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

Class JCudaKernels

java.lang.Object

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

public class JCudaKernels
extends Object

Utility class that allows LibMatrixCUDA as well as GPUObject to invoke custom CUDA kernels.

The utility org.apache.sysml.runtime.instructions.gpu.context.JCudaKernels simplifies the launching of the kernels. For example: to launch a kernel copyUpperToLowerTriangleDense<<1,1,32,32>>(jcudaDenseMatrixPtr, dim, dim*dim) , the user has to call: kernels.launchKernel("copyUpperToLowerTriangleDense", new ExecutionConfig(1,1,32,32), jcudaDenseMatrixPtr, dim, dim*dim)

Method Summary

Modifier and Type	Method and Description
static void	checkResult(int cuResult)
void	launchKernel(String name, ExecutionConfig config, Object... arguments) Setups the kernel parameters and launches the kernel using cuLaunchKernel API.

Methods inherited from class java.lang.Object

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

Method Detail

launchKernel

public void launchKernel(String name,
                         ExecutionConfig config,
                         Object... arguments)
                  throws DMLRuntimeException

Setups the kernel parameters and launches the kernel using cuLaunchKernel API. This function currently supports two dimensional grid and blocks.

Parameters:

name - name of the kernel

config - execution configuration

arguments - can be of type Pointer, long, double, float and int

Throws:

DMLRuntimeException - if DMLRuntimeException occurs

checkResult

public static void checkResult(int cuResult)
                        throws DMLRuntimeException

Throws:

DMLRuntimeException