serial_no
int64
1
24.2k
cuda_source
stringlengths
11
9.01M
17,301
#include <stdio.h> #include <cuda_runtime.h> #include <cuda.h> #include <stdlib.h> #include "device_launch_parameters.h" #include <thrust/scan.h> /* __global__ void dfs_parallel(int *d_frontier1,int *d_vertex,int *d_loc,int *d_edge,int *d_frontier2) { int i = blockIdx.x * gridDim.y + blockIdx.y,j=i*blockDim.x*blockDim...
17,302
#include<cuda.h> #include<stdio.h> #include<math.h> #include<ctime> __global__ void vecMulMatrixKernel(float* A, float* B, float* C, int n){ // clock_t start = clock(); int column = threadIdx.x + blockDim.x * blockIdx.x; int row = threadIdx.y + blockDim.y * blockIdx.y; //printf("%d ",blockDim.x); if(row<n && column...
17,303
#include "includes.h" #define NUMBER_OF_BLOCKS 256 #define NUMBER_OF_THREADS 64 // ========== // Macro taken from: // https://stackoverflow.com/questions/14038589/what-is-the-canonical-way-to-check-for-errors-using-the-cuda-runtime-api __device__ double dotProduct(double *a, double *b, int size) { double result = 0; ...
17,304
/* Now we make the matrix much bigger g++ -pg seq_matrix_big_mul.c -o seq_matrix_big_mul */ #include <stdio.h> #include <string.h> #include <stdlib.h> #include <time.h> #define N_THREADS 32 int num_rows_A = 2000; int num_rows_B = 2000; int num_rows_C = 2000; int num_cols_A = 2000; int num_cols_B = 600; int num_co...
17,305
// Using different memory spaces in CUDA #include <stdio.h> /********************** * using local memory * **********************/ // a __device__ or __global__ function runs on the GPU __global__ void use_local_memory_GPU(float in) { float f; // variable "f" is in local memory and private to each thread ...
17,306
#include<stdio.h> #define TBP 256 __global__ void hello_world() { printf("Hello World! My threadId is %d\n",threadIdx.x); __syncthreads(); } int main() { hello_world<<<1,TBP>>>(); cudaDeviceSynchronize(); return 0; }
17,307
#include<iostream> #include<stdio.h> #include<stdlib.h> #include <cuda.h> #include <math.h> #define BLOCK_SIZE 1024 // Kernel for the first iteration of parallel scan __global__ void parallelScan(float *d_out, float *d_in, int length) { volatile extern __shared__ double sharedData[]; int tid = threadIdx.x + blo...
17,308
/* Swap the elements of a vector: the first with the last and so on... */ #include<cuda.h> #include<stdlib.h> #include<stdio.h> #include<sys/time.h> void checkCUDAError(const char* msg); __global__ void rebalta (float *dati, int n) { int id; int t; id=blockIdx.x*blockDim.x+threadIdx.x; t=dati[n-id-1]; da...
17,309
#include "includes.h" __global__ void cunnx_BlockSparse_updateGradOutput_kernel( float *_gradOutput, float* gradOutputScale, const float *gradOutput, const float *output, const float *outputScale, int outputWindowSize, int outputSize) { __shared__ float buffer[BLOCKSPARSE_THREADS]; int tx = threadIdx.x; int i_step = bl...
17,310
/* * Copyright (c) 2009 Steve Worley < m a t h g e e k@(my last name).com > * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PR...
17,311
#include "includes.h" using namespace std; // GPU Code // __global__ indicates that it is a GPU kernel, that can be called from the CPU // CPU Code __global__ void Add(float* d_a, float* d_b, float* d_c, int N) { int id = blockIdx.x * blockDim.x + threadIdx.x; if(id < N) d_c[id] = d_a[id] + d_b[id]; }
17,312
#include <stdio.h> // Matrices are stored in row-major order: // M(row, col) = *(M.elements + row * M.width + col) typedef struct { int width; int height; float* elements; } Matrix; void print_matrix(const Matrix mat) { for(int r=0; r<mat.height; ++r) { for(int c=0; c<mat.width; ++c) { ...
17,313
#define COALESCED_NUM 16 #define blockDimX 256 #define blockDimY 1 #define gridDimX (gridDim.x) #define gridDimY (gridDim.y) #define idx (blockIdx.x*blockDimX+threadIdx.x) #define idy (blockIdx.y*blockDimY+threadIdx.y) #define bidy (blockIdx.y) #define bidx (blockIdx.x) #define tidx (threadIdx.x) #define tidy (threadId...
