こんにちは、JOCL (opencl) は初めてです。画像ごとの強度の合計を取るために、このコードを書きました。カーネルは、相互に配置されたすべての画像のすべてのピクセルの 1D 配列を取ります。画像は 300x300 であるため、画像あたり 90000 ピクセルです。現時点では、これを順番に行う場合よりも遅くなります。
私のコード
package PAR;
/*
* JOCL - Java bindings for OpenCL
*
* Copyright 2009 Marco Hutter - http://www.jocl.org/
*/
import IMAGE_IO.ImageReader;
import IMAGE_IO.Input_Folder;
import static org.jocl.CL.*;
import org.jocl.*;
/**
* A small JOCL sample.
*/
public class IPPARA {
/**
* The source code of the OpenCL program to execute
*/
private static String programSource =
"__kernel void "
+ "sampleKernel(__global uint *a,"
+ " __global uint *c)"
+ "{"
+ "__private uint intensity_core=0;"
+ " uint i = get_global_id(0);"
+ " for(uint j=i*90000; j < (i+1)*90000; j++){ "
+ " intensity_core += a[j];"
+ " }"
+ "c[i]=intensity_core;"
+ "}";
/**
* The entry point of this sample
*
* @param args Not used
*/
public static void main(String args[]) {
long numBytes[] = new long[1];
ImageReader imagereader = new ImageReader() ;
int srcArrayA[] = imagereader.readImages();
int size[] = new int[1];
size[0] = srcArrayA.length;
long before = System.nanoTime();
int dstArray[] = new int[size[0]/90000];
Pointer srcA = Pointer.to(srcArrayA);
Pointer dst = Pointer.to(dstArray);
// Obtain the platform IDs and initialize the context properties
System.out.println("Obtaining platform...");
cl_platform_id platforms[] = new cl_platform_id[1];
clGetPlatformIDs(platforms.length, platforms, null);
cl_context_properties contextProperties = new cl_context_properties();
contextProperties.addProperty(CL_CONTEXT_PLATFORM, platforms[0]);
// Create an OpenCL context on a GPU device
cl_context context = clCreateContextFromType(
contextProperties, CL_DEVICE_TYPE_CPU, null, null, null);
if (context == null) {
// If no context for a GPU device could be created,
// try to create one for a CPU device.
context = clCreateContextFromType(
contextProperties, CL_DEVICE_TYPE_CPU, null, null, null);
if (context == null) {
System.out.println("Unable to create a context");
return;
}
}
// Enable exceptions and subsequently omit error checks in this sample
CL.setExceptionsEnabled(true);
// Get the list of GPU devices associated with the context
clGetContextInfo(context, CL_CONTEXT_DEVICES, 0, null, numBytes);
// Obtain the cl_device_id for the first device
int numDevices = (int) numBytes[0] / Sizeof.cl_device_id;
cl_device_id devices[] = new cl_device_id[numDevices];
clGetContextInfo(context, CL_CONTEXT_DEVICES, numBytes[0],
Pointer.to(devices), null);
// Create a command-queue
cl_command_queue commandQueue =
clCreateCommandQueue(context, devices[0], 0, null);
// Allocate the memory objects for the input- and output data
cl_mem memObjects[] = new cl_mem[2];
memObjects[0] = clCreateBuffer(context,
CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
Sizeof.cl_uint * srcArrayA.length, srcA, null);
memObjects[1] = clCreateBuffer(context,
CL_MEM_READ_WRITE,
Sizeof.cl_uint * (srcArrayA.length/90000), null, null);
// Create the program from the source code
cl_program program = clCreateProgramWithSource(context,
1, new String[]{programSource}, null, null);
// Build the program
clBuildProgram(program, 0, null, null, null, null);
// Create the kernel
cl_kernel kernel = clCreateKernel(program, "sampleKernel", null);
// Set the arguments for the kernel
clSetKernelArg(kernel, 0,
Sizeof.cl_mem, Pointer.to(memObjects[0]));
clSetKernelArg(kernel, 1,
Sizeof.cl_mem, Pointer.to(memObjects[1]));
// Set the work-item dimensions
long local_work_size[] = new long[]{1};
long global_work_size[] = new long[]{(srcArrayA.length/90000)*local_work_size[0]};
// Execute the kernel
clEnqueueNDRangeKernel(commandQueue, kernel, 1, null,
global_work_size, local_work_size, 0, null, null);
// Read the output data
clEnqueueReadBuffer(commandQueue, memObjects[1], CL_TRUE, 0,
(srcArrayA.length/90000) * Sizeof.cl_float, dst, 0, null, null);
// Release kernel, program, and memory objects
clReleaseMemObject(memObjects[0]);
clReleaseMemObject(memObjects[1]);
clReleaseKernel(kernel);
clReleaseProgram(program);
clReleaseCommandQueue(commandQueue);
clReleaseContext(context);
long after = System.nanoTime();
System.out.println("Time: " + (after - before) / 1e9);
}
}
回答の提案の後、CPU を介した並列コードは、順次コードとほぼ同じ速さです。他にできる改善点はありますか?