mirror of
https://github.com/vortexgpgpu/vortex.git
synced 2025-04-23 21:39:10 -04:00
minor update
This commit is contained in:
Blaise Tine
parent
e7f8b40d93
commit
bd18b03cc3
5 changed files with 0 additions and 313 deletions
|
|
@ -1,7 +0,0 @@
|
|||
PROJECT = fft4
|
||||
|
||||
SRCS = main.cc
|
||||
|
||||
OPTS ?= -n32
|
||||
|
||||
include ../common.mk
|
||||
|
|
@ -1,3 +0,0 @@
|
|||
#pragma once
|
||||
|
||||
#define LOCAL_SIZE 16
|
||||
|
|
@ -1,63 +0,0 @@
|
|||
#include "common.h"
|
||||
|
||||
__kernel void fft_radix4(__global float2* input, __global float2* output, const unsigned int N) {
|
||||
int globalId = get_global_id(0);
|
||||
int localId = get_local_id(0);
|
||||
int groupId = get_group_id(0);
|
||||
|
||||
// Allocate local memory to store intermediate results and twiddle factors
|
||||
__local float2 localData[LOCAL_SIZE];
|
||||
__local float2 twiddleFactors[LOCAL_SIZE / 4];
|
||||
|
||||
// Calculate twiddle factors for this FFT stage and store in local memory
|
||||
if (localId < LOCAL_SIZE / 4) {
|
||||
float angle = -2 * M_PI * localId / LOCAL_SIZE;
|
||||
twiddleFactors[localId] = (float2)(cos(angle), sin(angle));
|
||||
}
|
||||
barrier(CLK_LOCAL_MEM_FENCE);
|
||||
|
||||
// Calculate the offset for the data this work-group will process
|
||||
int offset = groupId * LOCAL_SIZE;
|
||||
|
||||
// Load a chunk of input into local memory for faster access
|
||||
if (globalId < N) {
|
||||
localData[localId] = input[globalId];
|
||||
}
|
||||
barrier(CLK_LOCAL_MEM_FENCE);
|
||||
|
||||
// Perform the Radix-4 FFT on the data chunk in local memory
|
||||
for (unsigned int stride = 1; stride < LOCAL_SIZE; stride *= 4) {
|
||||
int twiddleIndex = (localId / stride) % 4;
|
||||
float2 twiddle = twiddleFactors[twiddleIndex * (LOCAL_SIZE / (4 * stride))];
|
||||
|
||||
// Load data
|
||||
float2 data0 = localData[localId];
|
||||
float2 data1 = localData[localId + stride];
|
||||
float2 data2 = localData[localId + 2 * stride];
|
||||
float2 data3 = localData[localId + 3 * stride];
|
||||
|
||||
// Apply twiddle factors
|
||||
data1 *= twiddle;
|
||||
data2 *= twiddle * twiddle;
|
||||
data3 *= twiddle * twiddle * twiddle;
|
||||
|
||||
// Radix-4 butterfly operations
|
||||
float2 t0 = data0 + data2;
|
||||
float2 t1 = data0 - data2;
|
||||
float2 t2 = data1 + data3;
|
||||
float2 t3 = (data1 - data3) * (float2)(0, -1);
|
||||
|
||||
// Store results
|
||||
localData[localId] = t0 + t2;
|
||||
localData[localId + stride] = t1 + t3;
|
||||
localData[localId + 2 * stride] = t0 - t2;
|
||||
localData[localId + 3 * stride] = t1 - t3;
|
||||
|
||||
barrier(CLK_LOCAL_MEM_FENCE);
|
||||
}
|
||||
|
||||
// Write the results back to global memory
|
||||
if (globalId < N) {
|
||||
output[globalId] = localData[localId];
|
||||
}
|
||||
}
|
||||
|
|
@ -1,240 +0,0 @@
|
|||
#include <stdio.h>
|
||||
#include <stdlib.h>
|
||||
#include <assert.h>
|
||||
#include <CL/opencl.h>
|
||||
#include <string.h>
|
||||
#include <time.h>
|
||||
#include <unistd.h>
|
||||
#include <chrono>
|
||||
#include <vector>
|
||||
#include <cmath>
|
||||
#include "common.h"
|
||||
|
||||
#define KERNEL_NAME "fft_radix4"
|
||||
|
||||
#define FLOAT_ULP 6
|
||||
|
||||
struct float2 {
|
||||
float x;
|
||||
float y;
|
||||
|
||||
float2(float real = 0.0f, float imag = 0.0f) : x(real), y(imag) {}
|
||||
|
||||
float2 operator+(const float2& other) const {
|
||||
return {x + other.x, y + other.y};
|
||||
}
|
||||
|
||||
float2 operator-(const float2& other) const {
|
||||
return {x - other.x, y - other.y};
|
||||
}
|
||||
|
||||
float2 operator*(const float2& other) const {
|
||||
return {x * other.x - y * other.y, x * other.y + y * other.x};
|
||||
}
|
||||
};
|
||||
|
||||
#define CL_CHECK(_expr) \
|
||||
do { \
|
||||
cl_int _err = _expr; \
|
||||
if (_err == CL_SUCCESS) \
|
||||
break; \
|
||||
printf("OpenCL Error: '%s' returned %d!\n", #_expr, (int)_err); \
|
||||
cleanup(); \
|
||||
exit(-1); \
|
||||
} while (0)
|
||||
|
||||
#define CL_CHECK2(_expr) \
|
||||
({ \
|
||||
cl_int _err = CL_INVALID_VALUE; \
|
||||
decltype(_expr) _ret = _expr; \
|
||||
if (_err != CL_SUCCESS) { \
|
||||
printf("OpenCL Error: '%s' returned %d!\n", #_expr, (int)_err); \
|
||||
cleanup(); \
|
||||
exit(-1); \
|
||||
} \
|
||||
_ret; \
|
||||
})
|
||||
|
||||
static int read_kernel_file(const char* filename, uint8_t** data, size_t* size) {
|
||||
if (nullptr == filename || nullptr == data || 0 == size)
|
||||
return -1;
|
||||
|
||||
FILE* fp = fopen(filename, "r");
|
||||
if (NULL == fp) {
|
||||
fprintf(stderr, "Failed to load kernel.");
|
||||
return -1;
|
||||
}
|
||||
fseek(fp , 0 , SEEK_END);
|
||||
long fsize = ftell(fp);
|
||||
rewind(fp);
|
||||
|
||||
*data = (uint8_t*)malloc(fsize);
|
||||
*size = fread(*data, 1, fsize, fp);
|
||||
|
||||
fclose(fp);
|
||||
|
||||
return 0;
|
||||
}
|
||||
|
||||
static std::vector<float2> referenceDFT(const std::vector<float2>& input) {
|
||||
std::vector<float2> output(input.size());
|
||||
for (unsigned int k = 0; k < input.size(); ++k) { // For each output element
|
||||
output[k] = {0, 0}; // Initialize to zero
|
||||
for (unsigned int n = 0; n < input.size(); ++n) { // For each input element
|
||||
float angle = -2 * M_PI * k * n / input.size();
|
||||
float2 twiddle = {cos(angle), sin(angle)};
|
||||
output[k].x += input[n].x * twiddle.x - input[n].y * twiddle.y;
|
||||
output[k].y += input[n].x * twiddle.y + input[n].y * twiddle.x;
|
||||
}
|
||||
}
|
||||
return output;
|
||||
}
|
||||
|
||||
static int verifyOutput(const std::vector<float2>& output,
|
||||
const std::vector<float2>& reference,
|
||||
unsigned int N) {
|
||||
int errors = 0;
|
||||
for (unsigned int i = 0; i < N; ++i) {
|
||||
float2 diff = {output[i].x - reference[i].x, output[i].y - reference[i].y};
|
||||
float error = sqrt(diff.x * diff.x + diff.y * diff.y);
|
||||
if (error > 1e-5) {
|
||||
printf("*** error: [%d] expected=(%f,%f), actual=(%f,%f)\n", i, reference[i].x, reference[i].y, output[i].x, output[i].y);
|
||||
++errors;
|
||||
}
|
||||
}
|
||||
return errors;
|
||||
}
|
||||
|
||||
cl_device_id device_id = NULL;
|
||||
cl_context context = NULL;
|
||||
cl_command_queue commandQueue = NULL;
|
||||
cl_program program = NULL;
|
||||
cl_kernel kernel = NULL;
|
||||
cl_mem i_memobj = NULL;
|
||||
cl_mem o_memobj = NULL;
|
||||
uint8_t *kernel_bin = NULL;
|
||||
|
||||
static void cleanup() {
|
||||
if (commandQueue) clReleaseCommandQueue(commandQueue);
|
||||
if (kernel) clReleaseKernel(kernel);
|
||||
if (program) clReleaseProgram(program);
|
||||
if (i_memobj) clReleaseMemObject(i_memobj);
|
||||
if (o_memobj) clReleaseMemObject(o_memobj);
|
||||
if (context) clReleaseContext(context);
|
||||
if (device_id) clReleaseDevice(device_id);
|
||||
if (kernel_bin) free(kernel_bin);
|
||||
}
|
||||
|
||||
int size = 64;
|
||||
|
||||
static void show_usage() {
|
||||
printf("Usage: [-n size] [-h: help]\n");
|
||||
}
|
||||
|
||||
static void parse_args(int argc, char **argv) {
|
||||
int c;
|
||||
while ((c = getopt(argc, argv, "n:h?")) != -1) {
|
||||
switch (c) {
|
||||
case 'n':
|
||||
size = atoi(optarg);
|
||||
break;
|
||||
case 'h':
|
||||
case '?': {
|
||||
show_usage();
|
||||
exit(0);
|
||||
} break;
|
||||
default:
|
||||
show_usage();
|
||||
exit(-1);
|
||||
}
|
||||
}
|
||||
|
||||
printf("Workload size=%d\n", size);
|
||||
}
|
||||
