Message: Re: Simulation time increase much faster than direct proportional with events number Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

Feedback Re: Simulation time increase much faster than direct proportional with events number 

Forum: Multithreading
Re: Question Simulation time increase much faster than direct proportional with events number (Alex Kav)
Re: Question Re: Simulation time increase much faster than direct proportional with events number (Alex Kav)
Date: 13 Dec, 2016
From: Andrea Dotti <Andrea Dotti>

Hello, I am sorry for the late answer.

We have never observed this behavior before. We perform testing on several architectures and never seen this. The fact that you do not have problems on the laptop but you do have on the cluster may indicate an issue with the cluster itself. Another possibility is a memory leak somewhere in the application. If this memory leak is important and depending on how the cluster is configured you could have slow downs with the number of events. You can use a monitoring utility (maybe even ps or top) to monitor the memory usage as a function of time, if you observe very large increase this may indicate that there is a memory leak in the application. We do perform memory leak tests for each release and we do not have any important one that can explain what you see. However please note that a small increase in memory in G4 is expected due to the use of "allocators" that optimize CPU. This increase usually reaches a "plateau" after few tens of events (depending on the complexity of the events).

Hope this helps, Andrea

Inline Depth:
 1 1
 All All
Outline Depth:
 1 1
 2 2
 All All
Add message: (add)

1 None: Re: Simulation time increase much faster than direct proportional with events number   (Alex Kav - 14 Dec, 2016)
(_ Feedback: Re: Simulation time increase much faster than direct proportional with events number   (Andrea Dotti - 14 Dec, 2016)
 Add Message Add Message
to: "Re: Simulation time increase much faster than direct proportional with events number"

 Subscribe Subscribe

This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews