From c79ad236b164b7498d87b45a58d1ffa257f882f1 Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:48:34 +0800
Subject: [PATCH 01/10] ADD file via upload

---
 Chen | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)
 create mode 100644 Chen

diff --git a/Chen b/Chen
new file mode 100644
index 0000000..aba197a
--- /dev/null
+++ b/Chen
@@ -0,0 +1,27 @@
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void vector_add_optimized(float* A, float* B, float* C, int size);
+int main() {
+	float A[SIZE], B[SIZE], C[SIZE];
+	int i;
+	srand(time(0));
+	for(i=0; i<SIZE; i++) {
+		A[i] = rand() % 100;
+		B[i] = rand() % 100;
+	}
+
+	clock_t start = clock();
+	vector_add_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("初始向量加法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+}
+
+void vector_add_optimized(float* A, float* B, float* C, int size) {
+	int i;
+	for(i=0; i<size; i++) {
+		C[i] = A[i] + B[i];
+	}
+}
\ No newline at end of file

From 3778299bf8bffb9b76e20f7c959733730a08124e Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:48:47 +0800
Subject: [PATCH 02/10] Delete 'Chen'

---
 Chen | 27 ---------------------------
 1 file changed, 27 deletions(-)
 delete mode 100644 Chen

diff --git a/Chen b/Chen
deleted file mode 100644
index aba197a..0000000
--- a/Chen
+++ /dev/null
@@ -1,27 +0,0 @@
-#include <stdio.h>
-#include <time.h>
-#include <stdlib.h>
-#define SIZE 1024
-
-void vector_add_optimized(float* A, float* B, float* C, int size);
-int main() {
-	float A[SIZE], B[SIZE], C[SIZE];
-	int i;
-	srand(time(0));
-	for(i=0; i<SIZE; i++) {
-		A[i] = rand() % 100;
-		B[i] = rand() % 100;
-	}
-
-	clock_t start = clock();
-	vector_add_optimized(A, B, C, SIZE);
-	clock_t end = clock();
-	printf("初始向量加法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
-}
-
-void vector_add_optimized(float* A, float* B, float* C, int size) {
-	int i;
-	for(i=0; i<size; i++) {
-		C[i] = A[i] + B[i];
-	}
-}
\ No newline at end of file

From fc101d3b2d4f3eb7d47e883aff554f8680e19685 Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:48:58 +0800
Subject: [PATCH 03/10] ADD file via upload

---
 Chen/CHEN_1.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)
 create mode 100644 Chen/CHEN_1.c

diff --git a/Chen/CHEN_1.c b/Chen/CHEN_1.c
new file mode 100644
index 0000000..aba197a
--- /dev/null
+++ b/Chen/CHEN_1.c
@@ -0,0 +1,27 @@
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void vector_add_optimized(float* A, float* B, float* C, int size);
+int main() {
+	float A[SIZE], B[SIZE], C[SIZE];
+	int i;
+	srand(time(0));
+	for(i=0; i<SIZE; i++) {
+		A[i] = rand() % 100;
+		B[i] = rand() % 100;
+	}
+
+	clock_t start = clock();
+	vector_add_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("初始向量加法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+}
+
+void vector_add_optimized(float* A, float* B, float* C, int size) {
+	int i;
+	for(i=0; i<size; i++) {
+		C[i] = A[i] + B[i];
+	}
+}
\ No newline at end of file

From 2856eb29eb825ca87e4f4633199171a091443223 Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:49:10 +0800
Subject: [PATCH 04/10] ADD file via upload

---
 Chen/CHEN_2.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)
 create mode 100644 Chen/CHEN_2.c

diff --git a/Chen/CHEN_2.c b/Chen/CHEN_2.c
new file mode 100644
index 0000000..312196c
--- /dev/null
+++ b/Chen/CHEN_2.c
@@ -0,0 +1,31 @@
+#include <arm_neon.h>
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void vector_add_optimized(float* A, float* B, float* C, int size);
+int main() {
+	float A[SIZE], B[SIZE], C[SIZE];
+	int i;
+	srand(time(0));
+	for(i=0; i<SIZE; i++) {
+		A[i] = rand() % 100;
+		B[i] = rand() % 100;
+	}
+
+	clock_t start = clock();
+	vector_add_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("使用NEON 优化向量加法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+}
+
+void vector_add_optimized(float* A, float* B, float* C, int size) {
+	int i;
+	for(i=0; i<size; i+=4) {
+		float32x4_t vecA = vld1q_f32(&A[i]);
+		float32x4_t vecB = vld1q_f32(&B[i]);
+		float32x4_t vecC = vaddq_f32(vecA, vecB);
+		vst1q_f32(&C[i], vecC);
+	}
+}
\ No newline at end of file

