<p>如果你想要更快的pi算法,试试这个。我以前从未使用过Decimal模块;我通常使用mpmath来进行任意精度的计算,它附带了很多函数,并为pi和e内置了“常量”。但我想Decimal是一个方便的工具,因为它是一个标准模块。在</p>
<pre><code>''' The Salamin / Brent / Gauss Arithmetic-Geometric Mean pi formula.
Let A[0] = 1, B[0] = 1/Sqrt(2)
Then iterate from 1 to 'n'.
A[n] = (A[n-1] + B[n-1])/2
B[n] = Sqrt(A[n-1]*B[n-1])
C[n] = (A[n-1]-B[n-1])/2
PI[n] = 4A[n+1]^2 / (1-(Sum (for j=1 to n; 2^(j+1))*C[j]^2))
See http://stackoverflow.com/q/26477866/4014959
Written by PM 2Ring 2008.10.19
Converted to use Decimal 2014.10.21
Converted to Python 3 2018.07.17
'''
import sys
from decimal import Decimal as D, getcontext, ROUND_DOWN
def AGM_pi(m):
a, b = D(1), D(2).sqrt() / 2
s, k = D(0), D(4)
for i in range(m):
c = (a - b) / 2
a, b = (a + b) / 2, (a * b).sqrt()
s += k * c * c
#In case we want to see intermediate results
#if False:
#pi = 4 * a * a / (1 - s)
#print("%2d:\n%s\n" % (i, pi))
k *= 2
return 4 * a * a / (1 - s)
def main():
prec = int(sys.argv[1]) if len(sys.argv) > 1 else 50
#Add 1 for the digit before the decimal point,
#plus a few more to compensate for rounding errors.
#delta == 7 handles the Feynman point, which has six 9s followed by an 8
delta = 3
prec += 1 + delta
ctx = getcontext()
ctx.prec = prec
#The precision of the AGM value doubles on every loop
pi = AGM_pi(prec.bit_length())
#Round down so all printed digits are (usually) correct
ctx.rounding = ROUND_DOWN
ctx.prec -= delta
print("pi ~=\n%s" % +pi)
if __name__ == '__main__':
main()
</code></pre>