1. 16 Aug, 2013 1 commit
    • Mirko Stoffers's avatar
      Made cDynamicExpression::evaluate (and therefore cPar::doubleValue) thread-safe · 7ff3c017
      Mirko Stoffers authored
      Problem: cDynamicExpression::evaluate was not thread-safe (also noted in the
      comment), but used by cPar::doubleValue from a threaded context. This
      caused using par() for random variables to yield wrong results in rare
      cases.
      
      Solution: Made the global variables used by cDynamicExpression::evaluate
      threadlocal.
      
      Conflicts:
      
      	src/sim/cdynamicexpression.cc
      7ff3c017
  2. 14 May, 2013 1 commit
    • Mirko Stoffers's avatar
      Moved context switching in main loop · 49c66c64
      Mirko Stoffers authored
      Problem: The context was set in the main loop just prior to the event
      offloading. This yielded the problem that it was not set correctly
      during determination of event durations thus picking the duration from
      wrong RNGs.
      
      Solution: Moved context switching prior to duration determination.
      49c66c64
  3. 18 Apr, 2013 1 commit
    • Mirko Stoffers's avatar
      Fixed bug with not-removed events · e1750c09
      Mirko Stoffers authored
      Problem: With Horizon the semantics of cScheduler::getNextEvent() and
      cSimulator::doOneEvent() where changed. Without Horiozn, getNextEvent()
      peeked the event in the FES, and doOneEvent() removed it. This was
      changed for Horizon, but the cParsimScheduler::getNextEvent() function
      was missed.
      
      Solution: Reverted the semantics back to the stock OMNeT++ semantics.
      This introduces a problem with the cEEFScheduler since the EEF scheduler
      needs to pick a bunch of events from the FES and select one. This was
      fixed by putting the selected event back into the heap in
      cEEFScheduler::getNextEvent(). I prefered this solution since the
      changed semantics would also break the TkEnv.
      e1750c09
  4. 16 Apr, 2013 2 commits
    • Mirko Stoffers's avatar
      Fixed bug in seed generation · fc839da5
      Mirko Stoffers authored
      fc839da5
    • Mirko Stoffers's avatar
      Modified random number genaration · 08285dff
      Mirko Stoffers authored
      Instead of deriving from cNumberGenerator, models now use a distinct RNG
      in the global RNG array. This ignores the user specific mapping, but
      introduces a mapping that creates <num-rngs> RNGs per module and maps
      them accordingly.
      
      This solves the problem that RNGs could not be called from inside a
      module before.
      
      You may need to change your models and you'll have to update the example
      model as well!
      08285dff
  5. 03 Apr, 2013 1 commit
    • Mirko Stoffers's avatar
      Renamed duration -> eventDuration · cbd2c699
      Mirko Stoffers authored
      Problem: We had a variable cMessage::duration for Horizon and a
      variable cPacket::duration from OMNeT++ itself while cPacket inherits
      from cMessage. This might lead to the wrong variable being accessed
      by packets.
      
      Solution: Renamed the Horizon variable cMessage::duration in
      cMessage::eventDuration
      cbd2c699
  6. 07 Mar, 2013 2 commits
    • Mirko Stoffers's avatar
      Moved cAsyncModule functionality to cSimpleModule, dropped cAsyncModule · fbe00e73
      Mirko Stoffers authored
      We eliminated the need for special cAsyncModule objects
      since it is more convenient to just stick to cSimpleModule in the
      model implementation and active Horizon through omnetpp.ini switches.
      
      This means, that you need to change all inheritances of cAsyncModule
      to cSimpleModule for your Horizon-enabled models. For new models, you
      can use Horizon w/o changing cSimpleModule to cAsyncModule.
      fbe00e73
    • Mirko Stoffers's avatar
      Moved inheritance of cNumberGenerator from cAsyncModule to cSimpleModule · ffa9279f
      Mirko Stoffers authored
      We are about to eliminate the need for special cAsyncModule objects
      since it is more convenient to just stick to cSimpleModule in the
      model implementation and active Horizon through omnetpp.ini switches.
      
      Therefore, we first shifted the inheritance of cNumberGenerator from
      cAsyncModule to cSimpleModule. This means that everytime you call
      number generator functions inside a module, a local number generator
      of the calling module is used rather than the one global RNG.
      ffa9279f
  7. 08 Sep, 2012 1 commit
  8. 05 Jun, 2012 1 commit
  9. 30 May, 2012 1 commit
  10. 24 May, 2012 2 commits
  11. 21 Apr, 2012 1 commit
  12. 19 Apr, 2012 1 commit
  13. 29 Mar, 2012 1 commit
  14. 15 Mar, 2012 3 commits
    • Simon Tenbusch's avatar
      [master] · bb311670
      Simon Tenbusch authored
      - added two MMN Queueing sample traces
      bb311670
    • Simon Tenbusch's avatar
      [master] · b5650378
      Simon Tenbusch authored
      - adding new binaries: JARs for nativelib and sequencechart and 64bit libopplibs.so
      - still missing: 32bit libopplibs.so
      b5650378
    • Simon Tenbusch's avatar
      Merge branch 'sequencechart' · 2338867c
      Simon Tenbusch authored
      Conflicts:
      	src/eventlog/ievent.h
      	src/eventlog/sequencechartfacade.cc
      
      [sequencechart]
      merging sequencechart branch into master branch
      2338867c
  15. 13 Feb, 2012 1 commit
  16. 01 Feb, 2012 1 commit
    • Simon Tenbusch's avatar
      [sequencechart] · 7c7d95e5
      Simon Tenbusch authored
      - fixing hopefully last bugs with the new event order: events are drawn correctly and now, only those events are drawn that are actually visible
      7c7d95e5
  17. 27 Jan, 2012 1 commit
  18. 25 Jan, 2012 3 commits
    • Simon Tenbusch's avatar
      [sequencechart] · 9bd9afa7
      Simon Tenbusch authored
      -started to use new ordering in some methods
      9bd9afa7
    • Simon Tenbusch's avatar
      [sequencechart] · 12288d0e
      Simon Tenbusch authored
      - implemented second ordering for the events (getNextEventRealTime / getPreviousEventRealTime) for better performance
      12288d0e
    • Simon Tenbusch's avatar
      [sequencechart] · 50468190
      Simon Tenbusch authored
      - changed bottleneck finding buttons to look for bottlenecks on the critical path and according to overlapping coefficient
      50468190
  19. 18 Jan, 2012 3 commits
  20. 13 Jan, 2012 1 commit
    • Simon Tenbusch's avatar
      [sequencechart] · 97e29f40
      Simon Tenbusch authored
      - fixed behaviour of eventLenghtMode Toggle: when switching from REAL_TIME to SIMULATION_TIME, the event length mode is switched to DURATION
      97e29f40
  21. 11 Jan, 2012 1 commit
    • Simon Tenbusch's avatar
      [sequencechart] · a2c79680
      Simon Tenbusch authored
      - fixed many bugs that were introduced due to potentially different ordering of events in REAL_TIME mode, and due to calculation of the simulation end timestamp when durations are turned on.
      - ticks are now drawn correctly
      a2c79680
  22. 21 Dec, 2011 2 commits
  23. 16 Dec, 2011 1 commit
  24. 14 Dec, 2011 1 commit
    • Simon Tenbusch's avatar
      [sequencechart] · 0596ce67
      Simon Tenbusch authored
      - further work on REAL TIME timeline mode
      - fixed bug in critical path computation (see wiki)
      0596ce67
  25. 09 Dec, 2011 1 commit
  26. 08 Dec, 2011 2 commits
  27. 07 Dec, 2011 3 commits