Page MenuHomec4science
Files F65309556

grid_map_mass_GPU.cu
No OneTemporary
Actions

File Metadata

Created
Sun, Jun 2, 20:01

grid_map_mass_GPU.cu
View Options

/**
Lenstool-HPC: HPC based massmodeling software and Lens-map generation
Copyright (C) 2017 Christoph Schaefer, EPFL (christophernstrerne.schaefer@epfl.ch), Gilles Fourestey (gilles.fourestey@epfl.ch)
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
@brief: Function for mass computation over a grid
*/
#include <fstream>
#include "grid_map_mass_GPU.cuh"
#include "gradient2_GPU.cuh"
#include <structure_hpc.hpp>
#define BLOCK_SIZE_X 32
#define BLOCK_SIZE_Y 16
//#define ROT
#define _SHARED_MEM
#ifdef _SHARED_MEM
#define SHARED __shared__
#warning "shared memory"
extern __shared__ type_t shared[];
#else
#define SHARED
#endif
#define Nx 1
#define Ny 0
#define cudasafe
extern "C"
{
type_t myseconds();
}
//GPU mapping function declaration to change when figured out linkage problems
__global__ void mass_grid_GPU(type_t *ampli,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t mult, type_t ds, type_t dl, type_t h, type_t z,int nbgridcells);
////Map function selection
map_gpu_function_t select_map_mass_function(const struct runmode_param* runmode){
if(runmode->mass == 1){
return &mass_1_grid_CPU_GPU;
}
else if(runmode->mass == 2){
return &mass_2_grid_CPU_GPU;
}
else if(runmode->mass == 3){
return &mass_3_grid_CPU_GPU;
}
else if(runmode->mass == 4){
return &mass_4_grid_CPU_GPU;
}
else{
fprintf(stderr, "ERROR: Mass mode %d not supported yet \n",runmode->mass);
exit(-1);
}
return 0;
}
////Mass Map calculation, doesnt fit the bloody template...
void map_grid_mass_GPU(map_gpu_function_t mapfunction, type_t *map, const struct cosmo_param *cosmo, const struct grid_param *frame, const struct Potential_SOA *lens, int nhalos ,int nbgridcells,int mode_amp, type_t zl, type_t zs )
{
type_t dx = (frame->xmax - frame->xmin)/(nbgridcells - 1);
type_t dy = (frame->ymax - frame->ymin)/(nbgridcells - 1);
//
map_grid_mass_GPU(mapfunction,map,cosmo, frame, lens, nhalos,mode_amp,zl, zs, dx, dy, nbgridcells, nbgridcells, 0, 0);
}
//
void map_grid_mass_GPU(map_gpu_function_t mapfunction, type_t *map,const struct cosmo_param *cosmo, const struct grid_param *frame, const struct Potential_SOA *lens, int nhalos, int mode_amp, type_t zl, type_t zs, type_t dx, type_t dy, int nbgridcells_x, int nbgridcells_y, int istart, int jstart)
{
grid_param *frame_gpu;
Potential_SOA *lens_gpu,*lens_kernel;
int *type_gpu;
type_t *lens_x_gpu, *lens_y_gpu, *b0_gpu, *angle_gpu, *epot_gpu, *rcore_gpu, *rcut_gpu, *anglecos_gpu, *anglesin_gpu, *dlsds_gpu;
type_t *grid_grad2_a_gpu, *grid_grad2_b_gpu , *grid_grad2_c_gpu, *grid_grad2_d_gpu, *map_gpu;
if(zl == 0) {
zl = lens->z[0];
}
type_t dl0s = module_cosmodistances_objectObject(zl, zs, *cosmo);
type_t dos = module_cosmodistances_observerObject(zs, *cosmo);
type_t dol = module_cosmodistances_observerObject(zl, *cosmo);
//select_ratio_function(std::string mode, const struct runmode_param* runmode, type_t dls, type_t ds);
