Skip to content
This repository was archived by the owner on Sep 11, 2024. It is now read-only.

Commit 9dea882

Browse files
lyubolplyubolp
committed
Added some examples to 09
1 parent 4e33f01 commit 9dea882

File tree

1 file changed

+195
-2
lines changed

1 file changed

+195
-2
lines changed

‎09 - Multithreading/09 - Multithreading.ipynb

+195-2
Original file line numberDiff line numberDiff line change
@@ -1387,7 +1387,200 @@
13871387
"cell_type": "markdown",
13881388
"metadata": {},
13891389
"source": [
1390-
"WIP"
1390+
"### Пример 1"
1391+
]
1392+
},
1393+
{
1394+
"cell_type": "markdown",
1395+
"metadata": {},
1396+
"source": [
1397+
"Нека имаме функция, която умножава две матрици. \n",
1398+
"\n",
1399+
"Нека напишем функция, която умножава две матрици паралелно. Нека използваме `multiprocessing` библиотеката за целта."
1400+
]
1401+
},
1402+
{
1403+
"cell_type": "code",
1404+
"execution_count": 4,
1405+
"metadata": {},
1406+
"outputs": [],
1407+
"source": [
1408+
"import time\n",
1409+
"def timeit():\n",
1410+
" def wrapper(func):\n",
1411+
" def inner(*args, **kwargs):\n",
1412+
" start = time.time()\n",
1413+
" result = func(*args, **kwargs)\n",
1414+
" end = time.time()\n",
1415+
" print(f'The function took {(end-start):.2f} seconds')\n",
1416+
" return result\n",
1417+
" return inner\n",
1418+
" return wrapper\n",
1419+
"\n",
1420+
"\n",
1421+
"@timeit()\n",
1422+
"def multiply_matrix(a: list[list[int]], b: list[list[int]]) -> list[list[int]]:\n",
1423+
" size_of_a = (len(a), len(a[0]))\n",
1424+
" size_of_b = (len(b), len(b[0]))\n",
1425+
"\n",
1426+
" if size_of_a[1] != size_of_b[0]:\n",
1427+
" raise ValueError(\"The matrixes cannot be multiplied\")\n",
1428+
" \n",
1429+
" m, n, p = size_of_a[0], size_of_a[1], size_of_b[1]\n",
1430+
"\n",
1431+
" result = [[0 for _ in range(p)] for _ in range(m)]\n",
1432+
"\n",
1433+
" for i in range(m):\n",
1434+
" for j in range(p):\n",
1435+
" result[i][j] = sum(a[i][k] * b[k][j] for k in range(n))\n",
1436+
"\n",
1437+
" return result"
1438+
]
1439+
},
1440+
{
1441+
"cell_type": "markdown",
1442+
"metadata": {},
1443+
"source": [
1444+
"### Решение на пример 1"
1445+
]
1446+
},
1447+
{
1448+
"cell_type": "code",
1449+
"execution_count": 17,
1450+
"metadata": {},
1451+
"outputs": [],
1452+
"source": [
1453+
"from multiprocessing import Array, Process\n",
1454+
"\n",
1455+
"@timeit()\n",
1456+
"def multithreaded_multiply_matrix(a: list[list[int]], b: list[list[int]]) -> list[list[int]]:\n",
1457+
" size_of_a = (len(a), len(a[0]))\n",
1458+
" size_of_b = (len(b), len(b[0]))\n",
1459+
"\n",
1460+
" if size_of_a[1] != size_of_b[0]:\n",
1461+
" raise ValueError(\"The matrixes cannot be multiplied\")\n",
1462+
" \n",
1463+
" m, n, p = size_of_a[0], size_of_a[1], size_of_b[1]\n",
1464+
" \n",
1465+
"\n",
1466+
" result = Array('i', m * p)\n",
1467+
"\n",
1468+
" threads = [Process(target=multiply_row, args=(result, n, p, i, a, b)) for i in range(m)]\n",
1469+
"\n",
1470+
" for thread in threads:\n",
1471+
" thread.start()\n",
1472+
"\n",
1473+
" for thread in threads:\n",
1474+
" thread.join()\n",
1475+
"\n",
1476+
" return [result[i*p:(i+1)*p] for i in range(m)]\n",
1477+
"\n",
1478+
"def multiply_row(shared_memory: Array, n: int, p: int, i: int, a: list[list[int]], b: list[list[int]]):\n",
1479+
" for j in range(p):\n",
1480+
" target_index = i * p + j\n",
1481+
" shared_memory[target_index] = sum(a[i][k] * b[k][j] for k in range(n))"
1482+
]
1483+
},
1484+
{
1485+
"cell_type": "code",
1486+
"execution_count": 18,
1487+
"metadata": {},
1488+
"outputs": [
1489+
{
1490+
"name": "stdout",
1491+
"output_type": "stream",
