Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[FEATURE] #58

Open
Moh9774 opened this issue Feb 21, 2024 · 3 comments
Open

[FEATURE] #58

Moh9774 opened this issue Feb 21, 2024 · 3 comments
Labels
enhancement New feature or request

Comments

@Moh9774
Copy link

Moh9774 commented Feb 21, 2024

Hello,

I have a question regarding the possibility of parallelizing ESA by altering the number of parallel collisional timesteps, denoted as "n_par_ct". To explore this possibility, I adjusted the "n_par_ct" values from 1 to 8. However, the outcomes were unexpected:

With n_par_ct= 1, the elapsed time was 1330.11 seconds, projecting to approximately 89.97 hours to simulate 20 years.
For n_par_ct= 2, it reduced to 1136.46 seconds, or around 76.87 hours for the 20-year simulation.
Increasing to n_par_ct= 3, the time slightly improved to 1121.81 seconds, equivalent to 75.88 hours.
At n_par_ct= 4, the duration increased again to 1290.32 seconds, or 87.28 hours.
With n_par_ct= 5, it further rose to 1434.44 seconds, translating to 97.02 hours.
At n_par_ct= 6, the time was 1235.26 seconds, or 83.55 hours.
For n_par_ct= 7, it was slightly less at 1228.97 seconds, approximately 83.13 hours.
Finally, at n_par_ct= 8, the time was 1254.69 seconds, or 84.87 hours.
I was anticipating that increasing the number of n_par_ct would lead to decreased simulation times. Has anyone else experienced similar results, or could someone provide guidance on this matter? Your insights would be greatly appreciated.

@Moh9774 Moh9774 added the enhancement New feature or request label Feb 21, 2024
@bluescarni
Copy link
Member

@Moh9774 in order to understand better, we would need to know the specifics of the setup of your simulation (e.g., how many objects, how big, etc.).

In any case, the first thing you should try to understand is whether cascade is already making an optimal use of the multiple cores on your system. E.g., you could try to run the top or htop tools while the simulation is running and observe if all cores are being used to their fullest. If that is already the case, it is unlikely that tweaking n_par_ct will lead to noticeable speedups.

@Moh9774
Copy link
Author

Moh9774 commented Mar 14, 2024

Hello,

Thank you for your advice. Following your suggestion, I monitored the system resources using top and I noticed that only one core is actively working during the simulation (99-100% cpu usage), which suggests that parallelization is not being used.

This observation was found by adjusting the n_par_ct values from 1 to 8. The specifics of my simulation setup, in response to your inquiry, are as follows:

Implemented Example: Simulation of 20 years in LEO, focusing on collision tracking (example was found on github webpage).

Given this, I'm interested to know if the parallelization works for your current development enviroment,

if so, would mean it is a specific problem of our environment (we installed cascade as shown in the github page) or we are missing some important parameters to distribute the workload across cores.

Thank you once again for your support.

@chu1995
Copy link

chu1995 commented Apr 29, 2024

Why can't I install this package using conda

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

3 participants