17,314
#include <iostream> #include <vector> #include <math.h> #include <assert.h> #include <memory> #include <random> #include "gMat.cuh" #include "gpuerrchk.cuh" #include "real.h" __global__ void matMulKernel(real* A, real* B, real* P, int m, int n, int s, int tile_size){ //each thread in the block will be responsible for...
17,315
#include "includes.h" __global__ void sobelEdgeDetectionSharedMem(int *input, int *output, int width, int height, int thresh) { int blockSize = 32; static __shared__ int shMem[34][34]; int i = blockIdx.x * blockDim.x + threadIdx.x; int j = blockIdx.y * blockDim.y + threadIdx.y; int index = j * width + i; int xind = ...
17,316
#include "includes.h" cudaError_t addWithCuda(int *c, const int *a, const int *b, unsigned int size); __global__ void global_scan(float* d_out, float* d_in) { int index = threadIdx.x; float out = 0.00f; d_out[index] = d_in[index]; __syncthreads(); for (int i = 1; i < sizeof(d_in); i*=2) { if (index - i >= 0) {...
17,317
#define CHECK(call) \ { \ const cudaError_t error = call; \ if( error != cudaSuccess ) \ ...
17,318
#include <stdio.h> #include <stdlib.h> #include <sys/time.h> #include <unistd.h> double get_time() { struct timeval tv; gettimeofday(&tv, NULL); return (double)tv.tv_sec + (double)1e-6 * tv.tv_usec; } __global__ void vec_add(int *x, int *y, int *z, int n) { int i = blockDim.x * blockIdx.x + threadIdx....
17,319
#include <cuda_runtime.h> #include <stdio.h> #include <stdlib.h> #include <limits.h> #include <math.h> #include <time.h> #define ll long long #define MAX_THREAD_PER_BLOCK 1024 #define MAX_NB_CITIES 50 #define DEBUG 1 __device__ int calc_perm_cost(ll idx, int n, ll nb_perm, int * dist, ll * fact) { // Test for valid ...
17,320
#include <cuda.h> #include <stdio.h> #include <iostream> #include <sys/time.h> using namespace std; const int BLOCK_WIDTH = 16; const int BLOCK_HEIGHT = 16; const int DEFAULT_ELE = 16; typedef struct __align__(8) linkNode { int edge; linkNode* next; } linkNode; //HACK: This will be incredibly slow on CUDA! __dev...
17,321
#include "includes.h" /*********************************************************** By Huahua Wang, the University of Minnesota, twin cities ***********************************************************/ __global__ void vecInit(float* X, unsigned int size, float value) { const unsigned int idx = blockIdx.x * b...
17,322
#include<stdio.h> #include<math.h> #define SIZE 1024 __global__ void sum(int * A, int * C) { int i=blockIdx.x*blockDim.x+threadIdx.x; C[i] =A[2*i+1]+A[2*i]; } /*__global__ void avg(int * A, int * C) { int i=blockIdx.x*blockDim.x+threadIdx.x; A[2*i] < A[2*i+1]?C[i]=A[2*i]:C[i]=A[2*i+1];...
17,323
//__global__ void gpuRecursiveReduce(int *g_idata, int *g_odata, // unsigned int isize) //{ // int tid = threadIdx.x; // // int *idata = g_idata + blockIdx.x*blockDim.x; // int *odata = &g_odata[blockIdx.x]; // // // stop condition // if (isize == 2 && tid == 0) // { // g_odata[blockIdx.x] = idata[0] + idata[1]; //...
17,324
#include "includes.h" __global__ void LoadVec(float *vector , float2 *FFT) { int idx = threadIdx.x + blockIdx.x*blockDim.x; // this should span the full range of the vector FFT[idx].x = vector[idx]; // The real part is replaced by the vector value FFT[idx].y = 0.0f; // The imaginary part is zero. The following k...
17,325
#include<stdio.h> __global__ void hellocuda(int* tidx ){ /* *tidx = 100; */ int x = threadIdx.x; tidx[x]=threadIdx.x; } int main(){ int i; int * d_tidx; int * h_tidx; cudaError_t err = cudaMalloc((void**) &d_tidx, 20*sizeof(int)); if (err != cudaSuccess){ printf("%s on ...
