CMSSW I/O out-of-the-box

This page discusses CMSSW I/O under the following conditions:
  • CMSSW_3_5_4, unpatched
  • Running PAT tuple creation.
  • cacheSize set to 0.
  • 300 events run

Is a graphs showing the I/O activity:

  • I/O activity on analyzed file over the whole run of the application:
    trace_nocache.png
    There are four main bands of activity - first, when ROOT opens the file. Second, when ROOT starts opening and processing the TTree itself (figuring out what it needs to deserialize, event metadata). Third is the read of the first event. After this, there's a long pause in I/O when the conditions data is loaded. Finally, there is the run through the file itself.
  • Notice there is a lot of blocking and a very low level of parallelism.

In general, the I/O behavior is bad because:

  1. Completely random I/O (this is true, but not visible from blot)
  2. Too many small reads (this is true, but not visible from blot)
  3. Stalling on basically every read - we do no read hints.
Topic attachments
I Attachment History Action Size Date Who Comment
PNGpng trace_nocache.png r1 manage 71.9 K 2010-03-30 - 20:13 BrianBockelman  
Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r2 - 2010-05-13 - BrianBockelman
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback