controlprogram.caching.MatrixObject ret, double epsilon)
      Performs the forward BatchNormalization layer computation for inference
      static void LibMatrixCUDA.batchNormalizationForwardTraining(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject scale, org.apache.sysml.runtime.controlprogram.caching.MatrixObject bias, org.apache.sysml.runtime.controlprogram.caching.MatrixObject runningMean, org.apache.sysml.runtime.controlprogram.caching.MatrixObject runningVar, org.apache.sysml.runtime.controlprogram.caching.MatrixObject ret, org.apache.sysml.runtime.controlprogram.caching.MatrixObject retRunningMean, org.apache.sysml.runtime.controlprogram.caching.MatrixObject retRunningVar, double epsilon, double exponentialAverageFactor)
      Performs the forward BatchNormalization layer computation for training
      static void LibMatrixCUDA.biasAdd(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject input, org.apache.sysml.runtime.controlprogram.caching.MatrixObject bias, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock)
      Performs the operation corresponding to the DML script: ones = matrix(1, rows=1, cols=Hout*Wout) output = input + matrix(bias %*% ones, rows=1, cols=F*Hout*Wout) This operation is often followed by conv2d and hence we have introduced bias_add(input, bias) built-in function
      static void LibMatrixCUDA.biasMultiply(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject input, org.apache.sysml.runtime.controlprogram.caching.MatrixObject bias, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock)
      Performs the operation corresponding to the DML script: ones = matrix(1, rows=1, cols=Hout*Wout) output = input * matrix(bias %*% ones, rows=1, cols=F*Hout*Wout) This operation is often followed by conv2d and hence we have introduced bias_add(input, bias) built-in function
      static void LibMatrixCUDA.cbind(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName) 
      static void LibMatrixCUDA.ceil(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "ceil" operation on a matrix on the GPU
      static void LibMatrixCUDA.conv2d(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject filter, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q) 
      static void LibMatrixCUDA.conv2d(GPUContext gCtx, String instName, jcuda.Pointer image, jcuda.Pointer filter, jcuda.Pointer output, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      Performs 2D convolution Takes up an insignificant amount of intermediate space when CONVOLUTION_PREFERENCE is set to CUDNN_CONVOLUTION_FWD_NO_WORKSPACE Intermediate space is required by the filter descriptor and convolution descriptor which are metadata structures and don't scale with the size of the input
      static void LibMatrixCUDA.conv2dBackwardData(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject filter, org.apache.sysml.runtime.controlprogram.caching.MatrixObject dout, org.apache.sysml.runtime.controlprogram.caching.MatrixObject output, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      This method computes the backpropogation errors for previous layer of convolution operation
      static void LibMatrixCUDA.conv2dBackwardFilter(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject dout, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      This method computes the backpropogation errors for filter of convolution operation
      static void LibMatrixCUDA.conv2dBiasAdd(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject bias, org.apache.sysml.runtime.controlprogram.caching.MatrixObject filter, org.apache.sysml.runtime.controlprogram.caching.MatrixObject output, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      Does a 2D convolution followed by a bias_add
      static void LibMatrixCUDA.cos(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "cos" operation on a matrix on the GPU
      static void LibMatrixCUDA.denseDenseMatmult(GPUContext gCtx, String instName, jcuda.Pointer output, int leftRows1, int leftCols1, int rightRows1, int rightCols1, boolean isLeftTransposed1, boolean isRightTransposed1, jcuda.Pointer leftPtr, jcuda.Pointer rightPtr)
      Dense-dense matrix multiply C = op(A) * op(B), A and B are dense matrices On the host, the matrices are in row-major format; cuBLAS expects them in column-major format.
      static void LibMatrixCUDA.exp(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "exp" operation on a matrix on the GPU
      static void LibMatrixCUDA.floor(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "floor" operation on a matrix on the GPU
      static boolean LibMatrixCUDA.isInSparseFormat(GPUContext gCtx, org.apache.sysml.runtime.controlprogram.caching.MatrixObject mo) 
      static void LibMatrixCUDA.log(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "log" operation on a matrix on the GPU
      static org.apache.sysml.runtime.controlprogram.caching.MatrixObject LibMatrixCUDA.matmult(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject left, org.apache.sysml.runtime.controlprogram.caching.MatrixObject right, String outputName, boolean isLeftTransposed, boolean isRightTransposed)
      Matrix multiply on GPU Examines sparsity and shapes and routes call to appropriate method from cuBLAS or cuSparse C = op(A) x op(B)
      static void LibMatrixCUDA.matmultTSMM(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject left, String outputName, boolean isLeftTransposed)
      Performs tsmm, A %*% A' or A' %*% A, on GPU by exploiting cublasDsyrk(...)