17,326
#include "includes.h" __global__ void scan_workefficient(float *g_odata, float *g_idata, int n) { // Dynamically allocated shared memory for scan kernels extern __shared__ float temp[]; int thid = threadIdx.x; int offset = 1; // Cache the computational window in shared memory temp[2*thid] = g_idata[2*thid]; temp...
17,327
#include <iostream> #include <cuda.h> #include <math.h> #include <stdio.h> typedef unsigned int uint; // O kernel a seguir obtém o índice da thread operante e verifica se ela está apta para manipular o array. // Assim, por meio dele, executamos em paralelo (CUDA Cores * SMs) operações em paralelo, dado o particioname...
17,328
#include <bits/stdc++.h> #include <curand_kernel.h> using namespace std; constexpr int POPULATION_SIZE = 128; constexpr int GENERATIONS = 100; constexpr double MUTATION_RATE = 0.1; //constexpr int MAX_SIZE = 1000; #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); } inline void gpuAssert(cudaError_t...
17,329
#include <stdio.h> #include <stdlib.h> #include <time.h> #include <math.h> #include <cuda_runtime.h> #include <limits.h> #define nums 200 void kthSmallest(int arr[], int k){ for (int i = 0; i < nums; ++i){ int upper_sum = 0,down_sum = 0,pivot = arr[i]; for (int j = 0; j < nums; ++j){ upper_sum += (pivot>a...
17,330
// Dan Rolfe #define BLOCKSIZE 32 // general structure for version 1 requested by prof // section 5.3 of the cuda pogramming guide /** * first load from device mem to shared mem * sync threads after read * do the processing from shared mem * sync threads after processing * write the results back to device mem **/ ...
17,331
/** * CUDA implementation of a fully-connected feed forward neural network * * Authors: Jevgenija Aksjonova (jevaks@kth.se) * Beatrice Ionascu (bionascu@kth.se) * * Last changed: 04/30/2017 */ #include <stdio.h> #include <assert.h> #include <stdlib.h> #include <string.h> #include <math.h> #define TOL ...
17,332
#include "includes.h" __global__ void computeExCovX(float *trans_x, float *trans_y, float *trans_z, int *valid_points, int *starting_voxel_id, int *voxel_id, int valid_points_num, double *centr_x, double *centr_y, double *centr_z, double gauss_d1, double gauss_d2, double *e_x_cov_x, double *icov00, double *icov01, doub...
17,333
// kernel definition #include "cuda_runtime.h" #include "device_launch_parameters.h" __global__ void vecAdd(float* A, float* B, float* C) { int idx = threadIdx.x; C[idx] = A[idx] + B[idx]; } //int main(void) //{ // // //vecAdd << <1, N >> > (A, B, C); //}
17,334
#include "../include/commons.cuh" __host__ void usage() { cout << "./main [cpu/gpu] [number_of_spheres]\n"; exit(EXIT_FAILURE); } __host__ void controls() { cout << "[] Controls:\n"; cout << "|- View Position: A, D, Mouse-Scroll\n"; cout << "|- Light Position: W, S, Q, E\n"; ...
17,335
#include <iostream> #include <iomanip> #include <new> #include <stdio.h> #include <cuda_runtime.h> #include <sys/time.h> #include <limits.h> #define imin(a,b) (a<b?a:b) const int N = 1<<15; const int threadsPerBlock = 256; const int blocksPerGrid = imin(32, (N + threadsPerBlock - 1) / threadsPerBlock); // 计时器函数 doub...
17,336
extern "C" { /* * Kernel for mirroring the image parallely usng CUDA */ __global__ void mirror(const uchar4* const inputChannel, uchar4* outputChannel, int numRows, int numCols, bool vertical) { int col = blockIdx.x * blockDim.x + threadIdx.x; int row = blockIdx.y * blockDim.y + th...
17,337
#include "includes.h" __global__ void sub_mul_kernel(double *g_out, double *a, double *b1, double *b2, double *ct, int n) { const int j2 = blockIdx.x * blockDim.x + threadIdx.x; double wkr, wki, xr, xi, yr, yi, ajr, aji, akr, aki, bjr, bji, bkr, bki; double new_ajr, new_aji, new_akr, new_aki; const int m = n >> 1; cons...