1492+
"text": [
1493+
"The function took 3.87 seconds\n",
1494+
"The function took 0.99 seconds\n"
1495+
]
1496+
}
1497+
],
1498+
"source": [
1499+
"from random import randint\n",
1500+
"\n",
1501+
"m, n, p = 400, 400, 400\n",
1502+
"value_range = (0, 10)\n",
1503+
"\n",
1504+
"a = [[randint(*value_range) for _ in range(n)] for _ in range(m)]\n",
1505+
"b = [[randint(*value_range) for _ in range(p)] for _ in range(n)]\n",
1506+
"\n",
1507+
"res_1 = multiply_matrix(a, b)\n",
1508+
"res_2 = multithreaded_multiply_matrix(a, b)\n"
1509+
]
1510+
},
1511+
{
1512+
"cell_type": "markdown",
1513+
"metadata": {},
1514+
"source": [
1515+
"### Пример 2"
1516+
]
1517+
},
1518+
{
1519+
"cell_type": "markdown",
1520+
"metadata": {},
1521+
"source": [
1522+
"Нека напишем функция `parallel_file_search`, която приема път до файл, низ по който търсим и брой на нишките, които ще използваме за търсенето. Функцията трябва да запише редовете, които съдържат низа в нов файл.\n",
1523+
"\n",
1524+
"Файла трябва да се раздели на `N` части, в които да се търси паралелно.\n",
1525+
"\n",
1526+
"Казваме, че един ред съдържа търсения низ, ако той се среща в него."
1527+
]
1528+
},
1529+
{
1530+
"cell_type": "markdown",
1531+
"metadata": {},
1532+
"source": [
1533+
"### Решение на пример 2"
1534+
]
1535+
},
1536+
{
1537+
"cell_type": "code",
1538+
"execution_count": 38,
1539+
"metadata": {},
1540+
"outputs": [],
1541+
"source": [
1542+
"from multiprocessing import Process, Semaphore\n",
1543+
"\n",
1544+
"def search_in_lines(lines: list[str], to_search: str, output_path: str, semaphore: Semaphore):\n",
1545+
" for line in lines:\n",
1546+
" if to_search in line:\n",
1547+
" results.append(line)\n",
1548+
"\n",
1549+
"@timeit()\n",
1550+
"def parallel_file_search(input_path: str, to_search: str, n: int, output_path: str) -> list[str]:\n",
1551+
" with open(input_path, encoding='utf-8') as input_file_descriptor:\n",
1552+
" amount_of_lines = len(input_file_descriptor.readlines())\n",
1553+
" \n",
1554+
" chunk_size = amount_of_lines // n\n",
1555+
" regions = [(i * chunk_size, (i+1) * chunk_size) for i in range(n - 1)] + [((n - 1) * chunk_size, amount_of_lines + 1)]\n",
1556+
"\n",
1557+
" procs = [Process(target=search_string_in_file, args=(to_search, input_path, region, output_path)) for region in regions]\n",
1558+
"\n",
1559+
" for proc in procs:\n",
1560+
" proc.start()\n",
1561+
"\n",
1562+
" for proc in procs:\n",
1563+
" proc.join()\n",
1564+
"\n",
1565+
"def search_string_in_file(string: str, file_path: str, region: tuple[int, int], output_file_path: str, semaphore: Semaphore):\n",
1566+
" results = []\n",
1567+
" start, end = region\n",
1568+
" with open(file_path, encoding='utf-8') as input_file_descriptor:\n",
1569+
" for line_number, line in enumerate(input_file_descriptor):\n",
1570+
" if start <= line_number < end:\n",
1571+
" column = line.find(string)\n",
1572+
" \n",
1573+
" while column != -1:\n",
1574+
" results.append(line.strip())\n",
1575+
" column = line.find(string, column+1)\n",
1576+
"\n",
1577+
" semaphore.acquire()\n",
1578+
"\n",
1579+
" with open(output_file_path, encoding='utf-8', mode='w+') as output_file_descriptor:\n",
1580+
" for result in results:\n",
1581+
" output_file_descriptor.write(result + '\\n')\n",
1582+
"\n",
1583+
" semaphore.release()"
13911584
]
13921585
}
13931586
],
@@ -1407,7 +1600,7 @@
14071600
"name": "python",
14081601
"nbconvert_exporter": "python",
14091602
"pygments_lexer": "ipython3",
1410-
"version": "3.10.8"
1603+
"version": "3.10.12"
14111604
},
14121605
"vscode": {
14131606
"interpreter": {

0 commit comments

Comments
 (0)