Message: AnalysisManager root performance for large numbers of ntuples Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

None AnalysisManager root performance for large numbers of ntuples 

Forum: Analysis
Date: 01 Jan, 2014
From: Sebastian Heil <Sebastian Heil>

Hi everyone,

I have a problem with the file writing of the AnalysisManager. The simulation contains about 1000 Detectors, for which I want to safe the deposited energy individually. Previously this was done via TFile and TTree in SensitiveDetector. Due to version 4.10 and the conflicts of this solution and the multithreading I decided to let the AnalysisManager handle the root files by storing the data as nTuples. This is implemented just like Example B4c.
The problem is that calling analysisManager->Write() takes very long. For 1500 nTuples with four threads it takes about 30 s real time to write the files for a single event. The required time increases with both, number of events and number of threads. This converges to somewhat 3 s per event with four threads, which bears no relation to the actual simulation time of about 3 ms per event.

Is this an expected/known behaviour for the AnalysisManager for large numbers of nTuples?
Or is there probably something wrong with my approach? (in which case I would provide more information about the actual implementation of course)

Thank you very much for your help and a happy new year,
Sebastian Heil

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

1 None: Re: AnalysisManager root performance for large numbers of ntuples   (Ivana Hrivnacova - 02 Jan, 2014)
1 None: Re: AnalysisManager root performance for large numbers of ntuples   (Sebastian Heil - 02 Jan, 2014)
2 None: Re: AnalysisManager root performance for large numbers of ntuples   (Ivana Hrivnacova - 02 Jan, 2014)
1 None: Re: AnalysisManager root performance for large numbers of ntuples   (Sebastian Heil - 02 Jan, 2014)
(_ None: Re: AnalysisManager root performance for large numbers of ntuples   (Sebastian Heil - 02 Jan, 2014)
(_ None: Re: AnalysisManager root performance for large numbers of ntuples   (Ivana Hrivnacova - 03 Jan, 2014)
3 None: Re: AnalysisManager root performance for large numbers of ntuples   (Sebastian Heil - 15 Jan, 2014)
4 None: Re: AnalysisManager root performance for large numbers of ntuples   (Guy Barrand - 21 Jan, 2014)
1 None: Re: AnalysisManager root performance for large numbers of ntuples   (Sebastian Heil - 23 Jan, 2014)
 Add Message Add Message
to: "AnalysisManager root performance for large numbers of ntuples"

 Subscribe Subscribe

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


[ Geant 4 Home | Geant 4 HyperNews | Search | Request New Forum | Feedback ]