|
||||
int main (int argc, char **argv) {
|
||||
// parse command arguments
|
||||
parse_args(argc, argv);
|
||||
|
||||
cl_platform_id platform_id;
|
||||
size_t kernel_size;
|
||||
|
||||
// Getting platform and device information
|
||||
CL_CHECK(clGetPlatformIDs(1, &platform_id, NULL));
|
||||
CL_CHECK(clGetDeviceIDs(platform_id, CL_DEVICE_TYPE_DEFAULT, 1, &device_id, NULL));
|
||||
|
||||
printf("Create context\n");
|
||||
context = CL_CHECK2(clCreateContext(NULL, 1, &device_id, NULL, NULL, &_err));
|
||||
|
||||
printf("Allocate device buffers\n");
|
||||
size_t nbytes = size * sizeof(float2);
|
||||
i_memobj = CL_CHECK2(clCreateBuffer(context, CL_MEM_READ_ONLY, nbytes, NULL, &_err));
|
||||
o_memobj = CL_CHECK2(clCreateBuffer(context, CL_MEM_WRITE_ONLY, nbytes, NULL, &_err));
|
||||
|
||||
printf("Create program from kernel source\n");
|
||||
#ifdef HOSTGPU
|
||||
if (0 != read_kernel_file("kernel.cl", &kernel_bin, &kernel_size))
|
||||
return -1;
|
||||
program = CL_CHECK2(clCreateProgramWithSource(
|
||||
context, 1, (const char**)&kernel_bin, &kernel_size, &_err));
|
||||
#else
|
||||
if (0 != read_kernel_file("kernel.pocl", &kernel_bin, &kernel_size))
|
||||
return -1;
|
||||
program = CL_CHECK2(clCreateProgramWithBinary(
|
||||
context, 1, &device_id, &kernel_size, (const uint8_t**)&kernel_bin, NULL, &_err));
|
||||
#endif
|
||||
|
||||
// Build program
|
||||
CL_CHECK(clBuildProgram(program, 1, &device_id, NULL, NULL, NULL));
|
||||
|
||||
// Create kernel
|
||||
kernel = CL_CHECK2(clCreateKernel(program, KERNEL_NAME, &_err));
|
||||
|
||||
// Set kernel arguments
|
||||
CL_CHECK(clSetKernelArg(kernel, 0, sizeof(cl_mem), (void *)&i_memobj));
|
||||
CL_CHECK(clSetKernelArg(kernel, 1, sizeof(cl_mem), (void *)&o_memobj));
|
||||
CL_CHECK(clSetKernelArg(kernel, 2, sizeof(int), (void *)&size));
|
||||
|
||||
// Allocate memories for input arrays and output arrays.
|
||||
std::vector<float2> h_i(size);
|
||||
std::vector<float2> h_o(size);
|
||||
|
||||
// Generate input values
|
||||
for (int i = 0; i < size; ++i) {
|
||||
h_i[i].x = sin(2 * M_PI * i / size); // Sine wave as an example
|
||||
h_i[i].y = 0.0f; // Zero imaginary part
|
||||
}
|
||||
|
||||
// Creating command queue
|
||||
commandQueue = CL_CHECK2(clCreateCommandQueue(context, device_id, 0, &_err));
|
||||
|
||||
printf("Upload source buffers\n");
|
||||
CL_CHECK(clEnqueueWriteBuffer(commandQueue, i_memobj, CL_TRUE, 0, nbytes, h_i.data(), 0, NULL, NULL));
|
||||
|
||||
printf("Execute the kernel\n");
|
||||
size_t global_work_size[1] = {size};
|
||||
size_t local_work_size[1] = {LOCAL_SIZE};
|
||||
auto time_start = std::chrono::high_resolution_clock::now();
|
||||
CL_CHECK(clEnqueueNDRangeKernel(commandQueue, kernel, 1, NULL, global_work_size, local_work_size, 0, NULL, NULL));
|
||||
CL_CHECK(clFinish(commandQueue));
|
||||
auto time_end = std::chrono::high_resolution_clock::now();
|
||||
double elapsed = std::chrono::duration_cast<std::chrono::milliseconds>(time_end - time_start).count();
|
||||
printf("Elapsed time: %lg ms\n", elapsed);
|
||||
|
||||
printf("Download destination buffer\n");
|
||||
CL_CHECK(clEnqueueReadBuffer(commandQueue, o_memobj, CL_TRUE, 0, nbytes, h_o.data(), 0, NULL, NULL));
|
||||
|
||||
printf("Verify result\n");
|
||||
std::vector<float2> reference = referenceDFT(h_i);
|
||||
auto errors = verifyOutput(h_o, reference, size);
|
||||
if (0 == errors) {
|
||||
printf("PASSED!\n");
|
||||
} else {
|
||||
printf("FAILED! - %d errors\n", errors);
|
||||
}
|
||||
|
||||
// Clean up
|
||||
cleanup();
|
||||
|
||||
return errors;
|
||||
}
|
||||
Binary file not shown.
Loading…
Add table
Add a link
Reference in a new issue