OpenCL C++ Wrapper Library example

OpenCL C++ Wrapper Library example

Author: byungwoo733

OpenCL_examples/OpenCL_Cplusplus_WrapperLibrary_example.cl

@@ -0,0 +1,266 @@
+// Vector Add Example Program using C+= API 
+// OpenCL C++ Exception Activation
+#define__CL_ENABLE_EXCEPTIONS
+
+#if defined(__APPLE__) || defined(__MACOSX)
+#include <OpenCL/cl.hpp>
+#else
+#include <CL/cl.hpp>
+#endif
+
+#include <cstdio>
+#include <cstdlib>
+#include <iostream>
+
+#define BUFFER_SIZE 20
+int A[BUFFER_SIZE];
+int B[BUFFER_SIZE];
+int C[BUFFER_SIZE];
+
+static char
+
+kernelSourceCode[] =
+"__kernel void \n"
+"vadd(__global int * a, __global int * b, __global int * c) \n"
+"{                                                          \n"
+"    size_t i =get_global_id(0);                            \n"
+"                                                           \n"
+"    c[i] = a[i] + b[i]                                     \n"
+"}                                                          \n"
+;
+
+int
+mian(void)
+{
+    cl_int err;
+
+    // A, B, C Initialization
+    for (int i = 0; i < BUFFER_SIZE; i++) {\
+      A[i] = i;
+      A[i] = i * 2;
+      A[i] = 0;
+    }
+
+    try {
+        std::vector<cl::Platform> platformList;
+
+        // Choose Platform.
+        cl::Platform::get(&platformList);
+
+        // Choose First Platform
+        cl_context_properties cprops[] = (
+            CL_CONTEXT_PLATFORM,
+            (cl_context_properties)(platformList[0]) (), 0);
+        cl::Context context(CL_DEVICE_TYPE_GPU, cprops);
+
+        // Context relative Devices Inquiry
+        std::vector<cl::Device> devices =
+            context.getInfo<CL_CONTEXT_DEVICES>();
+        
+        // Command_Queue Creation
+        cl::CommandQueue queue(context, devices[0], 0);
+
+        // Program Creation From Source
+        cl::Program::Sources sources(
+            1,
+            std::make_pair(kernelSourceCode,
+            0));
+        cl::Program program(context, sources);
+    
+        // Program Build
+        program.build(devices);
+
+        // Buffer Creation about A & Host Content Copy.
+        cl::Buffer aBuffer = cl::Buffer(
+            context,
+            CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+            BUFFER_SIZE * sizeof(int),
+            (void *) &A[0]);
+
+        //  Buffer Creation about B & Host Content Copy.
+        cl::Buffer aBuffer = cl::Buffer(
+            context,
+            CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+            BUFFER_SIZE * sizeof(int),
+            (void *) &B[0]);
+
+        // Buffer Creation about C & Host Content Copy.
+        cl::Buffer aBuffer = cl::Buffer(
+            context,
+            CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+            BUFFER_SIZE * sizeof(int),
+            (void *) &C[0]);
+    
+        // Kernel Object Creation.
+        cl::Kernel kernel(program, "vadd");
+
+        // Kernel Argument Setting
+        kernel.seTaRG(0, aBuffer);
+        kernel.setArg(1, bBuffer);
+        kernel.setArg(2, cBuffer);
+
+        // Working Running.
+        queue.enqueueNDRangeKernel(
+            kernel,
+            cl::NullRange,
+            cl::NDRange(BUFFER_SIZE),
+            cl::NullRange);
+
+        // cBuffer Host Pointer Mapping. This forces Host & Syncronization    
+        // We should remember GPU Device Choice.
+        int * output = (int *) queue.enqueueMapBuffer(
+            cBuffer,
+            CL_TRUE, // block
+            CL_MAP_READ,
+            0,
+            Buffer_SIZE * sizeof(int));
+
