指数溢出

Question

I have this piece of code:

from __future__ import division
import matplotlib.pyplot as plt
import os
import numpy as np
import copy
import random
import math

A_bb=9.2572264*10**(-45)
E_bb=4.8043478/10**11

x=np.arange(0,20,0.01)

def func(x,T):
    x=np.array(x)
    return A_bb*(10**x)**3/(np.exp(10**x*E_bb/T)-1)
popt,pcov=curve_fit(func,frec, Flux_57442_27165)
print (popt)

Basically I want to fit a blackbody curve to some data points. But I get this error:

overflow encountered in exp w=A_bb*(10**x)**3/(np.exp(10**x*E_bb/T1)-1)

The value of TI expect is of order 10^4. What is the problem and how can I solve it?

Answer 1

Python's native int type, from the documentation is only capable of storing integers upto 2^32:

Plain integers (also just called integers) are implemented using long in C, which gives them at least 32 bits of precision

To get around this, Python also provides the long datatype:

Long integers have unlimited precision.

Simply apply long() to T if an overflow is being encountered:

T = long(T)

and hope that it works.

---

If even that fails, I strongly recommend using numpy instead, which has much more powerful datatypes .