lens_gpu = (Potential_SOA *) malloc(sizeof(Potential_SOA));
lens_gpu->type = (int *) malloc(sizeof(int));
// Allocate variables on the GPU
cudasafe(cudaMalloc( (void**)&(lens_kernel), sizeof(Potential_SOA)),"Gradient2gpu.cu : Alloc Potential_SOA: " );
cudasafe(cudaMalloc( (void**)&(type_gpu), nhalos*sizeof(int)),"Gradient2gpu.cu : Alloc type_gpu: " );
cudasafe(cudaMalloc( (void**)&(lens_x_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc x_gpu: " );
cudasafe(cudaMalloc( (void**)&(lens_y_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc y_gpu: " );
cudasafe(cudaMalloc( (void**)&(b0_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc b0_gpu: " );
cudasafe(cudaMalloc( (void**)&(angle_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc angle_gpu: " );
cudasafe(cudaMalloc( (void**)&(epot_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc epot_gpu: " );
cudasafe(cudaMalloc( (void**)&(rcore_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc rcore_gpu: " );
cudasafe(cudaMalloc( (void**)&(rcut_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc rcut_gpu: " );
cudasafe(cudaMalloc( (void**)&(anglecos_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc anglecos_gpu: " );
cudasafe(cudaMalloc( (void**)&(anglesin_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc anglesin_gpu: " );
cudasafe(cudaMalloc( (void**)&(dlsds_gpu), nhalos*sizeof(type_t)),"Gradient2gpu.cu : Alloc dlsds_gpu: " );
cudasafe(cudaMalloc( (void**)&(frame_gpu), sizeof(grid_param)),"Gradient2gpu.cu : Alloc frame_gpu: " );
cudasafe(cudaMalloc( (void**)&(grid_grad2_a_gpu), (nbgridcells_x) * (nbgridcells_y) *sizeof(type_t)),"Gradient2gpu.cu : Alloc source_a_gpu: " );
cudasafe(cudaMalloc( (void**)&(grid_grad2_b_gpu), (nbgridcells_x) * (nbgridcells_y) *sizeof(type_t)),"Gradient2gpu.cu : Alloc source_b_gpu: " );
cudasafe(cudaMalloc( (void**)&(grid_grad2_c_gpu), (nbgridcells_x) * (nbgridcells_y) *sizeof(type_t)),"Gradient2gpu.cu : Alloc source_c_gpu: " );
cudasafe(cudaMalloc( (void**)&(grid_grad2_d_gpu), (nbgridcells_x) * (nbgridcells_y) *sizeof(type_t)),"Gradient2gpu.cu : Alloc source_d_gpu: " );
cudasafe(cudaMalloc( (void**)&(map_gpu), (nbgridcells_x) * (nbgridcells_y) *sizeof(type_t)),"Gradient2gpu.cu : Alloc map: " );
// Copy values to the GPU
//
cudasafe(cudaMemcpy(type_gpu,lens->type , nhalos*sizeof(int),cudaMemcpyHostToDevice ),"Gradient2gpu.cu : Copy type_gpu: " );
cudasafe(cudaMemcpy(lens_x_gpu,lens->position_x , nhalos*sizeof(type_t),cudaMemcpyHostToDevice ),"Gradient2gpu.cu : Copy x_gpu: " );
cudasafe(cudaMemcpy(lens_y_gpu,lens->position_y , nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy y_gpu: " );
cudasafe(cudaMemcpy(b0_gpu,lens->b0 , nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2pu.cu : Copy b0_gpu: " );
cudasafe(cudaMemcpy(angle_gpu,lens->ellipticity_angle , nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy angle_gpu: " );
