Python variable memory management

Question

I just wrote this primitive script:

from sys import getsizeof as g
x = 0
s = ''
while s != 'q':
    x = (x << 8) + 0xff
    print(str(x) + "    [" + str(g(x)) + "]")
    s = input("Enter to proceed, 'q' to quit  ")

The output is as follows - and quite surprising, as I perceive it:

255                    [28]
65535                  [28]
16777215               [28]
4294967295             [32]
1099511627775          [32]
281474976710655        [32]
72057594037927935      [32]
18446744073709551615   [36]

And so on. My point is: it seems that the variable x has some sort of 'overhead' with a size of 25 bytes. Where does this come from? Thanks in advance for any attempt to help me.

Everything is an object in Python.

juanpa.arrivillaga
– juanpa.arrivillaga

2017-07-11 18:39:39 +00:00
Commented Jul 11, 2017 at 18:39 — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Jul 11, 2017 at 18:39

Jon Kiparsky · Accepted Answer · 2017-07-11 18:03:06Z

1

A python int is an object, so it's not surprising that it has a small overhead. If this overhead starts to become meaningful for you then this implies you're manipulating substantial collections of ints, which suggests to me that the numpy library is probably something you should consider.

answered Jul 11, 2017 at 18:03

Jon Kiparsky

7,8242 gold badges26 silver badges42 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

juanpa.arrivillaga Over a year ago

Or even array.array, or struct

Jon Kiparsky Over a year ago

@juanpa.arrivillaga Yes, those are good for the raw representation, but numpy brings a lot of tools for manipulation.

juanpa.arrivillaga Over a year ago

Yep, I agree. But sometimes, those alternatives are all you need, and they are part of the standard library.

Collectives™ on Stack Overflow

Python variable memory management

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related