antonov_dmitry_lab_5/benchmark.py

@@ -2,13 +2,45 @@ import numpy as np
 import time
 import concurrent.futures
 
-def multiply_matrices(matrix_a, matrix_b):
-    return np.dot(matrix_a, matrix_b)
 
-def multiply_matrices_parallel(matrix_a, matrix_b):
-    with concurrent.futures.ThreadPoolExecutor() as executor:
-        result = executor.submit(np.dot, matrix_a, matrix_b)
-    return result.result()
+def multiply_matrices(matrix_a, matrix_b):
+    if len(matrix_a[0]) != len(matrix_b):
+        raise ValueError("Incompatible matrix dimensions for multiplication")
+
+    result = [[0 for _ in range(len(matrix_b[0]))] for _ in range(len(matrix_a))]
+
+    for i in range(len(matrix_a)):
+        for j in range(len(matrix_b[0])):
+            for k in range(len(matrix_b)):
+                result[i][j] += matrix_a[i][k] * matrix_b[k][j]
+
+    return result
+
+
+def multiply_matrices_parallel(matrix_a, matrix_b, num_threads):
+    if len(matrix_a[0]) != len(matrix_b):
+        raise ValueError("Incompatible matrix dimensions for multiplication")
+
+    result = [[0 for _ in range(len(matrix_b[0]))] for _ in range(len(matrix_a))]
+
+    with concurrent.futures.ThreadPoolExecutor(max_workers=num_threads) as executor:
+        futures = []
+        for i in range(len(matrix_a)):
+            futures.append(executor.submit(_multiply_row, matrix_a, matrix_b, i))
+
+        for i, future in enumerate(concurrent.futures.as_completed(futures)):
+            result[i] = future.result()
+
+    return result
+
+
+def _multiply_row(matrix_a, matrix_b, i):
+    row_result = [0 for _ in range(len(matrix_b[0]))]
+    for j in range(len(matrix_b[0])):
+        for k in range(len(matrix_b)):
+            row_result[j] += matrix_a[i][k] * matrix_b[k][j]
+    return row_result
+
 
 def benchmark_sequential(size):
     matrix_a = np.random.rand(size, size)
@@ -20,24 +52,30 @@ def benchmark_sequential(size):
 
     return end_time - start_time
 
-def benchmark_parallel(size):
+
+def benchmark_parallel(size, num_threads):
     matrix_a = np.random.rand(size, size)
     matrix_b = np.random.rand(size, size)
 
     start_time = time.time()
-    multiply_matrices_parallel(matrix_a, matrix_b)
+    multiply_matrices_parallel(matrix_a, matrix_b, num_threads)
     end_time = time.time()
 
     return end_time - start_time
 
+
 if __name__ == "__main__":
-    sizes = [100, 300, 500, 700, 900, 1000, 1200, 1400, 1700, 2000]
+    sizes = [300]
+    threads = [2, 8]
 
     for size in sizes:
         sequential_time = benchmark_sequential(size)
-        parallel_time = benchmark_parallel(size)
-
-        print(f"Размер матрицы: {size}x{size}")
         print(f"Время обычное: {sequential_time} с")
-        print(f"Время параллельное: {parallel_time} с")
-        print(f"Ускорение: {sequential_time / parallel_time}\n")
+        print(f"Размер матрицы: {size}x{size}")
+
+    for thread in threads:
+        for size in sizes:
+            parallel_time = benchmark_parallel(size, thread)
+            print(f"Размер матрицы: {size}x{size}")
+            print(f"Время параллельное: {parallel_time} с")
+            print(f"Потоков:  {thread}")

antonov_dmitry_lab_5/simple_app.py

@@ -6,14 +6,41 @@ app = Flask(__name__)
 
 
 def multiply_matrices(matrix_a, matrix_b):
-    result = np.dot(matrix_a, matrix_b)
+    if len(matrix_a[0]) != len(matrix_b):
+        raise ValueError("Incompatible matrix dimensions for multiplication")
+
+    result = [[0 for _ in range(len(matrix_b[0]))] for _ in range(len(matrix_a))]
+
+    for i in range(len(matrix_a)):
+        for j in range(len(matrix_b[0])):
+            for k in range(len(matrix_b)):
+                result[i][j] += matrix_a[i][k] * matrix_b[k][j]
+
     return result
 
+def multiply_matrices_parallel(matrix_a, matrix_b, num_threads):
+    if len(matrix_a[0]) != len(matrix_b):
+        raise ValueError("Incompatible matrix dimensions for multiplication")
+
+    result = [[0 for _ in range(len(matrix_b[0]))] for _ in range(len(matrix_a))]
+
+    with concurrent.futures.ThreadPoolExecutor(max_workers=num_threads) as executor:
+        futures = []
+        for i in range(len(matrix_a)):
+            futures.append(executor.submit(_multiply_row, matrix_a, matrix_b, i))
+
+        for i, future in enumerate(concurrent.futures.as_completed(futures)):
+            result[i] = future.result()
+
+    return result
+
+def _multiply_row(matrix_a, matrix_b, i):
+    row_result = [0 for _ in range(len(matrix_b[0]))]
+    for j in range(len(matrix_b[0])):
+        for k in range(len(matrix_b)):
+            row_result[j] += matrix_a[i][k] * matrix_b[k][j]
+    return row_result
 
-def multiply_matrices_parallel(matrix_a, matrix_b):
-    with concurrent.futures.ThreadPoolExecutor() as executor:
-        result = executor.submit(np.dot, matrix_a, matrix_b)
-    return result.result()
 
 
 @app.route('/')