#ifndef _GPUTOOL_H_
#define _GPUTOOL_H_
#include "GPUBaseLibAPI.h"
#include "BaseConstVariable.h"

#ifdef __CUDANVCC___
#include <cuda_runtime.h>
#include <device_launch_parameters.h>
#include <cublas_v2.h>
#include <cuComplex.h>

#define __CUDADEBUG__

#define CUDAMEMORY Memory1MB*100

#define LAMP_CUDA_PI 3.141592653589793238462643383279
#define PI4POW2 157.91367041742973

// SHAREMEMORY_FLOAT_HALF_STEP * BLOCK_SIZE = SHAREMEMORY_FLOAT_HALF
/**  CUDA 调用参数 ************************************************************************************/
#define BLOCK_SIZE 256
#define SHAREMEMORY_BYTE 49152
#define SHAREMEMORY_FLOAT_HALF_STEP 2
#define SHAREMEMORY_FLOAT_HALF SHAREMEMORY_FLOAT_HALF_STEP*BLOCK_SIZE

#define SHAREMEMORY_DEM_STEP 768
#define SHAREMEMORY_Reflect 612




enum LAMPGPUDATETYPE {
 LAMP_LONG,
 LAMP_FLOAT,
 LAMP_COMPLEXFLOAT
};

extern "C" struct CUDAVector {
 float x;
 float y;
 float z;
};

extern "C" struct CUDAVectorEllipsoidal {
 float theta;
 float phi;
 float Rho;
};

extern "C" struct CUDAComplex {
 float x;
 float y;
};


//// 定义设备函数
//extern __device__ cuComplex cuCexpf(cuComplex x);
//extern __device__ CUDAVector GPU_VectorAB(CUDAVector A, CUDAVector B);
//extern __device__ float GPU_VectorNorm2(CUDAVector A);
//extern __device__ float GPU_dotVector(CUDAVector A, CUDAVector B);
//extern __device__ float GPU_CosAngle_VectorA_VectorB(CUDAVector A, CUDAVector B);
//
//// 定义全局函数
//extern __global__ void CUDA_DistanceAB(float* Ax, float* Ay, float* Az, float* Bx, float* By, float* Bz, float* R, long len);
//extern __global__ void CUDA_B_DistanceA(float* Ax, float* Ay, float* Az, float Bx, float By, float Bz, float* R, long len);
//extern __global__ void CUDA_make_VectorA_B(float sX, float sY, float sZ, float* tX, float* tY, float* tZ, float* RstX, float* RstY, float* RstZ, long len);
//extern __global__ void CUDA_Norm_Vector(float* Vx, float* Vy, float* Vz, float* R, long len);
//extern __global__ void CUDA_cosAngle_VA_AB(float* Ax, float* Ay, float* Az, float* Bx, float* By, float* Bz, float* anglecos, long len);
//extern __global__ void CUDA_GridPoint_Linear_Interp1(float* v, float* q, float* qv, long xlen, long qlen);
//extern __global__ void CUDA_D_sin(double* y, double* X, int n);
//extern __global__ void CUDA_D_cos(double* y, double* X, int n);
//extern __global__ void CUDAKernel_MemsetBlock(cuComplex* data, cuComplex init0, long len);
//
//extern __global__ void CUDAKernel_MemsetBlock(float* data, float init0, long len);


// 打印GPU参数
extern "C" GPUBASELIBAPI void printDeviceInfo(int deviceId);
// 误差处理函数
extern "C" GPUBASELIBAPI void checkCudaError(cudaError_t err, const char* msg);

// GPU 内存函数
extern "C" GPUBASELIBAPI void* mallocCUDAHost(size_t memsize); // 主机内存声明
extern "C" GPUBASELIBAPI void FreeCUDAHost(void* ptr);
extern "C" GPUBASELIBAPI void* mallocCUDADevice(size_t memsize, int device_id = 0); // GPU内存声明
extern "C" GPUBASELIBAPI void FreeCUDADevice(void* ptr);
extern "C" GPUBASELIBAPI void HostToDevice(void* hostptr, void* deviceptr, size_t memsize);//GPU 内存数据转移 设备 -> GPU
extern "C" GPUBASELIBAPI void DeviceToHost(void* hostptr, void* deviceptr, size_t memsize);//GPU 内存数据转移 GPU -> 设备
extern "C" GPUBASELIBAPI void DeviceToDevice(void* s_deviceptr, void* t_deviceptr, size_t memsize);//GPU 内存数据转移 GPU -> 设备

extern "C" GPUBASELIBAPI void CUDA_MemsetBlock(cuComplex* data, cuComplex init0, long len);
// 矢量基础运算函数
extern "C" GPUBASELIBAPI void CUDAdistanceAB(float* Ax, float* Ay, float* Az, float* Bx, float* By, float* Bz, float* R, long member);
extern "C" GPUBASELIBAPI void CUDABdistanceAs(float* Ax, float* Ay, float* Az, float Bx, float By, float Bz, float* R, long member);
extern "C" GPUBASELIBAPI void CUDAmake_VectorA_B(float sX, float sY, float sZ, float* tX, float* tY, float* tZ, float* RstX, float* RstY, float* RstZ, long member);
extern "C" GPUBASELIBAPI void CUDANorm_Vector(float* Vx, float* Vy, float* Vz, float* R, long member);
extern "C" GPUBASELIBAPI void CUDAcosAngle_VA_AB(float* Ax, float* Ay, float* Az, float* Bx, float* By, float* Bz, float* anglecos, long len);

// 常见插值算法

extern "C" GPUBASELIBAPI void CUDAGridPointLinearInterp1(float* v, float* q, float* qv, long xlen, long qlen);
extern "C" GPUBASELIBAPI void CUDADSin(double* y, double* X, int n);
extern "C" GPUBASELIBAPI void CUDADCos(double* y, double* X, int n);

// 估算分块整数
extern "C" GPUBASELIBAPI long NextBlockPad(long num, long blocksize);


extern "C" GPUBASELIBAPI void PrintLasterError(const char* s);


extern "C" GPUBASELIBAPI void CUDAIFFTScale(cuComplex* inArr, cuComplex* outArr,long InRowCount,long InColCount,long outColCount);

extern "C" GPUBASELIBAPI void CUDAIFFT(cuComplex* inArr, cuComplex* outArr, long InRowCount, long InColCount, long outColCount);

extern "C" GPUBASELIBAPI void FFTShift1D(cuComplex* d_data, int batch_size, int signal_length);
extern "C" GPUBASELIBAPI void shared_complexPtrToHostCuComplex(std::complex<double>* src, cuComplex* dst, size_t len);


extern "C" GPUBASELIBAPI void HostCuComplexToshared_complexPtr(cuComplex* src, std::complex<double>* dst, size_t len);


#endif
#endif