bitwise left shift different results for identical operands

Question

Consider the array

a = np.array([64,  8, 16, 24, 32, 40, 48, 56, 64])

Its first and last elements are identical

a[0] == a[-1]
# True

Now let's left shift by a

b = 1<<a

And compare the first and last elements

b[0] == b[-1]
# False

What the heck is going on here?

Can anyone reproduce this? Is this a bug?

The actual values are

b
# array([                0,               256,             65536,
#                 16777216,        4294967296,     1099511627776,
#          281474976710656, 72057594037927936,                 1])

numpy version is 1.17.0 on Python 3.6.5

Expected Values

Either value has its logic: 0 = 1 * 2^64 mod 2^64 and 1 = 1 * 2^(64 mod 64)

Related

https://github.com/numpy/numpy/issues/10299

Answer 1

For me its showing true: python 3.7.4 numpy 1.17.x

import numpy
 a = numpy.array([10,20,15,10])
 a[0]==a[-1]
True
 b = 1<<a
 b[0] == b[-1]
True
