Full-Speed Simulator Meeting

			      13 May 2015


Project Labor Budget Remaining: 53%

********Recent Progress and News:


********Next Tasks:

**Rob: Schedule with Hans:
       - Lab-test of the event-callback version for real-time flight on an
       	 AUV. 
       - Piggyback sea trial?
       - Merge with the CVS head.

**Rob: Will finalize a diagram of both approaches for TRN implementation.

**Rob/Rich: Will choose implementation.

**All: Massage the big three GNC tasks + simulator into a single task that
       runs real-time for flight, full speed for simulation.

* Rob: Get CVS running on the Mac.  Either install a Linux VM, or xCode,
              (xFree86?)
* Rob: Continue investigating why the VM doesn't fully utilize the host CPU.

*******Questions:

******Overall Objectives:

     1. Qnx Vm (Rob)  (DONE)
     	-Desktop
	-Laptop
     2. Synchronize Layered Control(Rich)  (DONE)
     	-Done using event call backs.
     3. Non blocking operation (Rich, Rob, Kent, Hans)
     	-First pass using event callbacks.
	-Second pass moves GNC processes into a single process.
     4. TRN Integration (Rich, Rob)
     5. Testing and Validation (Rob)


*******Past Meeting Notes:

15April:

Rich will have preliminary code next week.

Kent: tests
 * Sim-to-sim comparisons.
 * Simple Lab1.cfg test.
 * Other lab test?  Ask Hans.
 * How to revert code at sea?
   - Parallel head and branch trees on the vehicle?
 * Select potential piggy-back sea days (Hans).

22 April:

Talk to Hans.

Kent may become available the second half of this year.

Afternoon meeting w/ Kent and Hans:

Hans thinks that our design is probably OK, but we need to verify that the
event/callback code doesn't fragment the heap or have a memory leak.  This,
if happening, is probably not apparent in short runs.  So we need to first
run a 24 hr. fast sim while monitoring Qnx sin.  If that looks good, then we
try a 24 hr. deck test.
 
***Meeting Notes; 29 April:

Rob to talk to Hans about deck test and merging.
Rob make a diagram of the flight-code and Trn timing.
Rob If time permits will modify the SysDataLogger to log memory and cpu load
    by process.
Rich to consider other options on how to integrate TRN into the fast sim.

***6 May:

Worked with Rich on how to implement fast TRN. Discussed code and rough-draft
diagram.

Talked with Todd Ruston about how to get the Qnx VM to run faster.  His
suggestion was to download a free trial of the "Workstation" onto one of the
loaner laptops, load this Qnx VM, and see if there are more options or ways
to speed it up.  IF it works, then I can upgrade my workstation similarly.

***13 May:

Rob & Kent: Do a local preliminary merge.

Rich: Check in TRN.

Rob: Test the simulated TRN.
  
H:  The procedure is:
   0) Tag the head.
   1) Down-merge the head into the branch
   2) Verify simulator still works. Run sims.
   3) Run a deck test, look at loop timings.  Make sure something isn't
      getting blocked.

H: Should be faster than 60x.  VM problem or our architecture?
   Just take the dynamics alone, and loop them.  See how long the dynamics
   alone takes.