From 6c050e832f6757c478b47aaaa766e9ac746b0369 Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:49:36 +0800
Subject: [PATCH 05/10] ADD file via upload

---
 chen/CHEN_3.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)
 create mode 100644 chen/CHEN_3.c

diff --git a/chen/CHEN_3.c b/chen/CHEN_3.c
new file mode 100644
index 0000000..210cb7b
--- /dev/null
+++ b/chen/CHEN_3.c
@@ -0,0 +1,53 @@
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void matmul_optimized(float** A, float** B, float** C, int n);
+int main() {
+	int n = SIZE;
+
+	// 分配内存空间
+	float** A = (float**)malloc(sizeof(float*) * n);
+	float** B = (float**)malloc(sizeof(float*) * n);
+	float** C = (float**)malloc(sizeof(float*) * n);
+	int i, j;
+	for(i=0; i<n; i++) {
+		A[i] = (float*)malloc(sizeof(float) * n);
+		B[i] = (float*)malloc(sizeof(float) * n);
+		C[i] = (float*)malloc(sizeof(float) * n);
+	}
+	
+	// 随机化
+	srand(time(0));
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		A[i][j] = rand() % 100;
+		B[i][j] = rand() % 100;
+	}
+
+	// 计算并计时
+	clock_t start = clock();
+	matmul_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("初始向量乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+
+	// 释放内存空间
+	for(i=0; i<n; i++) {
+		free(A[i]);
+		free(B[i]);
+		free(C[i]);
+	}
+	free(A); free(B); free(C);
+}
+
+void matmul_optimized(float** A, float** B, float** C, int n) {
+	int i, j, k;
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		C[i][j] = 0;
+		for(k=0; k<n; k++) {
+			C[i][j] = A[i][k] * B[k][j];
+		}
+	}
+}
\ No newline at end of file

From b7750de60ae07476932e87cebd3c41414836276a Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:49:52 +0800
Subject: [PATCH 06/10] Delete 'chen/CHEN_3.c'

---
 chen/CHEN_3.c | 53 ---------------------------------------------------
 1 file changed, 53 deletions(-)
 delete mode 100644 chen/CHEN_3.c

diff --git a/chen/CHEN_3.c b/chen/CHEN_3.c
deleted file mode 100644
index 210cb7b..0000000
--- a/chen/CHEN_3.c
+++ /dev/null
@@ -1,53 +0,0 @@
-#include <stdio.h>
-#include <time.h>
-#include <stdlib.h>
-#define SIZE 1024
-
-void matmul_optimized(float** A, float** B, float** C, int n);
-int main() {
-	int n = SIZE;
-
-	// 分配内存空间
-	float** A = (float**)malloc(sizeof(float*) * n);
-	float** B = (float**)malloc(sizeof(float*) * n);
-	float** C = (float**)malloc(sizeof(float*) * n);
-	int i, j;
-	for(i=0; i<n; i++) {
-		A[i] = (float*)malloc(sizeof(float) * n);
-		B[i] = (float*)malloc(sizeof(float) * n);
-		C[i] = (float*)malloc(sizeof(float) * n);
-	}
-	
-	// 随机化
-	srand(time(0));
-	for(i=0; i<n; i++)
-	for(j=0; j<n; j++) {
-		A[i][j] = rand() % 100;
-		B[i][j] = rand() % 100;
-	}
-
-	// 计算并计时
-	clock_t start = clock();
-	matmul_optimized(A, B, C, SIZE);
-	clock_t end = clock();
-	printf("初始向量乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
-
-	// 释放内存空间
-	for(i=0; i<n; i++) {
-		free(A[i]);
-		free(B[i]);
-		free(C[i]);
-	}
-	free(A); free(B); free(C);
-}
-
-void matmul_optimized(float** A, float** B, float** C, int n) {
-	int i, j, k;
-	for(i=0; i<n; i++)
-	for(j=0; j<n; j++) {
-		C[i][j] = 0;
-		for(k=0; k<n; k++) {
-			C[i][j] = A[i][k] * B[k][j];
-		}
-	}
-}
\ No newline at end of file

