You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
63 lines
1.5 KiB
63 lines
1.5 KiB
#include <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <ctime>
|
|
#include <arm_neon.h>
|
|
#define SIZE 1024
|
|
|
|
void matmul_optimized(float** A, float** B, float** C, int n) {
|
|
for(int i=0;i<n;i++){
|
|
for(int j=0;j<n;j++){
|
|
float32x4_t vecC=vdupq_n_f32(0.0);
|
|
for(int k=0;k<n;k+=4){
|
|
float32x4_t vecA=vld1q_f32(&A[i][k]);
|
|
float32x4_t vecB=vld1q_f32(&B[k][j]);
|
|
vecC=vmlaq_f32(vecC,vecA,vecB);
|
|
|
|
}
|
|
C[i][j]=vgetq_lane_f32(vecC,0)+
|
|
vgetq_lane_f32(vecC,1)+
|
|
vgetq_lane_f32(vecC,2)+
|
|
vgetq_lane_f32(vecC,3);
|
|
}
|
|
}
|
|
}
|
|
|
|
int main(){
|
|
//分配矩阵内存
|
|
float** A=new float*[SIZE];
|
|
float** B=new float*[SIZE];
|
|
float** C=new float*[SIZE];
|
|
|
|
for(int i=0;i<SIZE;i++){
|
|
A[i]=new float[SIZE];
|
|
B[i]=new float[SIZE];
|
|
C[i]=new float[SIZE];
|
|
}
|
|
|
|
//初始化
|
|
srand(time(NULL));
|
|
for(int i=0;i<SIZE;i++){
|
|
for(int j=0;j<SIZE;j++){
|
|
A[i][j]=rand()%10;
|
|
B[i][j]=rand()%10;
|
|
}
|
|
}
|
|
|
|
//计时并输出
|
|
clock_t start=clock();
|
|
matmul_optimized(A,B,C,SIZE);
|
|
clock_t end=clock();
|
|
printf("优化稠密矩阵乘法时间：%f秒\n",(double)(end-start)/CLOCKS_PER_SEC);
|
|
|
|
for(int i=0;i<SIZE;i++){
|
|
delete[] A[i];
|
|
delete[] B[i];
|
|
delete[] C[i];
|
|
}
|
|
delete[] A;
|
|
delete[] B;
|
|
delete[] C;
|
|
|
|
return 0;
|
|
}
|