I've tried multiple code to calculate execution time for a numerical sample of an algorithm in python
the first one(%%timeit) gives the execution time in microsecond however the second one which is shown below give the execution time of the same sample of algorithm in second ( the numerical sample is not generated randomely)
the second code is:
import time
start = time.time()
end = time.time()
total_time = end - start
print("\n"+ str(total_time))
I'm confused there is a significant different between the two results and I hope someone could help me with the correct answer
As far as I know, %%timeit (as a magic command in a notebook) runs the code multiple times and computes the average.
Your direct method using time.time() simply compares timestamps, so the result may fluctuate a lot more depending on available system ressources at that moment.
Also, depending on the actually timed code, since %%timeit runs that many times, python may be clever and recycle some intermediate results to make it run faster after the first time.
Instead of the manual method, you can also use %time (for one line of code) or %%time (for a cell) to just get the computation time without bothering with timestamps...
The command below may be applied:
import time
start_time = time.time()
main()
print("--- %s seconds ---" % (time.time() - start_time))
before your coding, using datetime.now, after coding ,using datetime.now
import datetime
start_time = datetime.datetime.now()
...
end_time = datetime.datetime.now()
calculate_time = (end_time-start_time).total_seconds()
print("calculate_time:", calculate_time)
This should work:
from datetime import datetime
start_time = datetime.now()
******** **
****** your script
********
end_time = datetime.now()
print("Duration: {}".format(end_time - start_time))
Maybe this would help:
https://docs.python.org/3/library/timeit.html
https://pynative.com/python-get-execution-time-of-program/
One can use time.perf_counter() (https://docs.python.org/3/library/time.html)
It would be something like the following:
import time
start = time.perf_counter()
[OPERATION]
end = time.perf_counter()
However, there are various ways to do that, depending on one's requirements.
See this thread: https://stackoverflow.com/q/1557571/7109869
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python