Bug #1431

Memory leaks in MAUS

Added by Rajaram, Durga about 10 years ago. Updated about 10 years ago.

Code Management
Target version:
Start date:
12 March 2014
Due date:
% Done:


Estimated time:
New Issue


At CM38 Adam mentioned a slow memory leak in MAUS.

Indeed, there do seem to be leaks.
On my machine, running offline reconstruction on a smallish data-set [ 04995.000 ], the memory usage climbed from 180MB to 550MB
[ memory monitor log attached ]

Valgrind shows some stuff coming from Python, ROOT and GEANT.
On the MAUS side, there seem to be lots of leaks from G4 DetectorConstruction, Globals, JsonCpp converters

Valgrind memcheck log attached


memory_monitor.log (1.22 KB) memory_monitor.log memory monitor log Rajaram, Durga, 12 March 2014 06:45
valgrind.log (7.06 MB) valgrind.log valgrind log Rajaram, Durga, 12 March 2014 06:45

Related issues

Related to MAUS - Bug #1456: Run 4000 job on GRID failed with memory leakClosedRogers, Chris16 April 2014


Updated by Rogers, Chris about 10 years ago

See also
  • #1278 - there was a problem where tracker folks were creating a new instance of the datastructure on every tracker recon step (without deleting it)
  • #1293 (item 19) - I list the bits of Geant4 that I couldn't figure out how to clean up. I believe the Geant4 memory leaks are only when the geometry is instantiated i.e. at start of run. I guess this was done against G4.9.2, I haven't checked G4.9.6.

Updated by Rajaram, Durga about 10 years ago

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

Fixed in trunk r710


Updated by Rajaram, Durga about 10 years ago

  • Target version changed from Future MAUS release to MAUS-v0.8.4

Also available in: Atom PDF