commit cf480b6d51501adae9cb99d41dfe15293ea037a4
Author: trivernis <trivernis@protonmail.com>
Date:   Mon Nov 23 22:03:50 2020 +0100

    Add simple prime calculation
    
    Signed-off-by: trivernis <trivernis@protonmail.com>

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..0d631dc
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,3 @@
+.idea
+cmake-build-debug
+primes.txt
\ No newline at end of file
diff --git a/CMakeLists.txt b/CMakeLists.txt
new file mode 100644
index 0000000..43fefcd
--- /dev/null
+++ b/CMakeLists.txt
@@ -0,0 +1,11 @@
+cmake_minimum_required(VERSION 3.17)
+project(opencl_demo)
+
+set(CMAKE_CXX_STANDARD 14)
+
+find_package(OpenCL REQUIRED)
+
+add_executable(opencl_demo main.cpp)
+
+target_include_directories (opencl_demo PUBLIC ${CMAKE_CURRENT_SOURCE_DIR})
+target_link_libraries(opencl_demo OpenCL::OpenCL)
\ No newline at end of file
diff --git a/main.cpp b/main.cpp
new file mode 100644
index 0000000..212fd68
--- /dev/null
+++ b/main.cpp
@@ -0,0 +1,121 @@
+#include <iostream>
+#include <fstream>
+#include <chrono>
+#include "CL/cl.h"
+
+#define MAX_SOURCE_SIZE (0x100000)
+
+using namespace std;
+
+int main() {
+    const int PRIME_LIST_SIZE = 1024 * 1024 * 64;
+    const int WORK_GROUP_SIZE = 128;
+    const auto LIST_MEM_SIZE = sizeof(int) * PRIME_LIST_SIZE;
+    auto *IN = (int *) malloc(LIST_MEM_SIZE);
+
+    for (int i = 0; i < PRIME_LIST_SIZE; i++) {
+        IN[i] = (i*2) + 1;
+    }
+
+    // Load the kernel source code into the array source_str
+    FILE *fp;
+    char *source_str;
+    size_t source_size;
+
+    fp = fopen("./prime_check_kernel.cl", "r");
+
+    if (!fp) {
+        fprintf(stderr, "Failed to load kernel.\n");
+        exit(1);
+    }
+
+    source_str = (char *) malloc(MAX_SOURCE_SIZE);
+    source_size = fread(source_str, 1, MAX_SOURCE_SIZE, fp);
+    fclose(fp);
+
+    // Get platform and device information
+    cl_platform_id platform_id = nullptr;
+    cl_device_id device_id = nullptr;
+    cl_uint ret_num_devices;
+    cl_uint ret_num_platforms;
+    cl_int ret = clGetPlatformIDs(1, &platform_id, &ret_num_platforms);
+    ret = clGetDeviceIDs(platform_id, CL_DEVICE_TYPE_GPU, 1,
+                         &device_id, &ret_num_devices);
+
+    cl_context context = clCreateContext(nullptr, 1, &device_id, nullptr, nullptr, &ret);
+
+    cl_command_queue command_queue = clCreateCommandQueueWithProperties(context, device_id, nullptr, &ret);
+
+    cl_mem in_mem_obj = clCreateBuffer(context, CL_MEM_READ_ONLY,
+                                      LIST_MEM_SIZE, nullptr, &ret);
+    cl_mem out_mem_obj = clCreateBuffer(context, CL_MEM_READ_WRITE, sizeof(bool) * PRIME_LIST_SIZE, nullptr, &ret);
+
+    ret = clEnqueueWriteBuffer(command_queue, in_mem_obj, CL_TRUE, 0,
+                               LIST_MEM_SIZE, IN, 0, nullptr, nullptr);
+
+    cl_program program = clCreateProgramWithSource(context, 1,
+                                                   (const char **) &source_str, (const size_t *) &source_size, &ret);
+
+    ret = clBuildProgram(program, 1, &device_id, nullptr, nullptr, nullptr);
+
+    if (ret != CL_SUCCESS) {
+        size_t log_size;
+        char *program_log;
+
+        clGetProgramBuildInfo(program, device_id, CL_PROGRAM_BUILD_LOG, 0, nullptr, &log_size);
+        program_log = (char*) malloc(log_size+1);
+        program_log[log_size] = '\0';
+        clGetProgramBuildInfo(program, device_id, CL_PROGRAM_BUILD_LOG,
+                              log_size+1, program_log, nullptr);
+        printf("\n=== ERROR ===\n\n%s\n=============\n", program_log);
+        free(program_log);
+        exit(1);
+    }
+
+    cl_kernel kernel = clCreateKernel(program, "check_prime", &ret);
+
+    ret = clSetKernelArg(kernel, 0, sizeof(cl_mem), (void *) &in_mem_obj);
+    ret = clSetKernelArg(kernel, 1, sizeof(cl_mem), (void *) & out_mem_obj);
+
+
+    auto start = chrono::high_resolution_clock::now();
+    size_t global_item_size = PRIME_LIST_SIZE;
+    size_t local_item_size = WORK_GROUP_SIZE;
+    printf("Enqueueing %d prime checks with a work group size of %d\n", PRIME_LIST_SIZE, WORK_GROUP_SIZE);
+    ret = clEnqueueNDRangeKernel(command_queue, kernel, 1, nullptr,
+                                 &global_item_size, &local_item_size, 0, nullptr, nullptr);
+
+    auto *OUT = (bool *) malloc(sizeof(bool) * PRIME_LIST_SIZE);
+    ret = clEnqueueReadBuffer(command_queue, out_mem_obj, CL_TRUE, 0,
+                              sizeof(bool) * PRIME_LIST_SIZE, OUT, 0, nullptr, nullptr);
+    auto finish = chrono::high_resolution_clock::now();
+
+    printf("Writing primes to file...\n");
+    ofstream prime_file;
+    prime_file.open("primes.txt");
+
+    int count = 0;
+    for (int i = 0; i < PRIME_LIST_SIZE; i++) {
+        if (OUT[i]) {
+            prime_file << IN[i] << "\n";
+            count++;
+        }
+    }
+    prime_file.close();
+    chrono::duration<double> elapsed = finish - start;
+
+    printf("Calculated %d primes in %f ms\n", count, elapsed.count() * 1000);
+
+    // Clean up
+    ret = clFlush(command_queue);
+    ret = clFinish(command_queue);
+    ret = clReleaseKernel(kernel);
+    ret = clReleaseProgram(program);
+    ret = clReleaseMemObject(in_mem_obj);
+    ret = clReleaseMemObject(out_mem_obj);
+    ret = clReleaseCommandQueue(command_queue);
+    ret = clReleaseContext(context);
+    free(IN);
+    free(OUT);
+    return 0;
+}
\ No newline at end of file
diff --git a/prime_check_kernel.cl b/prime_check_kernel.cl
new file mode 100644
index 0000000..fa05a40
--- /dev/null
+++ b/prime_check_kernel.cl
@@ -0,0 +1,18 @@
+__kernel void check_prime(__global int *IN, __global bool *OUT) {
+    int id = get_global_id(0);
+    int num = IN[id];
+    bool prime = true;
+
+    if (num < 3 || num % 2 == 0) {
+        prime = false;
+    } else {
+        for (int i = 3; i <= sqrt((float) num); i += 2) {
+            if (num % i == 0) {
+                prime = false;
+                break;
+            }
+        }
+    }
+
+    OUT[id] = prime;
+}
\ No newline at end of file