Tips, Recipies, and Notes#

Supressing output from the benchmark#

In your setup, use

import sys, os
sys.stdout = sys.stderr = open(os.devnull, 'a', encoding='utf-8')

Or in one line:

import sys, os; sys.stdout = sys.stderr = open(os.devnull, 'a', encoding='utf-8')

Using stdin as an input#

In fastero, "-" Is not used for stdin, you should use "file: stdin" instead

100-400ns overhead#

First I’ll show you this example

Python 3.11.0a6 (main, Mar  7 2022, 16:46:19) [MSC v.1929 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import timeit
>>> timeit.timeit(number=1) * 1_000_000
>>> timeit.timeit(number=1_000_000)

Although both of them are supposed to have the same value (in an ideal situation). The first one is almost 40 times lower.

This also applies to number=2 but the difference is almost cut in half (20 times slower).

>>> timeit.timeit(number=2) * 1_000_000
>>> timeit.timeit(number=2_000_000)

Due to this reason. Using very low run count will result in a 100-400 nanosecond overhead. This will not matter most of the time. Because, if you are dealing with code so fast that this is gonna matter, the run count will probably be very high so it won’t matter anymore. But I would still like to mention it here in case someone’s wondering

Programmatic Usage#

Although fastero isn’t meant to be used programmatically, you can use it as that, with the use of shell commands

import os
import json

data = json.loads(os.popen("fastero \"str(1)\" \"f'{1}'\" --json --quiet").read())

This will output the minimum time required to run “str(1)”. The data variable format is same as the format given by the –export-json flg