17,338
#include <cuda.h> #include <cuda_runtime_api.h> #include <stdio.h> #include <iostream> #include <string.h> #include <algorithm> #include <stdlib.h> //#define N 4 #define BLOCK_SIZE 4 #define GRID_SIZE 2 // threadfence(); using namespace std; __device__ volatile int Arrayin[100]; __device__ volatile int Arrayout[10...
17,339
#include "includes.h" __device__ inline float3 addCuda(float3 a, float3 b) { return{ a.x + b.x, a.y + b.y, a.z + b.z }; } __device__ inline float3 multiplyCuda(float a, float3 b) { return{ a * b.x, a * b.y, a * b.z }; } __device__ inline float euclideanLenCuda(float3 a, float3 b, float d) { float mod = (b.x - a.x) * (b...
17,340
// return the cuda compute capabiltiy for device 0 on the current hardware // this is not just for convenience, but also directly used by the cmake build #include <stdio.h> #include <stdlib.h> int main() { cudaDeviceProp prop; if (cudaGetDeviceProperties(&prop, 0)) { fprintf(stderr, "Failed to get cuda ...
17,341
#include <cuda.h> #include <stdio.h> #include <dlfcn.h> #include <stdlib.h> CUresult cuDeviceTotalMem(size_t* bytes, CUdevice dev) { void *handle; handle = dlopen("/usr/lib/x86_64-linux-gnu/libcuda.so.1", RTLD_LAZY); printf("%s\n", "I just want to tell you that cuDeviceTotalMem is STILL hijacked!"); ...
17,342
#include <stdlib.h> #include <stdio.h> #include <sys/time.h> #define initTimer struct timeval tv1, tv2; struct timezone tz #define startTimer gettimeofday(&tv1, &tz) #define stopTimer gettimeofday(&tv2, &tz) #define tpsCalcul (tv2.tv_sec-tv1.tv_sec)*1000000L + (tv2.tv_usec-tv1.tv_usec) #define MAX_DIM_GRID 65535 #de...
17,343
#include <ctype.h> #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <sys/times.h> #include <time.h> #include <cuda_runtime.h> #define PI 3.14159265358979323846 #define FactorArcosegRad 0.00000484814 clock_t timestart, timeend; /** @brief Función que transforma un valor en arco segundo a radianes @p...
17,344
// Babak Poursartip // 09/15/2020 // udemy CUDA // memory management in cude // start thread in a multiple of 32 #include <cstdio> #include <cstdlib> #include <time.h> // ================================= __global__ void mem_trs_test(int *input) { // 1d grid, 1d block int gid = blockDim.x * blockIdx.x + thread...
17,345
#include "includes.h" __global__ void cmin(float *d_in, float *min, int len) { extern __shared__ float smin[]; unsigned int tid = threadIdx.x; unsigned int i = blockIdx.x * blockDim.x + threadIdx.x; smin[tid] = d_in[i]<d_in[i+len] ? d_in[i] : d_in[i+len]; __syncthreads(); if(blockDim.x > 512 && tid<512) {if(smin[ti...
17,346
/** @brief Compare vector sum calculation functions in CPU vs GPU. @file 00.cu @author isquicha @version 0.1.0 */ #include <stdio.h> #include <time.h> // Cuda headers are on CUDA Toolkit instalation path/VERSION/include #include "cuda.h" #include "cuda_runtime.h" #include "device_launch_parameters.h" ...
17,347
#include "includes.h" __global__ void zupdate_stencil(float *zx, float *zy, float *zoutx, float *zouty, float *g, float tau, float invlambda, int nx, int ny) { int px = blockIdx.x * blockDim.x + threadIdx.x; int py = blockIdx.y * blockDim.y + threadIdx.y; int idx = px + py*nx; int tidx, tpx, tpy; float a, b, t; float ...
17,348
#include <stdlib.h> //#include <float.h> #include <stdio.h> //#include <string.h> //#include <math.h> #include <time.h> typedef struct { int width; int height; float *elements; } Matrix; void startSeed() { srand(time(NULL)); int seed = rand(); srand(seed); } void draw_random(Matrix mat) { for ...
17,349
/* Single Author info: arajend4 Ayushi Rajendra Kumar */ #include <stdio.h> #include <stdlib.h> #include <math.h> //#include <cuda_runtime.h> /* first grid point */ #define XI 0.0 /* last grid point */ #define XF M_PI typedef double FP_PREC; /* function declarations */ double fn...
