Project

General

Profile

Bug #473

Instability in MAUSSteppingActionTest

Added by Rogers, Chris over 12 years ago. Updated over 12 years ago.

Status:
Closed
Priority:
High
Assignee:
Rogers, Chris
Category:
common_cpp
Target version:
MAUS-v0.0.3
Start date:
26 May 2011
Due date:
% Done:

100%

Estimated time:
Workflow:

Description

Humm here's an awkward one - I occasionally get output like:

tests/cpp_unit/Simulation/MAUSSteppingActionTest.cpp:122: Failure          
Value of: 9.                                                               
  Actual: 9                                                                
Expected: out["momentum"]["x"].asDouble()                                  
Which is: 3.9693107475943875e+114                                          
tests/cpp_unit/Simulation/MAUSSteppingActionTest.cpp:123: Failure          
Value of: 18.                                                              
  Actual: 18                                                               
Expected: out["momentum"]["y"].asDouble()                                  
Which is: 7.938621495188775e+114                                           
tests/cpp_unit/Simulation/MAUSSteppingActionTest.cpp:124: Failure          
Value of: 27.                                                              
  Actual: 27                                                               
Expected: out["momentum"]["z"].asDouble()                                  
Which is: 1.1907932242783163e+115                                          
tests/cpp_unit/Simulation/MAUSSteppingActionTest.cpp:126: Failure          
Value of: 9.                                                               
  Actual: 9                                                                
Expected: out["energy"].asDouble()                                         
Which is: 8.753015450537953e+227

from the unit tests. Seems to occur only occasionally - which probably means something isn't being allocated that should be. (Check with valgrind?)

Related issues

Related to MAUS - Bug #449: valgrind testRejectedRogers, Chris11 May 2011

Actions
#1

Updated by Tunnell, Christopher over 12 years ago

Ivan Reid:


    Can I just mention igprof, which I've been using for a week or so now,
tracking memory leaks and CPU hogs in CMS?  It has nowhere near the overhead
of valgrind (a few percent for performance, 3 or 4 x for memory trace) and
provides a much more detailed calling sequence.

http://igprof.sourceforge.net/analysis.html

    The web-based navigator is quite slick, too, e.g.
http://cms-service-sdtweb.web.cern.ch/cms-service-sdtweb/ptf/igprof/digi/423/navigator/PoissonQ_perf
#2

Updated by Rogers, Chris over 12 years ago

  • Priority changed from Normal to High
#3

Updated by Rogers, Chris over 12 years ago

I commented out the issue - but see comment on valgrind output issue #449

#4

Updated by Rogers, Chris over 12 years ago

  • Status changed from Open to Closed
  • % Done changed from 0 to 100

Closed by r569 (valgrind no longer gives error messages related to this - there are some funny error messages I don't understand but they look to be related to gunit framework)

#5

Updated by Rogers, Chris over 12 years ago

  • Target version changed from Future MAUS release to MAUS-v0.0.3

Also available in: Atom PDF