Memory Management and Limits

sys includes several functions for understanding and controlling memory usage.

Reference Counts

The primary implementation of Python (CPython) uses reference counting and garbage collection for automatic memory management. An object is automatically marked to be collected when its reference count drops to zero. To examine the reference count of an existing object, use getrefcount().

import sys

one = []
print('At start         :', sys.getrefcount(one))

two = one

print('Second reference :', sys.getrefcount(one))

del two

print('After del        :', sys.getrefcount(one))

The value reported is actually one higher than expected because there is a temporary reference to the object held by getrefcount() itself.

$ python3

At start         : 2
Second reference : 3
After del        : 2

See also

  • gc – Control the garbage collector via the functions exposed in gc.

Object Size

Knowing how many references an object has may help find cycles or a memory leak, but it is not enough to determine what objects are consuming the most memory. That requires knowledge about how big objects are.

import sys

class MyClass:

objects = [
    [], (), {}, 'c', 'string', b'bytes', 1, 2.3,
    MyClass, MyClass(),

for obj in objects:
    print('{:>10} : {}'.format(type(obj).__name__,

getsizeof() reports the size of an object in bytes.

$ python3

      list : 64
     tuple : 48
      dict : 288
       str : 50
       str : 55
     bytes : 38
       int : 28
     float : 24
      type : 1016
   MyClass : 56

The reported size for a custom class does not include the size of the attribute values.

import sys

class WithoutAttributes:

class WithAttributes:
    def __init__(self):
        self.a = 'a'
        self.b = 'b'

without_attrs = WithoutAttributes()
print('WithoutAttributes:', sys.getsizeof(without_attrs))

with_attrs = WithAttributes()
print('WithAttributes:', sys.getsizeof(with_attrs))

This can give a false impression of the amount of memory being consumed.

$ python3

WithoutAttributes: 56
WithAttributes: 56

For a more complete estimate of the space used by a class, provide a __sizeof__() method to compute the value by aggregating the sizes of attributes of an object.

import sys

class WithAttributes:
    def __init__(self):
        self.a = 'a'
        self.b = 'b'

    def __sizeof__(self):
        return object.__sizeof__(self) + \
            sum(sys.getsizeof(v) for v in self.__dict__.values())

my_inst = WithAttributes()

This version adds the base size of the object to the sizes of all of the attributes stored in the internal __dict__.

$ python3



Allowing infinite recursion in a Python application may introduce a stack overflow in the interpreter itself, leading to a crash. To eliminate this situation, the interpreter provides a way to control the maximum recursion depth using setrecursionlimit() and getrecursionlimit().

import sys

print('Initial limit:', sys.getrecursionlimit())


print('Modified limit:', sys.getrecursionlimit())

def generate_recursion_error(i):
    generate_recursion_error(i + 1)

except RuntimeError as err:
    print('Caught exception:', err)

Once the recursion limit is reached, the interpreter raises a RuntimeError exception so the program has an opportunity to handle the situation.

$ python3

Initial limit: 1000
Modified limit: 10
Caught exception: maximum recursion depth exceeded while calling
a Python object

Maximum Values

Along with the runtime configurable values, sys includes variables defining the maximum values for types that vary from system to system.

import sys

print('maxsize   :', sys.maxsize)
print('maxunicode:', sys.maxunicode)

maxsize is the maximum size of a list, dictionary, string, or other data structure dictated by the C interpreter’s size type. maxunicode is the largest integer Unicode point supported by the interpreter as currently configured.

$ python3

maxsize   : 9223372036854775807
maxunicode: 1114111

Floating Point Values

The structure float_info contains information about the floating-point type representation used by the interpreter, based on the underlying system’s float implementation.

import sys

print('Smallest difference (epsilon):', sys.float_info.epsilon)
print('Digits (dig)              :', sys.float_info.dig)
print('Mantissa digits (mant_dig):', sys.float_info.mant_dig)
print('Maximum (max):', sys.float_info.max)
print('Minimum (min):', sys.float_info.min)
print('Radix of exponents (radix):', sys.float_info.radix)
print('Maximum exponent for radix (max_exp):',
print('Minimum exponent for radix (min_exp):',
print('Max. exponent power of 10 (max_10_exp):',
print('Min. exponent power of 10 (min_10_exp):',
print('Rounding for addition (rounds):', sys.float_info.rounds)

These values depend on the compiler and underlying system. These examples were produced on OS X 10.9.5 on an Intel Core i7.

$ python3

Smallest difference (epsilon): 2.220446049250313e-16

Digits (dig)              : 15
Mantissa digits (mant_dig): 53

Maximum (max): 1.7976931348623157e+308
Minimum (min): 2.2250738585072014e-308

Radix of exponents (radix): 2

Maximum exponent for radix (max_exp): 1024
Minimum exponent for radix (min_exp): -1021

Max. exponent power of 10 (max_10_exp): 308
Min. exponent power of 10 (min_10_exp): -307

Rounding for addition (rounds): 1

See also

  • The float.h C header file for the local compiler contains more details about these settings.

Integer Values

The structure int_info holds information about the internal representation of integers used by the interpreter.

import sys

print('Number of bits used to hold each digit:',
print('Size in bytes of C type used to hold each digit:',

These examples were produced on OS X 10.9.5 on an Intel Core i7.

$ python3

Number of bits used to hold each digit: 30
Size in bytes of C type used to hold each digit: 4

The C type used to store integers internally is determined when the interpreter is built. 64-bit architectures automatically use 30-bit integers by default, and they can be enabled for 32-bit architectures with the configuration flag --enable-big-digits.

See also

Byte Ordering

byteorder is set to the native byte order.

import sys


The value is either big for big-endian or little for little-endian.

$ python3


See also

  • Wikipedia: Endianness – Description of big and little endian memory systems.

  • array and struct – Other modules that depend on the byte order of data.

  • float.h – The C header file for the local compiler contains

    more details about these settings.