17,350
typedef double MYFLOAT; #define PI 3.14159265359 __device__ MYFLOAT gpu_ix2x(int ix, int nx, MYFLOAT lx){ return ((ix) - nx / 2.0)*lx/nx; } __device__ MYFLOAT gpu_iy2y(int iy, int ny, MYFLOAT ly){ return ((iy) - ny / 2.0)*ly/ny; } __global__ void eff_update_up_down(int nx, int ny, MYFLOAT *temp){ int ix; ...
17,351
#include "includes.h" __global__ void middle_to_right(float* data, const int nx, const int ny) { float tmp; for ( int r = 0; r < ny; ++r ) { float last_val = data[r*nx+nx/2]; for ( int c = nx-1; c >= nx/2; --c ){ int idx = r*nx+c; tmp = data[idx]; data[idx] = last_val; last_val = tmp; } } }
17,352
#include <iostream> using namespace std; int main() { return 0; }
17,353
#include<iostream> int main() { cudaDeviceProp prop; int count; cudaGetDeviceCount(&count); std::cout<<"GPU num:"<<count<<std::endl; cudaGetDeviceProperties(&prop,0); std::cout<<"Max threads/block:"<<prop.maxThreadsPerBlock<<std::endl; std::cout<<"Max threads/SM:"<<prop.maxThreadsPerMultiP...
17,354
/** * @file : params_kernelf.cu * @brief : Modified implementation of njuffa's; * CUDA kernel functions as parameters with CUDA C++14, CUDA Unified Memory Management * @details : Modified implementation of njuffa's, * std::function vs. function pointer in C++11, C++14, and now in CUDA * std::fun...
17,355
#include <math.h> #include <stdlib.h> #include <stdio.h> #include "unistd.h" #include "time.h" #include "string.h" #include <cuda_runtime.h> // ---------------------- Optimised Dedispersion Loop ------------------------------ __global__ void testAtomicCas(int *buffer, int nsamp, int factor) { if (blockIdx.x * bl...
17,356
#include "includes.h" __global__ void ComputeOffsetOfMatrixAB(const int32_t* row_sum, const int32_t* col_sum, int32_t* output, int32_t K_A_B, int32_t N) { for (int32_t i = threadIdx.x; i < N; i += blockDim.x) { *(output + blockIdx.x * N + i) = K_A_B - row_sum[blockIdx.x] - col_sum[i]; } }
17,357
#include "includes.h" __global__ void normalise(float* result, unsigned int resultLength, float* amps, unsigned int* hits) { int absoluteThreadIdx = blockDim.x * blockIdx.x + threadIdx.x; if(absoluteThreadIdx > resultLength) return; result[absoluteThreadIdx] = amps[absoluteThreadIdx] / hits[absoluteThreadIdx / 4]; }
17,358
#include <cuda.h> #include <stdlib.h> #include <iomanip> #include <iostream> using std::setw; const int N = 5; const int threadsPerBlock = N; int blocksPerGrid = 1; // device code __global__ void prefixSum(float* x, float* c) { __shared__ float cache[2 * threadsPerBlock]; // declaring a array in shared memor...
17,359
#include <iostream> #include <stdlib.h> #include <ctime> #include <cuda.h> struct float3_t { float x, y, z; }; __global__ void FindClosetGPU (float3_t* points, int* indices, int count) { if (count <= 1) return; int idx = blockIdx.x * blockDim.x + threadIdx.x; if (idx < count) { float3_t thisPoint = points...
17,360
#include <iostream> #include <cstdlib> #include <math.h> #include <chrono> #include <iomanip> #include <fstream> using namespace std; using namespace std::chrono; typedef unsigned long long ULL; ofstream primeresult; ofstream timeresult; __global__ void cuda_erastothenes_sieve (ULL *marked, ULL *limit, ULL *n, int *to...
17,361
#include <stdio.h> #include <time.h> #include <stdlib.h> #include <thrust/generate.h> #include <thrust/random.h> #include <thrust/iterator/counting_iterator.h> #include <thrust/functional.h> #include <thrust/transform_reduce.h> #include <cmath> struct montecarlo : public thrust::unary_function<unsigned int, float> { ...
17,362
#include <stdio.h> #include <string.h> #include <stdlib.h> #include <unistd.h> #define LIST_SIZE 100000 __device__ unsigned long long instCountList[LIST_SIZE]; __device__ int init_flag = 0; extern "C" __device__ void profileCount(long index){ if(init_flag == 0){ int i = 0; for(i=0;i<LIST_SIZE;i++){ inst...