From d3edcf589ac9892265e180165fc57cac3b9ea401 Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:50:02 +0800
Subject: [PATCH 07/10] ADD file via upload

---
 Chen/CHEN_3.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 53 insertions(+)
 create mode 100644 Chen/CHEN_3.c

diff --git a/Chen/CHEN_3.c b/Chen/CHEN_3.c
new file mode 100644
index 0000000..210cb7b
--- /dev/null
+++ b/Chen/CHEN_3.c
@@ -0,0 +1,53 @@
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void matmul_optimized(float** A, float** B, float** C, int n);
+int main() {
+	int n = SIZE;
+
+	// 分配内存空间
+	float** A = (float**)malloc(sizeof(float*) * n);
+	float** B = (float**)malloc(sizeof(float*) * n);
+	float** C = (float**)malloc(sizeof(float*) * n);
+	int i, j;
+	for(i=0; i<n; i++) {
+		A[i] = (float*)malloc(sizeof(float) * n);
+		B[i] = (float*)malloc(sizeof(float) * n);
+		C[i] = (float*)malloc(sizeof(float) * n);
+	}
+	
+	// 随机化
+	srand(time(0));
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		A[i][j] = rand() % 100;
+		B[i][j] = rand() % 100;
+	}
+
+	// 计算并计时
+	clock_t start = clock();
+	matmul_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("初始向量乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+
+	// 释放内存空间
+	for(i=0; i<n; i++) {
+		free(A[i]);
+		free(B[i]);
+		free(C[i]);
+	}
+	free(A); free(B); free(C);
+}
+
+void matmul_optimized(float** A, float** B, float** C, int n) {
+	int i, j, k;
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		C[i][j] = 0;
+		for(k=0; k<n; k++) {
+			C[i][j] = A[i][k] * B[k][j];
+		}
+	}
+}
\ No newline at end of file

From 5d750743a6ef4088b638f03de52148e7dd1cff0b Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:50:10 +0800
Subject: [PATCH 08/10] ADD file via upload

---
 Chen/CHEN_4.c | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)
 create mode 100644 Chen/CHEN_4.c

diff --git a/Chen/CHEN_4.c b/Chen/CHEN_4.c
new file mode 100644
index 0000000..882b53b
--- /dev/null
+++ b/Chen/CHEN_4.c
@@ -0,0 +1,59 @@
+#include <arm_neon.h>
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void matmul_optimized(float** A, float** B, float** C, int n);
+int main() {
+	int n = SIZE;
+
+	// 分配内存空间
+	float** A = (float**)malloc(sizeof(float*) * n);
+	float** B = (float**)malloc(sizeof(float*) * n);
+	float** C = (float**)malloc(sizeof(float*) * n);
+	int i, j;
+	for(i=0; i<n; i++) {
+		A[i] = (float*)malloc(sizeof(float) * n);
+		B[i] = (float*)malloc(sizeof(float) * n);
+		C[i] = (float*)malloc(sizeof(float) * n);
+	}
+	
+	// 随机化
+	srand(time(0));
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		A[i][j] = rand() % 100;
+		B[i][j] = rand() % 100;
+	}
+
+	// 计算并计时
+	clock_t start = clock();
+	matmul_optimized(A, B, C, SIZE);
+	clock_t end = clock();
+	printf("使用 NEON 优化稠密矩阵乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+
+	// 释放内存空间
+	for(i=0; i<n; i++) {
+		free(A[i]);
+		free(B[i]);
+		free(C[i]);
+	}
+	free(A); free(B); free(C);
+}
+
+void matmul_optimized(float** A, float** B, float** C, int n) {
+	int i, j, k;
+	float32x4_t vecA, vecB, vecC;
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		vecC = vdupq_n_f32(0.0);
+		for(k=0; k<n; k+=4) {
+			vecA = vld1q_f32(&A[i][k]);
+			vecB = vld1q_f32(&A[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);
+	}
+}
\ No newline at end of file

From 93d0eae3ae2cb299fa3b95961c2ddad7de2c5c3e Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:50:18 +0800
Subject: [PATCH 09/10] ADD file via upload