      static void LibMatrixCUDA.matrixMatrixArithmetic(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName, boolean isLeftTransposed, boolean isRightTransposed, org.apache.sysml.runtime.matrix.operators.BinaryOperator op)
      Performs elementwise arithmetic operation specified by op of two input matrices in1 and in2
      static void LibMatrixCUDA.matrixMatrixRelational(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName, org.apache.sysml.runtime.matrix.operators.BinaryOperator op)
      Performs elementwise operation relational specified by op of two input matrices in1 and in2
      static void LibMatrixCUDA.matrixScalarArithmetic(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in, String outputName, boolean isInputTransposed, org.apache.sysml.runtime.matrix.operators.ScalarOperator op)
      Entry point to perform elementwise matrix-scalar arithmetic operation specified by op
      static void LibMatrixCUDA.matrixScalarRelational(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in, String outputName, org.apache.sysml.runtime.matrix.operators.ScalarOperator op)
      Entry point to perform elementwise matrix-scalar relational operation specified by op
      static void LibMatrixCUDA.maxpooling(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      performs maxpooling on GPU by exploiting cudnnPoolingForward(...)
      static void LibMatrixCUDA.maxpoolingBackward(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject image, org.apache.sysml.runtime.controlprogram.caching.MatrixObject dout, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q)
      Performs maxpoolingBackward on GPU by exploiting cudnnPoolingBackward(...) This method computes the backpropogation errors for previous layer of maxpooling operation
      static void LibMatrixCUDA.performMaxpooling(GPUContext gCtx, String instName, jcuda.Pointer x, jcuda.jcudnn.cudnnTensorDescriptor xDesc, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock, int N, int C, int H, int W, int K, int R, int S, int pad_h, int pad_w, int stride_h, int stride_w, int P, int Q) 
      static void LibMatrixCUDA.rbind(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName) 
      static void LibMatrixCUDA.relu(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in, String outputName)
      Performs the relu operation on the GPU.
      static void LibMatrixCUDA.reluBackward(GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject input, org.apache.sysml.runtime.controlprogram.caching.MatrixObject dout, org.apache.sysml.runtime.controlprogram.caching.MatrixObject outputBlock)
      This method computes the backpropagation errors for previous layer of relu operation
      static void LibMatrixCUDA.round(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "round" operation on a matrix on the GPU
      static void LibMatrixCUDA.sign(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "sign" operation on a matrix on the GPU
      static void LibMatrixCUDA.sin(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "sin" operation on a matrix on the GPU
      static void LibMatrixCUDA.sliceOperations(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.util.IndexRange ixrange, String outputName)
      Method to perform rangeReIndex operation for a given lower and upper bounds in row and column dimensions.
      static void LibMatrixCUDA.solve(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in2, String outputName)
      Implements the "solve" function for systemml Ax = B (A is of size m*n, B is of size m*1, x is of size n*1)
      static void LibMatrixCUDA.sqrt(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "sqrt" operation on a matrix on the GPU
      static void LibMatrixCUDA.tan(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String outputName)
      Performs an "tan" operation on a matrix on the GPU
      static void LibMatrixCUDA.transpose(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in, String outputName)
      Transposes the input matrix using cublasDgeam
      static void LibMatrixCUDA.unaryAggregate(org.apache.sysml.runtime.controlprogram.context.ExecutionContext ec, GPUContext gCtx, String instName, org.apache.sysml.runtime.controlprogram.caching.MatrixObject in1, String output, org.apache.sysml.runtime.matrix.operators.AggregateUnaryOperator op)
      Entry point to perform Unary aggregate operations on the GPU.
Skip navigation links
  • Prev
  • Next

Copyright © 2017 The Apache Software Foundation. All rights reserved.