#define CL_TARGET_OPENCL_VERSION 120

#include <utility>
#include <CL/cl.hpp>
#include <cstdio>
#include <cstdlib>
#include <fstream>
#include <iostream>
#include <string>
#include <iterator>

const std::string hw("Hello World!");

inline void checkErr(cl_int err, const char* name) {
	if(err != CL_SUCCESS) {
		std::cerr << "ERROR: " << name << " (" << err << ")" << std::endl;
		exit(EXIT_FAILURE);
	}
}

void initMatrix(float *mat, int size) {
	for(int i = 0; i < size; i++) {
		mat[i] = rand() % 50;
	}
}

int main(void) {
	cl_int err;

	std::vector<cl::Platform> platformList;
	cl::Platform::get(&platformList);
	checkErr(platformList.size() != 0 ? CL_SUCCESS : -1, "cl::Platform::get");

	std::cerr << "Number of platforms: " << platformList.size() << std::endl;

	for(int i = 0; i < platformList.size(); i++) {
      std::string platformVendor;
   	platformList[i].getInfo((cl_platform_info)CL_PLATFORM_VENDOR, &platformVendor);
   	std::cerr << "Platform #" << i << " Vendor: " << platformVendor << std::endl;
      std::string platformName;
      platformList[i].getInfo((cl_platform_info)CL_PLATFORM_NAME, &platformName);
      std::cerr << "Platform #" << i << " Name: " << platformName << std::endl;
   }

	cl_context_properties cprops[3] = { CL_CONTEXT_PLATFORM, (cl_context_properties)(platformList[0])(), 0 };
	cl::Context context(CL_DEVICE_TYPE_GPU, cprops, NULL, NULL, &err);
	checkErr(err, "Context::Context()");

	// Matrices as seen on OpenCL device
	cl::Buffer deviceA;
	cl::Buffer deviceB;
	cl::Buffer deviceC;

	// Seed rand()
	srand(432414);

	// Allocate host side of memory for inputs
	int widthA = 1024;
	int widthB = 1024;

	unsigned int sizeA = widthA * widthA;
	unsigned int memSizeA = sizeA * sizeof(float);
	float* hostA = (float*) malloc(memSizeA);

	unsigned int sizeB = widthB * widthB;
	unsigned int memSizeB = sizeB * sizeof(float);
	float* hostB = (float*) malloc(memSizeB);

	initMatrix(hostA, sizeA);
	initMatrix(hostB, sizeB);

	// Allocate host memory for result
	unsigned int sizeC = widthA * widthB;
	unsigned int memSizeC = sizeC * sizeof(float);
	float* hostC = (float*) malloc(memSizeC);

	std::vector<cl::Device> devices;
	devices = context.getInfo<CL_CONTEXT_DEVICES>();
	checkErr(devices.size() > 0 ? CL_SUCCESS : -1, "devices.size() > 0");

	std::ifstream file("fresh_kernel.cl");
	checkErr(file.is_open() ? CL_SUCCESS : -1, "fresh_kernel.cl");

	std::string programSourceString(std::istreambuf_iterator<char>(file), (std::istreambuf_iterator<char>()));
	cl::Program::Sources programSource(1, std::make_pair(programSourceString.c_str(), programSourceString.length() + 1));
	cl::Program program(context, programSource);
	err = program.build(devices, "");
	checkErr(err, "Program::Build()");

	cl::Kernel kernel(program, "matrix_mult", &err);
	checkErr(err, "Kernel::Kernel()");

	deviceC = cl::Buffer(context, CL_MEM_WRITE_ONLY, memSizeC, NULL, &err);
	deviceA = cl::Buffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR, memSizeA, hostA, &err);
	deviceB = cl::Buffer(context, CL_MEM_READ_ONLY | CL_MEM_COPY_HOST_PTR, memSizeB, hostB, &err);

	size_t localWorkSize[2], globalWorkSize[2];

	err = kernel.setArg(0, deviceC);
	err |= kernel.setArg(1, deviceA);
	err |= kernel.setArg(2, deviceB);
	err |= kernel.setArg(3, widthA);
	err |= kernel.setArg(4, widthB);

	localWorkSize[0] = 16;
	localWorkSize[1] = 16;
	globalWorkSize[0] = 1024;
	globalWorkSize[1] = 1024;

	cl::CommandQueue queue(context, devices[0], 0, &err);
	checkErr(err, "CommandQueue::CommandQueue()");
	cl::Event event;
	err = queue.enqueueNDRangeKernel(kernel, cl::NullRange, cl::NDRange(1024, 1024), cl::NDRange(16, 16), NULL, &event);
	checkErr(err, "CommandQueue::enqueueNDRangeKernel()");

	event.wait();
	err = queue.enqueueReadBuffer(deviceC, CL_TRUE, 0, memSizeC, hostC);
	checkErr(err, "CommandQueue::enqueueReadBuffer()");

	// Verification
	int matches = 0;
	for(int y = 0; y < widthA; y++) {
		for(int x = 0; x < widthA; x++) {
			float value = 0;
			for(int k = 0; k < widthA; k++) {
		      float elemA = hostA[y * widthA + k];
		      float elemB = hostB[k * widthB + x];
		      value += (elemA * elemB);
		   }
			if(hostC[y * widthA + x] == value) {
				matches++;
			} else {
				std::cout << "Device gave " << hostC[y * widthA + x] << ", host gave " << value << std::endl;
			}
		}
	}
	std::cout << matches << "/" << widthA * widthA << " correct." << std::endl;

	return EXIT_SUCCESS;
}