---
 Chen/CHEN_5.c | 70 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 70 insertions(+)
 create mode 100644 Chen/CHEN_5.c

diff --git a/Chen/CHEN_5.c b/Chen/CHEN_5.c
new file mode 100644
index 0000000..c5f99a5
--- /dev/null
+++ b/Chen/CHEN_5.c
@@ -0,0 +1,70 @@
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#define SIZE 1024
+
+void sparce_matmul_coo(float*, int*, int*, int,
+	float*, int*, int*, int, float*, int*, int*, int*);
+int main() {
+	// 矩阵 A 的 COO 格式
+	float A_values[] = {1, 2, 3, 4, 5}; 
+	int A_rowIndex[] = {0, 0, 1, 2, 2}; 
+	int A_colIndex[] = {0, 2, 1, 0, 2}; 
+	int A_nonZeroCount = 5;
+	// 矩阵 B 的 Coo 格式
+	float B_values[] = {6, 8, 7, 9}; 
+	int B_rowIndex[] = {0, 2, 1, 2}; 
+	int B_colIndex[] = {0, 0, 1, 2}; 
+	int B_nonZeroCount = 4;// 结果矩阵 C 的 Coo 格式
+	float C_values[SIZE]; 
+	int C_rowIndex[SIZE]; 
+	int C_colIndex[SIZE]; 
+	int C_nonZeroCount = 0;
+
+	clock_t start = clock();
+	sparce_matmul_coo(A_values, A_rowIndex, A_colIndex, A_nonZeroCount,
+	                  B_values, B_rowIndex, B_colIndex, B_nonZeroCount,
+					  C_values, C_rowIndex, C_colIndex, &C_nonZeroCount);
+	clock_t end = clock();
+	printf("基础的稀疏矩阵乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+}
+
+void sparce_matmul_coo(float* A_values, int* A_rowIndex, int* A_colIndex, int A_nonZeroCount,
+                       float* B_values, int* B_rowIndex, int* B_colIndex, int B_nonZeroCount,
+					   float* C_values, int* C_rowIndex, int* C_colIndex, int* C_nonZeroCount) {
+    int currentIndex = 0;
+	int i, j, k;
+	int rowA, colA, rowB, colB;
+	float valueA, valueB, product;
+	// 遍历 A 的非零元素
+	for(i=0; i<A_nonZeroCount; i++) {
+		rowA = A_rowIndex[i];
+		colA = A_colIndex[i];
+		valueA = A_values[i];
+		// 遍历 B 的非零元素
+		for(j=0; j<B_nonZeroCount; j++) {
+			rowB = B_rowIndex[j];
+			colB = B_colIndex[j];
+			valueB = B_values[j];
+			// 如果 A 的列和 B 的行匹配，则计算乘积并存储结果
+			if (colA == rowB) {
+				product = valueA * valueB;
+				// 检查是否已有此(rowA, colB) 项
+				int found = 0;
+				for(k=0; k<currentIndex; k++) {
+					if(C_rowIndex[k] == rowA && C_colIndex[k] == colB) {
+						C_values[k] += product;
+						break;
+					}
+				}
+				if (!found) {
+					C_values[currentIndex] = product;
+					C_rowIndex[currentIndex] = rowA;
+					C_colIndex[currentIndex] = colB;
+					currentIndex++;
+				}
+			}
+		}
+	}
+	*C_nonZeroCount = currentIndex;
+}
\ No newline at end of file

From 80c222779f864b5a0b4c7fa81693a069df8aeb6b Mon Sep 17 00:00:00 2001
From: p8sljnpht <3178612685@qq.com>
Date: Sat, 23 Nov 2024 10:50:25 +0800
Subject: [PATCH 10/10] ADD file via upload

