Get the sample frequency on a loop call to `time.time()`

README.md

Time Sampling

I was curious to know what level of precision I could get when using time.sleep(x), like, if it made any sense to try and sleep for say 0.0132 seconds.

It turns out, it does. On a Windows machine with a 1.6GHz processor, I get a sample rate such:

$ python3 time_sampling.py 6
Average step over 1000000 (100.0%) iterations: 2.77e-07
Sample frequency: 3.61 MHz

That is, nearly four million samples.

First argument of the script is a decimal order of magnitude (power of 10, denoting the iterations to perform).

time_sampling.py

import time
import sys

all_times = []
t0 = time.time()
t1 = t0
iterations_max = 10**int(sys.argv[1])

try:
    for iterations in range(1,iterations_max+1):
        t1 = time.time()
        all_times.append(t1-t0)
        t0 = t1
except KeyboardInterrupt:
    # In case we passed an insanely high number of iterations
    # We want to still get values out for our iterations
    pass

avg = sum(all_times)/iterations

scaled_freq = 1/avg
scale = ["", "K", "M", "G", "T"]
scale_idx = 0
kstep = 1000

for magnitude in scale:
    if scaled_freq > kstep:
        scale_idx += 1
        scaled_freq = scaled_freq/kstep


print("Average step over {} ({}%) iterations: {:.2e}".format(iterations, (100.0*iterations/iterations_max), avg))
print("Sample frequency: {:.2f} {}Hz".format(scaled_freq, scale[scale_idx]))