Skip to content

jsearcy1/QuickAnaHPC

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

QuickAnaHPC

This package contains some testing code for running a QuickAna-based analysis application on HPCs such as Edison (and eventually Cori). For the sake of comparisons, functionality is included to run the application in various modes using a few different EventLoop drivers. The application can run directly on an interactive node, can run on the PDSF batch system, and can run using PROOF-lite on either a local node or a batch system like Edison.

Overview

The application is based on EventLoop and QuickAna. It processes systematics and applies the (mostly default) object definitions of QuickAna. The output is currently a set of kinematic histograms for each systematic. Python is used to configure, steer, and submit the application to various EventLoop backends.

Contents of the package

  • AnalysisAlg - A simple EventLoop Algorithm for running QuickAna. It has a configuration flag for disabling systematics. It books and fills the output histograms.
  • runQuickAnaHPC.py - Executable python script which configures and launches the application. Has several command line options for specifying the input samples, the driver to use, and various other job configuration options.
    • The --task ID:N option allows to specify a chunk of the input data for processing. The process will only use the ID-th chunk of size 1/N of the input samples. If samples are broken up into individual files, this is a useful way to distribute files evenly across separately submitted tasks.
  • Python modules - Utility code for sample management, logging, and loading RootCore libraries. Can be expanded as necessary. I may move some machine-specific or driver-specific setups in here.
  • Edison example submission scripts in scripts.

About

Testing code for running QuickAna analysis on Edison

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 47.8%
  • Python 39.9%
  • Shell 11.1%
  • C 1.2%