Files
LBPM/cuda/exe/lb1_MRT-swap.cu
2020-10-07 12:31:49 -04:00

249 lines
8.4 KiB
Plaintext

#include <stdio.h>
#include <iostream>
#include <fstream>
#include <cuda.h>
//#include <cutil.h>
using namespace std;
//*************************************************************************
extern "C" void dvc_InitD3Q19(char *ID, double *f_even, double *f_odd, int Nx,
int Ny, int Nz, int nblocks, int nthreads, int S);
//*************************************************************************
extern "C" void dvc_SwapD3Q19(char *ID, double *f_even, double *f_odd, int Nx,
int Ny, int Nz, int nblocks, int nthreads, int S);
//*************************************************************************
extern "C" void dvc_MRT(char *ID, double *f_even, double *f_odd, double rlxA, double rlxB, double Fx, double Fy, double Fz,
int Nx, int Ny, int Nz, int nblocks, int nthreads, int S);
//*************************************************************************
void Write_Out(double *array, int Nx, int Ny, int Nz){
int value;
FILE *output;
output = fopen("dist.list","w");
for (int k=0; k<Nz; k++){
for (int j=0; j<Ny; j++){
for (int i=0; i<Nx; i++){
int index = k*Nx*Ny+j*Nx+i;
value = int(array[index]);
fprintf(output, "| %i",value);
}
fprintf(output, " | \n");
}
fprintf(output,"************************************** \n");
}
fclose(output);
}
//**************************************************************************
// MRT implementation of the LBM using CUDA
//**************************************************************************
int main(void)
{
int deviceCount;
cudaGetDeviceCount(&deviceCount);
int device = 1;
printf("Number of devices = %i \n", deviceCount);
printf("Current device is = %i \n", device);
cudaSetDevice(device);
// BGK Model parameters
string FILENAME;
unsigned int nBlocks, nthreads;
int timestepMax, interval;
double tau,Fx,Fy,Fz,tol;
// Domain variables
int Nx,Ny,Nz;
ifstream input("MRT.in");
input >> FILENAME; // name of the input file
input >> Nz; // number of nodes (x,y,z)
input >> nBlocks;
input >> nthreads;
input >> tau; // relaxation time
input >> Fx; // External force components (x,y,z)
input >> Fy;
input >> Fz;
input >> timestepMax; // max no. of timesteps
input >> interval; // error interval
input >> tol; // error tolerance
double rlx_setA = 1.f/tau;
double rlx_setB = 8.f*(2.f-rlx_setA)/(8.f-rlx_setA);
printf("tau = %f \n", tau);
printf("Set A = %f \n", rlx_setA);
printf("Set B = %f \n", rlx_setB);
printf("Force(x) = %f \n", Fx);
printf("Force(y) = %f \n", Fy);
printf("Force(z) = %f \n", Fz);
Nx = Ny = Nz; // Cubic domain
int N = Nx*Ny*Nz;
int dist_mem_size = N*sizeof(double);
// unsigned int nBlocks = 32;
// int nthreads = 128;
int S = N/nthreads/nBlocks;
// unsigned int nBlocks = N/nthreads + (N%nthreads == 0?0:1);
dim3 grid(nBlocks,1,1);
printf("Number of blocks = %i \n", nBlocks);
printf("Threads per block = %i \n", nthreads);
printf("Sweeps per thread = %i \n", S);
printf("Number of nodes per side = %i \n", Nx);
printf("Total Number of nodes = %i \n", N);
//.......................................................................
printf("Read input media... \n");
// .......... READ THE INPUT FILE .......................................
int n;
char value;
char *id;
id = new char[N];
int sum = 0;
double porosity;
ifstream PM(FILENAME.c_str(),ios::binary);
for (int k=0;k<Nz;k++){
for (int j=0;j<Ny;j++){
for (int i=0;i<Nx;i++){
PM.read((char *) (&value), sizeof(value));
n = k*Nx*Ny+j*Nx+i;
id[n] = value;
if (value > 0) sum++;
}
}
}
PM.close();
printf("File porosity = %f\n", double(sum)/N);
//.......................................................................
//...........device phase ID.................................................
char *ID;
cudaMalloc((void **) &ID, N); // Allocate device memory
// Copy to the device
cudaMemcpy(ID, id, N, cudaMemcpyHostToDevice);
//...........................................................................