cudasafe(cudaMemcpy(epot_gpu, lens->ellipticity_potential, nhalos*sizeof(type_t),cudaMemcpyHostToDevice ),"Gradient2gpu.cu : Copy epot_gpu: " );
cudasafe(cudaMemcpy(rcore_gpu, lens->rcore, nhalos*sizeof(type_t),cudaMemcpyHostToDevice ),"Gradient2gpu.cu : Copy rcore_gpu: " );
cudasafe(cudaMemcpy(rcut_gpu, lens->rcut, nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy rcut_gpu: " );
cudasafe(cudaMemcpy(anglecos_gpu, lens->anglecos, nhalos*sizeof(type_t),cudaMemcpyHostToDevice ),"Gradient2gpu.cu : Copy anglecos: " );
cudasafe(cudaMemcpy(anglesin_gpu, lens->anglesin, nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy anglesin: " );
cudasafe(cudaMemcpy(dlsds_gpu, lens->dlsds, nhalos*sizeof(type_t), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy dlsds: " );
cudasafe(cudaMemcpy(frame_gpu, frame, sizeof(grid_param), cudaMemcpyHostToDevice),"Gradient2gpu.cu : Copy fame_gpu: " );
//
lens_gpu->type = type_gpu;
lens_gpu->position_x = lens_x_gpu;
lens_gpu->position_y = lens_y_gpu;
lens_gpu->b0 = b0_gpu;
lens_gpu->ellipticity_angle = angle_gpu;
lens_gpu->ellipticity_potential = epot_gpu;
lens_gpu->rcore = rcore_gpu;
lens_gpu->rcut = rcut_gpu;
lens_gpu->anglecos = anglecos_gpu;
lens_gpu->anglesin = anglesin_gpu;
lens_gpu->dlsds = dlsds_gpu;
//
cudaMemcpy(lens_kernel, lens_gpu, sizeof(Potential_SOA), cudaMemcpyHostToDevice);
//
type_t time = -myseconds();
//
#if 0
int shalos = 0;
while (shalos < nhalos)
{
int lens_type = lens->type[shalos];
type_t z = lens->z[shalos];
int count = 1;
if (shalos + count < nhalos){
std::cerr << shalos << " " << count << " " << lens->type[shalos + count] << " " << lens->z[shalos + count] << " " << std::endl;
while (lens->type[shalos + count] == lens_type and lens->z[shalos + count] == z ){
count++;
if(shalos + count >= nhalos)
break;
//std::cerr << shalos << " " << count << " " << lens->type[shalos + count] << " " << lens->z[shalos + count] << " " << std::endl;
}
std::cerr << shalos << " " << count << " " << lens->type[shalos + count] << " " << lens->z[shalos + count] << " " << std::endl;
}
//if (shalos < nhalos) std::cerr << shalos << " " << count << " " << lens->type[shalos + count] << " " << lens->z[shalos + count] << " " << std::endl;
shalos += count;
}
#endif
//module_potentialDerivatives_totalGradient2_SOA_CPU_GPU(grid_grad2_a_gpu, grid_grad2_b_gpu, grid_grad2_c_gpu, grid_grad2_d_gpu, frame_gpu, lens_kernel, nhalos, dx, dy, nbgridcells_x, nbgridcells_y, istart, jstart);
module_potentialDerivatives_Kmap_SOA_CPU_GPU(grid_grad2_a_gpu, grid_grad2_b_gpu, grid_grad2_c_gpu, grid_grad2_d_gpu, frame_gpu, lens_kernel, nhalos, dx, dy, nbgridcells_x, nbgridcells_y, istart, jstart);
//
cudasafe(cudaGetLastError(), "module_potentialDerivative_totalGradient_SOA_CPU_GPU");
//std::cerr << "ZS " << zs << " "<< dl0s << " "<< dos <<std::endl;
mapfunction(map_gpu,grid_grad2_a_gpu, grid_grad2_b_gpu, grid_grad2_c_gpu, grid_grad2_d_gpu,dl0s,dos,dol,cosmo->h,zs,nbgridcells_x,nbgridcells_y,frame);
cudasafe(cudaGetLastError(), "module_potentialDerivative_totalGradient_SOA_CPU_GPU");
