Full-Speed Simulator Meeting

			      15 Jan 2015

				Agenda


Discuss:

   * Diagrams of proposed synchronization designs: Rob, Rich, Hans
   * Talk to Steve R and Yanwu.  Who else?  Talk to Dave and Brett? Rob
   * Hans will prototype type some code for sychronization.
   * Kent may take a look at the architecture (Rob to help).
   * Audit the code; find all system time calls.  Consider redesigning the time
     object. (TimeP.h)  Rob, Rich (Hans?)


Main Tasks:

     1. Qnx Vm (Rob)
     2. Synchronize (Rich)
     3. Non blocking operation (Rich and Rob)
     4. TRN Integration
     5. Testing and Validation (Rob)

Meeting Notes:

     Hans points out that if Simulator exceeds 200 ms it will overrun
     Navigation.  We need to synchronize entire loop to simulator finishing.
     This moves the dotted line in slide 3 (see 15Jan2015.pptx) to the end of
     simulator.  We can do this with by either passing a message back or
     using a semaphore.

     Rob to fix diagram.  TRN not really blocking; needs to for full speed
     sim.

     We will meet again in two weeks.  Rob to send an e-mail.

Tasks for the next meeting:

   * Branch the code, or start using GIT (Rich, Rob)

   * Discuss system time calls, and how many.  (Rich).

   * Change Navigation so that, when in simulation mode, rather than pending
     on the next 200 ms tick (Periodic Task), it pends on
     DynamicControl/Simulator finishing. Hopefully it will start ripping
     through.  We won't yet have replaced clock_gettime with simulator time,
     but this will prove the concept. We should start to get an inkling of
     how fast it will go. (Rich, Rob).

   * Improve the hand-drawn timing/data-flow diagram (Rob).

   * Kent to look at the code. (that would be Kent).