---
 Chen/CHEN_6.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 88 insertions(+)
 create mode 100644 Chen/CHEN_6.c

diff --git a/Chen/CHEN_6.c b/Chen/CHEN_6.c
new file mode 100644
index 0000000..ad94b51
--- /dev/null
+++ b/Chen/CHEN_6.c
@@ -0,0 +1,88 @@
+#include <arm_neon.h>
+#include <stdio.h>
+#include <time.h>
+#include <stdlib.h>
+#include <string.h>
+#define SIZE 1024
+#define DENSE_MATRIX_SIZE 5
+
+float** sparce_matmul_coo(float* A_values, int* A_rowIndex, int* A_colIndex, int A_nonZeroCount,
+                          float* B_values, int* B_rowIndex, int* B_colIndex, int B_nonZeroCount);
+void matmul_optimized(float** A, float** B, float** C, int n);
+void print_matrix(float** m, int rows, int cols);
+int main() {
+	// 矩阵 A 的 COO 格式
+	float A_values[] = {1, 2, 3, 4, 5}; 
+	int A_rowIndex[] = {0, 0, 1, 2, 2}; 
+	int A_colIndex[] = {0, 2, 1, 0, 2}; 
+	int A_nonZeroCount = 5;
+	// 矩阵 B 的 Coo 格式
+	float B_values[] = {6, 8, 7, 9}; 
+	int B_rowIndex[] = {0, 2, 1, 2}; 
+	int B_colIndex[] = {0, 0, 1, 2}; 
+	int B_nonZeroCount = 4;// 结果矩阵 C 的 Coo 格式
+	float C_values[SIZE]; 
+	int C_rowIndex[SIZE]; 
+	int C_colIndex[SIZE]; 
+	int C_nonZeroCount = 0;
+
+	clock_t start = clock();
+	float** ans = sparce_matmul_coo(A_values, A_rowIndex, A_colIndex, A_nonZeroCount,
+	                  B_values, B_rowIndex, B_colIndex, B_nonZeroCount);
+	clock_t end = clock();
+	printf("优化的稀疏矩阵乘法时间：%lf\n", (double)(end-start) / CLOCKS_PER_SEC);
+	print_matrix(ans, DENSE_MATRIX_SIZE, DENSE_MATRIX_SIZE);
+}
+
+float** sparce_matmul_coo(float* A_values, int* A_rowIndex, int* A_colIndex, int A_nonZeroCount,
+                          float* B_values, int* B_rowIndex, int* B_colIndex, int B_nonZeroCount) {
+	// 分配内存空间
+	float** A = (float**)malloc(sizeof(float*) * DENSE_MATRIX_SIZE);
+	float** B = (float**)malloc(sizeof(float*) * DENSE_MATRIX_SIZE);
+	float** C = (float**)malloc(sizeof(float*) * DENSE_MATRIX_SIZE);
+	int i;
+	for(i=0; i<DENSE_MATRIX_SIZE; i++) {
+		A[i] = (float*)malloc(sizeof(float) * DENSE_MATRIX_SIZE);
+		B[i] = (float*)malloc(sizeof(float) * DENSE_MATRIX_SIZE);
+		C[i] = (float*)malloc(sizeof(float) * DENSE_MATRIX_SIZE);
+		memset(A[i], 0, sizeof(float) * DENSE_MATRIX_SIZE);
+		memset(B[i], 0, sizeof(float) * DENSE_MATRIX_SIZE);
+		memset(C[i], 0, sizeof(float) * DENSE_MATRIX_SIZE);
+	}
+	for (i=0; i<A_nonZeroCount; i++) A[A_rowIndex[i]][A_colIndex[i]] = A_values[i];
+	for (i=0; i<B_nonZeroCount; i++) B[B_rowIndex[i]][B_colIndex[i]] = B_values[i];
+	matmul_optimized(A, B, C, DENSE_MATRIX_SIZE);
+
+	// 释放内存空间
+	for(i=0; i<DENSE_MATRIX_SIZE; i++) {
+		free(A[i]);
+		free(B[i]);
+	}
+	free(A); free(B);
+	return C;
+}
+
+void matmul_optimized(float** A, float** B, float** C, int n) {
+	int i, j, k;
+	float32x4_t vecA, vecB, vecC;
+	for(i=0; i<n; i++)
+	for(j=0; j<n; j++) {
+		vecC = vdupq_n_f32(0.0);
+		for(k=0; k<n; k+=4) {
+			vecA = vld1q_f32(&A[i][k]);
+			vecB = vld1q_f32(&A[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);
+	}
+}
+void print_matrix(float** m, int rows, int cols) {
+	int i, j;
+	for(i=0; i<rows; i++) {
+		for(j=0; j<cols; j++) {
+			printf("%5.0f ", m[i][j]);
+		}
+		printf("\n");
+	}
+}
\ No newline at end of file