caicaixiehan的博客

1.1是一个单纯的块在图像上的划分，划分结果如下图1.2hello.cu#include<stdio.h>__global__ void hellofromGPU(void){printf("hello world from CPU\n");}int main(){printf("hello world from gpu\n");hellofromGPU<<<1,10>>>();cudaDeviceReset();return 0

//cublas计算矩阵乘法（sgemm）#include<cuda_runtime.h>#include<device_launch_parameters.h>#include<cublas_v2.h>#include<curand.h>#include<stdlib.h>#include<stdio.h>#include<assert.h>#include<time.h>#include.

//一些基本的cublas示例#include<cuda_runtime.h>#include<device_launch_parameters.h>#include<cublas_v2.h>#include<stdlib.h>#include<assert.h>#include<math.h>//初始化向量void vector_init(float*a, int n) { for (int i = 0; i &.

//矩阵乘法的优化版本#include<cuda_runtime.h>#include<device_launch_parameters.h>#include<stdlib.h>#include<stdio.h>#include<assert.h>#include<math.h>#define SHMEM_SIZE 16*16*4 //为方便起见，静态shmem计算 16x16矩阵__global__ void .