17,363
#include "includes.h" __global__ void getCrossingTimes(double *results, int *crossTimes, int N, int numSims, int lowerThreshold, int upperThreshold) { int tid = threadIdx.x + blockIdx.x * blockDim.x; while (tid < N * numSims) { if (crossTimes[tid/N] == 0) { if (results[tid] <= lowerThreshold) { crossTimes[tid/N] = tid ...
17,364
#include "includes.h" __global__ void InitArrays(float *ip, float *op, float *fp, int *kp, int ncols) { int i; float *fppos, *oppos, *ippos; int *kppos; int blockOffset; int rowStartPos; int colsPerThread; // Each block gets a row, each thread will fill part of a row // Calculate the offset of the row blockOffset = b...
17,365
#include "includes.h" __global__ void device_add(int *a, int *b, int *c) { int blockId = blockIdx.x; if (blockId < arrSize) c[blockId] = a[blockId] + b[blockId]; }
17,366
#include "includes.h" __global__ void prime( int *a, int *b, int *c ) { int tid = (blockIdx.x*blockDim.x) + threadIdx.x; // this thread handles the data at its thread id if (tid < vector_size){ c[tid] = a[tid] + b[tid]; // add vectors together } }
17,367
//#include "device_launch_parameters.h" //#include "cuda_runtime.h" //#include "Box.h" // //#define cudaCheckError() { \ // cudaError_t e=cudaGetLastError(); \ // if(e!=cudaSuccess) { \ // printf("Cuda failure, %s",cudaGetErrorString(e)); \ // exit(0); \ // }\ //} // //__global__ void check(int noOfCubes, box* boxes...
17,368
/* Very simple addition kernel */ __global__ void add_kernel(double *in, int N) { int tid = blockIdx.x * blockDim.x + threadIdx.x; if (tid < N) in[tid]++; } /* Kernel call wrapper, we can not use the <<<>>> syntax in MPI code. Arguments: data (double *) -- pointer to the device memory data...
17,369
#include "includes.h" __global__ void cudaKernel(int *n, int limit) { }
17,370
#ifdef __cplusplus extern "C" { #endif struct point{ float x; float y; }; __global__ void pi(const struct point* A, int* res, const int nbPoint, const float ray){ const int idx = 32*blockDim.x * blockIdx.x + threadIdx.x; if (idx < nbPoint-32*blockDim.x) #pragma unroll 16 for (...
17,371
#include <cstdio> __device__ __forceinline__ float addf(float* input1, float* input2) { return input1[blockIdx.x] + input2[blockIdx.x]; } extern "C" __global__ void add(float c, float* __restrict__ input1, float* __restrict__ input2, float* __restrict__ output) { output[blockIdx.x] = addf(input1, input2) + c; }
17,372
#include <stdio.h> __global__ void compute_data(int *a, int const x, int const n) { int idx = threadIdx.x + blockDim.x*blockIdx.x; if (idx<n) { int aa = a[idx]; int product = 0.0; for(int i = 0; i < x; i++) product += aa; a[idx] = product; } } extern "C" void ext_compute_data(int grid_size, int ...
17,373
#define THETA_N 4 #define SQRT_2 1.4142135623730951f #define PI 3.141592653589793f extern "C" { /** * Clears out the Gabor Energies Tensor, setting all of its values to zero. * The Gabor Energies Tensor is the data structure whose [y, x, theta] value contains the average magnitude response to * the different comple...
17,374
#include "includes.h" __global__ void ComputeSquareDistance(float* dOut, float* dIn, int n, int d) { // Load values that will be reused __shared__ float blockA[KNN_BLOCK_SIZE][KNN_BLOCK_SIZE]; __shared__ float blockB[KNN_BLOCK_SIZE][KNN_BLOCK_SIZE]; // A is responsible for points indexed between aStart and aEnd auto a...
17,375
// ######################################################################## // Practical Course: GPU Programming in Computer Vision // Technical University of Munich, Computer Vision Group // ######################################################################## #include <cuda_runtime.h> #include <iostream> using na...
17,376
#include <stdint.h> #include <stdio.h> #define N 32 #define THREADS_PER_BLOCK 32 __global__ void dotproduct(float* x, float* y, float* result) { // Compute the index this thread should use to access elements size_t index = threadIdx.x + blockIdx.x * THREADS_PER_BLOCK; // Create space for a shared array t...