//......................device distributions.................................
double *f_even,*f_odd;
//...........................................................................
cudaMalloc((void **) &f_even, 10*dist_mem_size); // Allocate device memory
cudaMalloc((void **) &f_odd, 9*dist_mem_size); // Allocate device memory
//...........................................................................
//...........................................................................
// cudaHostAlloc(&fa,dist_mem_size,cudaHostAllocPortable);
// cudaHostAlloc(&fb,dist_mem_size,cudaHostAllocPortable);
// cudaHostRegister(fa,dist_mem_size,cudaHostRegisterPortable);
// cudaHostRegister(fb,dist_mem_size,cudaHostRegisterPortable);
// cudaHostRegister(id,N*sizeof(char),cudaHostAllocPortable);
printf("Setting the distributions, size = : %i\n", N);
//...........................................................................
// INITIALIZE <<< grid, nthreads >>> (ID, f_even, f_odd, Nx, Ny, Nz, S);
//...........................................................................
dvc_InitD3Q19(ID,f_even,f_odd,Nx,Ny,Nz,nBlocks,nthreads,S);
//*************************************************************************
int timestep = 0;
printf("No. of timesteps: %i \n", timestepMax);
//.......create a stream for the LB calculation.......
cudaStream_t stream;
cudaStreamCreate(&stream);
//.......create and start timer............
cudaEvent_t start, stop;
float time;
cudaEventCreate(&start);
cudaEventCreate(&stop);
cudaEventRecord( start, 0 );
//.........................................
//************ MAIN ITERATION LOOP ***************************************/
while (timestep < timestepMax){
//...................................................................
//........ Execute the swap kernel (device) .........................
// SWAP <<< grid, nthreads >>> (ID, f_even, f_odd, Nx, Ny, Nz, S);
//...................................................................
dvc_SwapD3Q19(ID,f_even,f_odd,Nx,Ny,Nz,nBlocks,nthreads,S);
//........ Execute the collision kernel (device) ....................
// MRT <<< grid, nthreads >>> (ID, f_even, f_odd, Nx, Ny, Nz, S,
// rlx_setA, rlx_setB, Fx, Fy, Fz);
//............................................................
dvc_MRT(ID, f_even, f_odd, rlx_setA, rlx_setB, Fx, Fy, Fz,Nx,Ny,Nz,nBlocks,nthreads,S);
// Iteration completed!
timestep++;
//...................................................................
}
//************************************************************************/
cudaThreadSynchronize();
//.......... stop and destroy timer.............................
cudaEventRecord( stop, stream);
cudaEventSynchronize( stop );
cudaEventElapsedTime( &time, start, stop );
printf("CPU time = %f \n", time);
float MLUPS = 0.001*float(Nx*Ny*Nz)*timestep/time;
printf("MLUPS = %f \n", MLUPS);
cudaStreamDestroy(stream);
cudaEventDestroy( start );
cudaEventDestroy( stop );
//..............................................................
//..............................................................
//.........Compute the velocity and copy result to host ........
double *velocity;
velocity = new double[3*N];
//......................device distributions....................................
double *vel;
//..............................................................................
cudaMalloc((void **) &vel, 3*dist_mem_size); // Allocate device memory
//..............................................................................
// Compute_VELOCITY <<< grid, nthreads >>> (ID, f_even, f_odd, vel, Nx, Ny, Nz, S);
//..............................................................................
cudaMemcpy(velocity, vel, 3*dist_mem_size, cudaMemcpyDeviceToHost);
//..............................................................................
//............................................................
//....Write the z-velocity to test poiseuille flow............
double vz,vz_avg;
vz_avg = 0.0;
FILE *output;
output = fopen("velocity.out","w");
for (int k=0; k<1; k++){
for (int j=0; j<1; j++){
for (int i=0; i<Nx; i++){
int n = k*Nx*Ny+j*Nx+i;
//.....print value........
vz = velocity[2*N+n];
vz_avg += vz;
fprintf(output, " %e",vz);
}
}
}
fclose(output);
vz = vz_avg/double(sum);
printf("Average Velocity = %e\n", vz);
// cleanup
cudaFree(f_even); cudaFree(f_odd); cudaFree(vel); cudaFree(ID);
free (velocity); free(id);
}