Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Taking Runtime observation repeatedly gets slower over time

See original GitHub issue

🐛 Bug

After creating an environment, taking the runtime observation from the same environment progressively takes more time to compute.

Log from the MRE below:

Completed in 61.9ms
Completed in 60.3ms
Completed in 92.3ms
Completed in 63.4ms
Completed in 60.7ms
Completed in 64.7ms
...
Completed in 223.1ms
Completed in 252.4ms
Completed in 237.7ms
Completed in 219.0ms
Completed in 221.4ms
Completed in 235.5ms

To Reproduce

Steps to reproduce the behavior:

Create an environment.
Repeatedly take the runtime observation.

import gym
import compiler_gym
from compiler_gym.util.timer import Timer

with gym.make('llvm-v0') as env:
    env.reset(benchmark='benchmark://cbench-v1/bitcount')
    while True:
        with Timer(f'Completed'):
            env.observation.Runtime()

Expected behavior

Taking the runtime observation repeatedly shouldn’t get slower progressively.

Environment

CompilerGym: development branch
How you installed CompilerGym (conda, pip, source): source
OS: Ubuntu 18.04 LTS
Python version: 3.9.6
Build command you used (if compiling from source): make bdist_wheel-linux

Issue Analytics

State:
Created 2 years ago
Comments:5 (4 by maintainers)

Top GitHub Comments

1reaction

ChrisCumminscommented, Sep 11, 2021

Hi @aymanizz, thanks for the report, and for the minimal repro script! I was able to successfully reproduce the problem on my end. Let me take a look a look at this next week and get back to you.

Cheers, Chris

0reactions

ChrisCumminscommented, Sep 17, 2021

#398 has been merged into development. Please update and let me know if you are able to reproduce the problem.

Cheers, Chris