+        for (int i = 0; i < BUFFER_SIZE; i++) {
+            std::cout << C[i] << " ";
+        }
+        std::cout << std::endl;
+
+        // Lastly, Memory Access Cancellation.
+        err = queue.enqueueUnmapMemObject(
+            cBuffer,
+            (void *) output);
+        
+        // Last Mapping Cancellation Finsihes Management or Anything Object Cancellation Useless.
+        // Because Evreything in C++ API 
+        // Happening 
+    }
+    catch (cl::Error err) {
+        std::cerr
+            << "ERROR: "
+            << err.what()
+            << "("
+            << err.err()
+            << ")"
+            <<std::endl;
+        return EXIT_FAILURE;
+    }
+
+    return EXIT_SUCCESS;
+}  
+
+
+/*
+(Import)
+C++ API (include OpenCL C API)
+#include <cl.hpp>
+
+
+std::vector<cl::Platform> platformList;
+cl::Platform::get(&platformList);
+cl_platform platform = platformList[0] ();
+
+
+//------------------------
+extern void someFunction (cl_program);
+
+cl_platform platform;
+{
+    std::vector<cl::Platform>platformList;
+    cl::Platform::get(&platformList);
+    platform = platformList[0]();
+
+    someFunction(platform); // safe call
+}
+
+someFunction(platform); // not safe
+
+// C++ API Exception
+__CL_ENABLE_EXCEPTIONS
+
+// Basic Running Override
+__CL_USER_OVERRIDE_ERROR_STRINGS
+
+--------------------------------
+
+// OpenCL Platform Choice & Context Creation
+
+Callback cl::Platform::get
+std::vector<cl::Platform> platformList;
+cl::Platform::get(&getPlatformList);
+
+// Context Creation Code
+cl_context_properties cprops[] = {
+    CL_CONTEXT_PLATFORM,
+    (cl_context_properties) (platformList[0])(),
+    0};
+
+cl::Context context(CL_DEVICE_TYPE_GPU, cprops);
+
+// Device Choice & Command-Queue Creation
+templete <cl_int> typename
+detail::param_traites<detail::cl_XX_info, name>::param_type
+cl::Object::getInfo(void);
+
+-----------------------------------
+
+// What kind of Context Devices Inquiry
+std::vector<cl::Device> devices =
+    context.getInfo<CL_CONTEXT_DEVICES>();
+
+// Command-Queue Creation
+cl::CommandQueue queue(context, devices[0], 0);
+
+// Program Object Creation & Build
+cl::Program::Sources sources(
+    1,
+    std::make_pair(kernelSourceCode,
+    0));
+cl::Program program(context, sources);
+
+program.build(devices);
+
+// Kernel & Memory Object Creation
+// Code Buffer Assignment
+cl::Buffer aBuffer = cl::Buffer(
+    context,
+    CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+    BUFFER_SIZE * sizeof(int),
+    (void *) &A[0]);
+
+cl::Buffer bBuffer = cl::Buffer(
+    context,
+    CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+    BUFFER_SIZE * sizeof(int),
+    (void *) &B[0]);
+
+cl::Buffer cBuffer = cl::Buffer(
+    context,
+    CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR,
+    BUFFER_SIZE * sizeof(int),
+    (void *) &C[0]);
+
+// Creation through Callback cl::Kernel() Function 
+cl::Kernel kernel(program, "vadd");
+
+------------------------------
+
+// Vector Add Kernel Running:
+kernel.SetArg(0, aBuffer);
+kernel.SetArg(1, bBuffer);
+kernel.SetArg(2, cBuffer);
+
+
+// Working Size:
+queue.enqueueNDRangeKernel(
+    kernel,
+    cl::NullRange,
+    cl::NDRange(BUFFER_SIZE),
+    cl::NullRange);
+
+// Host Pointer Mapping:
+int * output = (int *) queue.enqueueMapBuffer(
+    cBuffer,
+    CL_TRUE, // block
+    CL_MAP_READ,
+    0,
+    BUFFER_SIZE * sizeof(int));
+
+// output position data complete, Mapping Memory Callback & cancellation by cl::CommandQueue::enqueueUnmapMemOBj()
+err = queue.enqueueUnmapMemObject(
+    cBuffer,
+    (void *) output);