cudaDeviceSynchronize();
//
cudasafe(cudaMemcpy( map, map_gpu, (nbgridcells_x)*(nbgridcells_y)*sizeof(type_t), cudaMemcpyDeviceToHost )," --- Gradient2gpu.cu : Copy source_a_gpu: " );
//
time += myseconds();
std::cout << " kernel time = " << time << " s." << std::endl;
//printf("-----> %f %f \n",grid_grad_x[Nx], grid_grad_y[Ny]);
// Free GPU memory
cudaFree(lens_kernel);
cudaFree(type_gpu);
cudaFree(lens_x_gpu);
cudaFree(lens_y_gpu);
cudaFree(b0_gpu);
cudaFree(angle_gpu);
cudaFree(epot_gpu);
cudaFree(rcore_gpu);
cudaFree(rcut_gpu);
cudaFree(anglecos_gpu);
cudaFree(anglesin_gpu);
cudaFree(dlsds_gpu);
cudaFree(grid_grad2_a_gpu);
cudaFree(grid_grad2_b_gpu);
cudaFree(grid_grad2_c_gpu);
cudaFree(grid_grad2_d_gpu);
cudaFree(map_gpu);
}
//Mass NR 1
void mass_1_grid_CPU_GPU(type_t *map,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t dl0s, type_t ds, type_t dl, type_t h, type_t z, int nbgridcells_x, int nbgridcells_y, const struct grid_param *frame)
{
int GRID_SIZE_X = (nbgridcells_x + BLOCK_SIZE_X - 1)/BLOCK_SIZE_X; // number of blocks
int GRID_SIZE_Y = (nbgridcells_y + BLOCK_SIZE_Y - 1)/BLOCK_SIZE_Y;
//
//printf("grid_size_x = %d, grid_size_y = %d, nbgridcells_x = %d, nbgridcells_y = %d, istart = %d, jstart = %d (split)\n", GRID_SIZE_X, GRID_SIZE_Y, nbgridcells_x, nbgridcells_y, istart, jstart);
//
dim3 threads(BLOCK_SIZE_X, BLOCK_SIZE_Y/1);
dim3 grid (GRID_SIZE_X , GRID_SIZE_Y);
//
//printf("nhalos = %d, size of shared memory = %lf (split)\n", nhalos, (type_t) (8*nhalos + BLOCK_SIZE_X*BLOCK_SIZE_Y)*sizeof(type_t));
//
cudaMemset(map, 0, nbgridcells_x*nbgridcells_y*sizeof(type_t));
//
mass_grid_GPU<<<grid, threads>>> (map,grid_grad2_a, grid_grad2_b,grid_grad2_c, grid_grad2_d,1, dl0s,ds,h,z,nbgridcells_x);
cudasafe(cudaGetLastError(), "amplif_grid_CPU_GPU");
//
cudaDeviceSynchronize();
printf("GPU kernel done...\n");
}
//Mass NR 2
void mass_2_grid_CPU_GPU(type_t *map,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t dl0s, type_t ds, type_t dl, type_t h, type_t z, int nbgridcells_x, int nbgridcells_y, const struct grid_param *frame)
{
int GRID_SIZE_X = (nbgridcells_x + BLOCK_SIZE_X - 1)/BLOCK_SIZE_X; // number of blocks
int GRID_SIZE_Y = (nbgridcells_y + BLOCK_SIZE_Y - 1)/BLOCK_SIZE_Y;
//
//printf("grid_size_x = %d, grid_size_y = %d, nbgridcells_x = %d, nbgridcells_y = %d, istart = %d, jstart = %d (split)\n", GRID_SIZE_X, GRID_SIZE_Y, nbgridcells_x, nbgridcells_y, istart, jstart);
type_t dcrit = cH4piG * h / dl / dl0s * ds; // in g/cm^2
//
dim3 threads(BLOCK_SIZE_X, BLOCK_SIZE_Y/1);
dim3 grid (GRID_SIZE_X , GRID_SIZE_Y);
//
//printf("nhalos = %d, size of shared memory = %lf (split)\n", nhalos, (type_t) (8*nhalos + BLOCK_SIZE_X*BLOCK_SIZE_Y)*sizeof(type_t));
//
cudaMemset(map, 0, nbgridcells_x*nbgridcells_y*sizeof(type_t));
//
mass_grid_GPU<<<grid, threads>>> (map,grid_grad2_a, grid_grad2_b,grid_grad2_c, grid_grad2_d,dcrit, dl0s,ds,h,z,nbgridcells_x);
cudasafe(cudaGetLastError(), "amplif_grid_CPU_GPU");
//
cudaDeviceSynchronize();
