PEP 703的内容是什么，意义又是什么呢？

可以说python的官方接受的no-GIL提议的PEP就是PEP 703给出的，如果GIL帧的从python中移除那么可以说对整个python生态圈将有着跨越性发展的意义。

====================================================

PEP 703地址：

https://peps.python.org/pep-0703/

PEP 703 – Making the Global Interpreter Lock Optional in CPython

在网上找到的一个中文的nogil项目的相关内容：

Python 官方研讨会：彻底移除 GIL 真的可行么？

===========================================================

nogil项目代码地址：

https://github.com/colesbury/nogil

https://hub.docker.com/r/nogil/python-cuda

为了测试nogil项目的效果，于是手动编译nogil-python3.9，并和新编译的python3.13进行性能对比：

nogil项目的编译命令：

./configure [--prefix=PREFIX] [--enable-optimizations]
make -j
make install

测试代码：

import threading
import time
 
def print_time( threadName, delay):
    s = 0

    for i in range(1000000000):
        s+=i

    print(threadName, s)

p1=threading.Thread( target=print_time, args=("Thread-1", 2, ) )
p2=threading.Thread( target=print_time, args=("Thread-2", 4, ) )

a = time.time()
p1.start()
p2.start()

p1.join()
p2.join()
print(time.time()-a)

测试结果：

可以看到nogil项目的多核心多线程确实可以较大程度提高多线程性能。

-----------------------------------------------------

再使用nogil项目的官方测试代码：

import sys
from concurrent.futures import ThreadPoolExecutor

print(f"nogil={getattr(sys.flags, 'nogil', False)}")

def fib(n):
    if n < 2: return 1
    return fib(n-1) + fib(n-2)

threads = 8
if len(sys.argv) > 1:
    threads = int(sys.argv[1])

with ThreadPoolExecutor(max_workers=threads) as executor:
    for _ in range(threads):
        executor.submit(lambda: print(fib(34)))

python3.13的表现：

(base) root@75f68d8ff029:~/nogil# time /root/python3.13/bin/python3.13 /tmp/x2.py
nogil=False
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m5.623s
user   0m5.638s
sys   0m0.004s
(base) root@75f68d8ff029:~/nogil# time /root/python3.13/bin/python3.13 /tmp/x2.py
nogil=False
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m5.608s
user   0m5.618s
sys   0m0.012s
(base) root@75f68d8ff029:~/nogil# time /root/python3.13/bin/python3.13 /tmp/x2.py
nogil=False
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m5.602s
user   0m5.591s
sys   0m0.028s

------------------------------------------

nogil python3.9的表现(8线程)：

(base) root@75f68d8ff029:~/nogil# time /root/python3.9_nogil/bin/python3.9 /tmp/x2.py
nogil=True
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m0.506s
user   0m3.816s
sys   0m0.012s
(base) root@75f68d8ff029:~/nogil# time /root/python3.9_nogil/bin/python3.9 /tmp/x2.py
nogil=True
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m0.514s
user   0m3.882s
sys   0m0.004s
(base) root@75f68d8ff029:~/nogil# time /root/python3.9_nogil/bin/python3.9 /tmp/x2.py
nogil=True
9227465
9227465
9227465
9227465
9227465
9227465
9227465
9227465

real   0m0.502s
user   0m3.815s
sys   0m0.004s