17,377
#include "includes.h" __global__ void linearLayerForward( float* W, float* A, float* Z, float* b, int W_x_dim, int W_y_dim, int A_x_dim, int A_y_dim) { int row = blockIdx.y * blockDim.y + threadIdx.y; int col = blockIdx.x * blockDim.x + threadIdx.x; int Z_x_dim = A_x_dim; int Z_y_dim = W_y_dim; float Z_value = 0; i...
17,378
/* Finds: size of the read only cache For Maxwell microarchitecture Source code based on paper https://arxiv.org/pdf/1509.02308.pdf Compile with nvcc -arch=sm_52 maxwell_readonly.cu -o readonly (__ldg() intrinsic is only available on compute capability 3.5+ architecture) */ #include <stdio.h> #include <stdin...
17,379
#include "includes.h" //#define array_size 100000000 #define array_size 101 //987459712 cudaError_t addWithCuda(int *total); __shared__ int temp[array_size]; __global__ void addKernel(int *tid_c, int *tid_total) { int tid = blockIdx.x * blockDim.x + threadIdx.x; tid_c[tid] = tid; if (tid <= array_size) { temp...
17,380
///////////////////////////////////////////////////////////////////////// // // // CUDA code which calculates PI using Monte-Carlo method // // It will get random points in the square between (0,0) and (1,1) // // Find whether it is i...
17,381
#include <math.h> #include <stdio.h> #include <stdlib.h> #define THREADS_PER_BLOCK 16 __global__ void set(int *A, int N) { int idx = threadIdx.x + blockIdx.x * blockDim.x; //index saved A[idx] = idx; //A[1] = 1, A[2] = 2, ..., A[N] = N } int main(void) { const i...
17,382
#include "kernels.cuh" struct node { char nodeType; int index; double vr; double dr; int child[2]; bool flag; }; __global__ void build_circuit(struct node** array, int n, int H, int *num) { unsigned int index = threadIdx.x + blockIdx.x*blockDim.x; unsigned int stride = gridDim.x*blockDim.x;...
17,383
#include <thrust/device_vector.h> #include <thrust/host_vector.h> #include <iostream> #include <chrono> #include <thrust/count.h> #include <thrust/functional.h> #include <thrust/iterator/constant_iterator.h> #include <thrust/transform_reduce.h> #include <math.h> struct var{ double media; int N; var(double...
17,384
// File name: add.cu #include <stdio.h> __global__ void add(int a, int b, int *c){ *c = a+b; } int main(void){ int c; int *device_c; cudaMalloc((void**)&device_c, sizeof(int)); add<<<1, 1>>>(2, 7, device_c); cudaMemcpy(&c, device_c, sizeof(int), cudaMemcpyDeviceToHost); printf("2+7 = %d\n", c); //cudeFree(devic...
17,385
#include "includes.h" #define BLOCKSIZE 4 #define CELLS_PER_THREAD 4 // Stride length __global__ void ShortestPath2(float *Arr1,float *Arr2,int N){ //Arr1 input array,Holds weights //Arr2 output array unsigned int k; int row=blockIdx.x; int col=threadIdx.x; if(row >= N || col >= N) return; int index=row*N+col; ...
17,386
#include <thrust/device_vector.h> #include <thrust/iterator/zip_iterator.h> #include <thrust/tuple.h> #include <thrust/reduce.h> int main() { // initialize vectors thrust::device_vector<int> A(3); thrust::device_vector<char> B(3); A[0] = 10; A[1] = 20; A[2] = 30; B[0] = 'x'; B[1] = 'y'; B[2] = 'z'; // crea...
17,387
#include "cuda_runtime.h" #include "device_launch_parameters.h" #define _USE_MATH_DEFINES #include <iostream> #include <math.h> #include <curand.h> using namespace std; __device__ __host__ __inline__ float N(float x) { return 0.5 + 0.5*erf(x*M_SQRT1_2); } __device__ __host__ void price(float k, float s, flo...
17,388
#pragma once #include <iostream> #include "tuple_utility.cu" namespace nearptd { template<size_t Dim> class Cell { public: typedef short int Cell_Index_T; typedef typename ntuple<Cell_Index_T, Dim>::tuple Cell_Tuple; ntuple<Cell_Index_T, Dim> Cell_Ntuple; Cell_Index_T c[Dim]; __host_...
