org.apache.sysml.runtime.instructions.gpu.context

Class ExecutionConfig

java.lang.Object

org.apache.sysml.runtime.instructions.gpu.context.ExecutionConfig

public class ExecutionConfig
extends Object

Java Wrapper to specify CUDA execution configuration for launching custom kernels

Field Summary

Fields

Modifier and Type	Field and Description
int	blockDimX
int	blockDimY
int	blockDimZ
int	gridDimX
int	gridDimY
int	gridDimZ
int	sharedMemBytes
jcuda.driver.CUstream	stream

Constructor Summary

Constructors

Constructor and Description
ExecutionConfig(int gridDimX, int blockDimX)
ExecutionConfig(int gridDimX, int blockDimX, int sharedMemBytes) Convenience constructor for setting the number of blocks, number of threads and the shared memory size
ExecutionConfig(int gridDimX, int gridDimY, int blockDimX, int blockDimY)

Method Summary

Modifier and Type	Method and Description
static ExecutionConfig	getConfigForSimpleMatrixOperations(int rlen, int clen) Use this for simple vector operations and use following in the kernel int index = blockIdx.x * blockDim.x + threadIdx.x
static ExecutionConfig	getConfigForSimpleVectorOperations(int numCells) Use this for simple vector operations and use following in the kernel int index = blockIdx.x * blockDim.x + threadIdx.x
String	toString()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

gridDimX

public int gridDimX

gridDimY

public int gridDimY

gridDimZ

public int gridDimZ

blockDimX

public int blockDimX

blockDimY

public int blockDimY

blockDimZ

public int blockDimZ

sharedMemBytes

public int sharedMemBytes

stream

public jcuda.driver.CUstream stream

Constructor Detail

ExecutionConfig

public ExecutionConfig(int gridDimX,
                       int blockDimX,
                       int sharedMemBytes)

Convenience constructor for setting the number of blocks, number of threads and the shared memory size

Parameters:

gridDimX - Number of blocks (for CUDA Kernel)

blockDimX - Number of threads per block (for CUDA Kernel)

sharedMemBytes - Amount of Shared memory (for CUDA Kernel)

ExecutionConfig

public ExecutionConfig(int gridDimX,
                       int blockDimX)

ExecutionConfig

public ExecutionConfig(int gridDimX,
                       int gridDimY,
                       int blockDimX,
                       int blockDimY)

Method Detail

getConfigForSimpleVectorOperations

public static ExecutionConfig getConfigForSimpleVectorOperations(int numCells)

Use this for simple vector operations and use following in the kernel int index = blockIdx.x * blockDim.x + threadIdx.x

This tries to schedule as minimum grids as possible.

Parameters:

numCells - number of cells

Returns:

execution configuration

getConfigForSimpleMatrixOperations

public static ExecutionConfig getConfigForSimpleMatrixOperations(int rlen,
                                                                 int clen)

Use this for simple vector operations and use following in the kernel int index = blockIdx.x * blockDim.x + threadIdx.x

Parameters:

rlen - number of rows

clen - number of columns

Returns:

execution configuration

toString

public String toString()

Overrides:

toString in class Object