profile result with unexpected big negative number #211
Description
I want to profile the memory usage of objects with the codes below:
from memory_profiler import profile
NUM = 100000
class Person(object):
def __init__(self, name):
self.name = name
@profile
def run1():
ps = []
for i in range(NUM):
ps.append(Person(str(i)))
@profile
def run2():
r = []
for _ in range(NUM):
r.append({})
if __name__ == "__main__":
run1()
run2()
And I get the result:
$ python -m memory_profiler demo.py
Filename: demo.py
Line # Mem usage Increment Line Contents
================================================
12 10.8 MiB 10.8 MiB @profile
13 def run1():
14 10.8 MiB 0.0 MiB ps = []
15 53.1 MiB -744.4 MiB for i in range(NUM):
16 53.1 MiB -711.9 MiB ps.append(Person(str(i)))
Filename: demo.py
Line # Mem usage Increment Line Contents
================================================
18 14.8 MiB 14.8 MiB @profile
19 def run2():
20 14.8 MiB 0.0 MiB r = []
21 42.4 MiB -47.0 MiB for _ in range(NUM):
22 42.4 MiB -19.5 MiB r.append({})
The run1 line:15 and line:16 have a big negative number.
But if I change the global NUM to 10000, the result is total different:
$ python -m memory_profiler demo.py
Filename: demo.py
Line # Mem usage Increment Line Contents
================================================
12 10.7 MiB 10.7 MiB @profile
13 def run1():
14 10.7 MiB 0.0 MiB ps = []
15 15.1 MiB 0.3 MiB for i in range(NUM):
16 15.1 MiB 4.0 MiB ps.append(Person(str(i)))
Filename: demo.py
Line # Mem usage Increment Line Contents
================================================
18 11.6 MiB 11.6 MiB @profile
19 def run2():
20 11.6 MiB 0.0 MiB r = []
21 14.2 MiB 0.0 MiB for _ in range(NUM):
22 14.2 MiB 2.6 MiB r.append({})
How to explain the different between the two cases?
Should I consider the loop or something else, thanks.