yum-mirror/VkFFT

git clone https://git.yummers.dev/yum-mirror/VkFFT

Tolmachev DmitriiVkFFT v1.3.2 release3fa0c21

master
branchesdevelopmasterall refs →
1.3 KiB49 linesraw
1//general parts
2#include <stdio.h>
3#include <vector>
4#include <memory>
5#include <string.h>
6#include <chrono>
7#include <thread>
8#include <iostream>
9#ifndef __STDC_FORMAT_MACROS
10#define __STDC_FORMAT_MACROS
11#endif
12#include <inttypes.h>
13
14//CUDA parts
15#include "cuda_runtime.h"
16#include "device_launch_parameters.h"
17#include <cufft.h>
18
19void launch_precision_cuFFT_single(void* inputC, void* output_cuFFT, int device_id, uint64_t* dims)
20{
21	cudaSetDevice(device_id);
22	cufftHandle planC2C;
23	cufftComplex* dataC;
24	cudaMalloc((void**)&dataC, sizeof(cufftComplex) * dims[0] * dims[1] * dims[2]);
25	cudaMemcpy(dataC, inputC, sizeof(cufftComplex) * dims[0] * dims[1] * dims[2], cudaMemcpyHostToDevice);
26	if (cudaGetLastError() != cudaSuccess) {
27		fprintf(stderr, "Cuda error: Failed to allocate\n");
28		return;
29	}
30	switch (dims[4]) {
31	case 1:
32		cufftPlan1d(&planC2C, dims[0], CUFFT_C2C, 1);
33		break;
34	case 2:
35		cufftPlan2d(&planC2C, dims[1], dims[0], CUFFT_C2C);
36		break;
37	case 3:
38		cufftPlan3d(&planC2C, dims[2], dims[1], dims[0], CUFFT_C2C);
39		break;
40	}
41	for (int i = 0; i < 1; i++) {
42		cufftExecC2C(planC2C, dataC, dataC, -1);
43	}
44	cudaDeviceSynchronize();
45	cudaMemcpy(output_cuFFT, dataC, sizeof(cufftComplex) * dims[0] * dims[1] * dims[2], cudaMemcpyDeviceToHost);
46	cudaDeviceSynchronize();
47	cufftDestroy(planC2C);
48	cudaFree(dataC);
49}