[Interpreter comparisons] A ChatGPT-generated script to compare the startup times of several interpreters in order to estimate the likely delays when scripts are used in Espanso triggers.. #python #ruby #pwsh #node #bash #espanso #perl #php #sh

benchmark.py

# https://chatgpt.com/share/685eb2a6-8dd4-8012-9fc9-d8f0ff137773
#!/usr/bin/env python3
import subprocess
import time
import shutil

interpreters = {
    "PowerShell (pwsh)": ["pwsh", "-NoProfile", "-Command", "exit"],
    "Ruby": ["ruby", "-e", "exit"],
    "Node.js": ["node", "-e", "process.exit()"],
    "Bash": ["bash", "-c", "exit"],
    "sh": ["sh", "-c", "exit"],
    "Python3": ["python3", "-c", "exit()"],
    "Perl": ["perl", "-e", "exit"],
    "PHP": ["php", "-r", "exit;"]
}

REPEATS = 10

print(f"Measuring average interpreter startup time over {REPEATS} runs...\n")

for name, cmd in interpreters.items():
    if not shutil.which(cmd[0]):
        print(f"{name:<20}: ❌ Not installed")
        continue

    durations = []
    for _ in range(REPEATS):
        try:
            start = time.perf_counter()
            subprocess.run(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, check=True)
            end = time.perf_counter()
            durations.append(end - start)
        except subprocess.CalledProcessError as e:
            print(f"{name:<20}: ❌ Error - {e}")
            durations = None
            break

    if durations:
        avg = sum(durations) / REPEATS
        print(f"{name:<20}: {avg:.4f} seconds (avg)")

My_data.txt

Measuring average interpreter startup time over 10 runs...

PowerShell (pwsh)   : 0.3438 seconds (avg)
Ruby                : 0.0778 seconds (avg)
Node.js             : 0.0910 seconds (avg)
Bash                : 0.0015 seconds (avg)
sh                  : 0.0009 seconds (avg)
Python3             : 0.0255 seconds (avg)
Perl                : 0.0021 seconds (avg)
PHP                 : ❌ Not installed

12 year-old PC, Linux Mint 21.2 Xfce, Intel(R) Core(TM) i3-3240 (4) @ 3.40 GHz, 24GB RAM, SSD

The first run, with just one instance of each (make line 18 `REPEATS = 1`), took roughly 
double the times, presumably because the interpreters weren't cached in memory. This is 
probably more representative for Espanso triggers.