17,389
#include <cuda.h> #include <stdio.h> #include <stdlib.h> #include <time.h> //max value for element of array #define MAX 100000 //defined threads per block for cims machines #define THREADS_PER_BLOCK 1024 //number of warp #define WARP 32 void generate(int *a, const int size); __global__ void get_max(int *array, int *m...
17,390
#include <stdio.h> #include <stdlib.h> #include <string.h> #include <time.h> #include "cuda_runtime.h" //#include "cuda_profiler_api.h" #define THREADS 32 //In each block THREADS*THREADS threads struct matrix { int ncols; int nrows; double* mat; }; void readMatrix(struct matrix* m, FILE* file); void printMatrix(s...
17,391
#include <stdio.h> __global__ void add(int a, int b, int *c) { *c = a + b; } int main() { //----------- cuda devices info --------------- int cuda_count; cudaDeviceProp prop; cudaGetDeviceCount(&cuda_count); printf("Exist %d device with cuda support\n",cuda_count); for(int device=0;...
17,392
#include "includes.h" __global__ void Predictor (const double TIME, double4 *p_pred, float4 *v_pred, float4 *a_pred, double4 *p_corr, double4 *v_corr, double *loc_time, double4 *acc, double4 *acc1, double4 *acc2, double4 *acc3, int istart, int* nvec, int ppgpus, unsigned int N){ int i = blockIdx.x*blockDim.x + th...
17,393
// // Created by moura on 30/12/2022. // #include <iostream> using namespace std; #define checkCudaErrors(val) check_cuda( (val), #val, __FILE__, __LINE__ ) void check_cuda(cudaError_t result, char const *const func, const char *const file, int const line) { if (result != cudaSuccess) { std::cerr << "CUD...
17,394
#include <stdio.h> #include <cuda.h> #include <math.h> #define BLOCK_DIM 16 __global__ void multiply(int *a, int *b, int *c, int wa) { int t_id = threadIdx.x; int i; c[t_id] = 0; for (i=0; i<wa; i++) c[t_id] += a[t_id * wa + i] * b[i]; } int main (int argc, char *argv[]) { // Initialize host variables ...
17,395
#include "Dummy.cuh" #include <cmath> __host__ __device__ Dummy::Dummy() : counter(0) { ; } __device__ void Dummy::incrementCounterDevice() { ++counter; magicNumber = counter * 5 * pow((double)3, __double2int_rd(magicNumber) % 100) + counter * magicNumber + counter * 2 * pow((double)2, __double2int_rd(magicNumber) % ...
17,396
__global__ void op1(float *mat, int mat_size) { int i=blockIdx.y*blockDim.y+threadIdx.y; int j=blockIdx.x*blockDim.x+threadIdx.x; float temp; if((i<mat_size) && (j<mat_size) && (j<mat_size-1)){ if(j%2==0){ temp = mat[i*mat_size+j]; mat[i*mat_size+j] = mat[i*mat_size+j+1]; mat[i*mat_size+j+1] = t...
17,397
/* This is a automatically generated test. Do not modify */ #include <stdio.h> #include <stdlib.h> #include <math.h> __global__ void compute(float comp, int var_1,float* var_2,float* var_3,int var_4,float var_5,float var_6,float var_7,float var_8,float var_9,float var_10,float var_11,float var_12,float var_13,float ...
17,398
#include <stdio.h> #include <stdlib.h> #include <time.h> #include <math.h> #include <sys/time.h> #include <stdint.h> #define MAX 100 #define MIN 1 __global__ void DUKernel(int *D_Level,int *D_Del,int n, int num); void IscomponentSame(int *L,int *D, int n,int num); uint64_t getTime(){ struct timeval t; gettimeofday(&...
17,399
// // main.cpp // Parallel Degree of Separation // // Created by Cary on 11/16/14. // Copyright (c) 2014 Cary. All rights reserved. // #include <iostream> #include <fstream> #include <cstdlib> #include <set> #include <stdio.h> #include <string.h> #include <stdlib.h> #include <assert.h> #include <sys/time.h> #inclu...
17,400
/* Write GPU kernels to compete the functionality of estimating the integral via the trapezoidal rule. */ #define BLOCK 64 #define GRID 16384 __global__ void trap_kernel(float a, float b, int n, float h, float *Result_fromGPU) { int tx = threadIdx.x; int bd = blockDim.x; int bi = blockIdx.x; int i; ...