printf("GPU kernel done...\n");
}
//Mass NR 3
void mass_3_grid_CPU_GPU(type_t *map,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t dl0s, type_t ds, type_t dl, type_t h, type_t z, int nbgridcells_x, int nbgridcells_y, const struct grid_param *frame)
{
int GRID_SIZE_X = (nbgridcells_x + BLOCK_SIZE_X - 1)/BLOCK_SIZE_X; // number of blocks
int GRID_SIZE_Y = (nbgridcells_y + BLOCK_SIZE_Y - 1)/BLOCK_SIZE_Y;
//
type_t dx = (frame->xmax - frame->xmin)/(nbgridcells_x - 1);
type_t dy = (frame->ymax - frame->ymin)/(nbgridcells_y - 1);
//printf("grid_size_x = %d, grid_size_y = %d, nbgridcells_x = %d, nbgridcells_y = %d, istart = %d, jstart = %d (split)\n", GRID_SIZE_X, GRID_SIZE_Y, nbgridcells_x, nbgridcells_y, istart, jstart);
type_t conv = MCRIT12 / h * dx * dy * dl / dl0s * ds; // in 10^12 M_sol/pixel
//
dim3 threads(BLOCK_SIZE_X, BLOCK_SIZE_Y/1);
dim3 grid (GRID_SIZE_X , GRID_SIZE_Y);
//
//printf("nhalos = %d, size of shared memory = %lf (split)\n", nhalos, (type_t) (8*nhalos + BLOCK_SIZE_X*BLOCK_SIZE_Y)*sizeof(type_t));
//
cudaMemset(map, 0, nbgridcells_x*nbgridcells_y*sizeof(type_t));
//
mass_grid_GPU<<<grid, threads>>> (map,grid_grad2_a, grid_grad2_b,grid_grad2_c, grid_grad2_d,conv, dl0s,ds,h,z,nbgridcells_x);
cudasafe(cudaGetLastError(), "amplif_grid_CPU_GPU");
//
cudaDeviceSynchronize();
printf("GPU kernel done...\n");
}
//Mass NR 4
void mass_4_grid_CPU_GPU(type_t *map,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t dl0s, type_t ds, type_t dl, type_t h, type_t z, int nbgridcells_x, int nbgridcells_y, const struct grid_param *frame)
{
int GRID_SIZE_X = (nbgridcells_x + BLOCK_SIZE_X - 1)/BLOCK_SIZE_X; // number of blocks
int GRID_SIZE_Y = (nbgridcells_y + BLOCK_SIZE_Y - 1)/BLOCK_SIZE_Y;
//
type_t dcritA = cH0_4piG * h / dl / dl0s * ds; // in 10^12 M_sol/kpc^2
//
dim3 threads(BLOCK_SIZE_X, BLOCK_SIZE_Y/1);
dim3 grid (GRID_SIZE_X , GRID_SIZE_Y);
//
//printf("nhalos = %d, size of shared memory = %lf (split)\n", nhalos, (type_t) (8*nhalos + BLOCK_SIZE_X*BLOCK_SIZE_Y)*sizeof(type_t));
//
cudaMemset(map, 0, nbgridcells_x*nbgridcells_y*sizeof(type_t));
//
mass_grid_GPU<<<grid, threads>>> (map,grid_grad2_a, grid_grad2_b,grid_grad2_c, grid_grad2_d,dcritA, dl0s,ds,h,z,nbgridcells_x);
cudasafe(cudaGetLastError(), "amplif_grid_CPU_GPU");
//
cudaDeviceSynchronize();
printf("GPU kernel done...\n");
}
//
__global__ void mass_grid_GPU(type_t *ampli,type_t *grid_grad2_a,type_t *grid_grad2_b,type_t *grid_grad2_c,type_t *grid_grad2_d, type_t mult, type_t dls, type_t ds,type_t h, type_t z,int nbgridcells)
{
////
int col = blockIdx.x*blockDim.x + threadIdx.x;
int row = blockIdx.y*blockDim.y + threadIdx.y;
//type_t dlsds = dls / ds ;
//
if ((row < nbgridcells) && (col < nbgridcells))
{
int index = row*nbgridcells + col;
//if(col == 0 and row == 0)printf(" MASS GPU %f %f\n",grid_grad2_a[index], grid_grad2_c[index]);
type_t ga1 = (grid_grad2_a[index] + grid_grad2_c[index]) * 0.5 ;//* dlsds ;
ampli[index] = ga1*mult;
}
}

Event Timeline

c4science · Help