VEGAS Team Meeting

April 22, 2013

Attendees

CALTECH, UC-B, & Columbia

NRAO:

Agenda:

• GBT Test Observation report
• Ongoing system issues

Meeting Notes:

• Everything seemed to go well once again. The test psr showed up just fine, and the flux cal looked good again. The 1.2 Hz is still there at the same (non-problematic) level as before. Otherwise no notable issues as far as I could tell.
  • Further GC observations will wait for installation of the K-Band receiver
• Packet collisions between GUPPI and VEGAS - maybe something that is in the switch. Wolfgang should take a first look and work as needed with Dave McMahon at UCB.
• Dual instance tests. John was looking at a problem where GPU slot was running at 1/2 speed on boot. Supermicro is looking at it. This needs to be resolved or at least understood by next week or else we should plan to buy more machines and eliminate the dual instance requirement. The space will double but the AD power & HVAC loads will not change greatly
• The split between the FPGA and GPU modes is coming along nicely
• Mode integration meeting at GB. Needs about 1 month warning. Meeting topic for next week.

Long-term action Items

1. Ray to make VEGAS simulator.
2. A BOF for mode 13, or some other mode that uses GPUs.
   • (04-Feb) Not available yet
   • (11-Feb) 16 input version compiling now for subsequent test
   • (04-Mar) Should be ready to test in GB ~7-Mar
     • Need to refine what tests are done in Berkeley v. GB
       • Noise + tone tests in Berkeley
       • Replicate in GB (one Roach)
       • Build-out to eight Roach boards
       • Test on GBT
3. A BOF with working switching signal logic.
   • (04-Feb) designed but still troubleshooting switching table loading issues
   • (11-Feb) Added to design for Mode #1
   • (11-Feb) Add sfp shift into manager
4. A test and decision on what computers to buy.
   • (04-Feb) Not yet determined - Jayanth should go ahead with his tests to complete selection
   • (04-Feb) Needs HPC11.
5. All Banks not starting at the same time.
   • (04-Feb) (above)
   • (04-Mar) Resolved
6. It will be good if we get the full spectral dynamic range corresponding to the 8 bit ADC in the next version of mode 1.
   • (04-Feb) Conduct initial tests at Berkeley, to see if there are any limiting issues for dynamic range and optimize
   • (04-Feb) what is/are the requirement(s) w.r.t full scale sin wave?
   • (04-Feb) is this a new requirement for the manager (yes)? is a parameter required to select between options? (parametrized shifting)
   • (04-Feb) what are requirements for shift parameter? depends on result of studying spectra...
   • (25-Feb)
7. KFPA observation with mode 13.
   • (04-Feb) awaits mode 13
8. Manager/VEGAS HPC code that will let two instances of the program be run simultaneously.
   • (04-Feb) not available, but not required for GC obs.
9. Remember to do the RFI test of the rack.
   • (04-Feb) Not yet scheduled
10. Roach to Valon serial port connection.
    • (04-Feb) Not yet implemented; not req'd for GC obs
11. The I2C bus software tested with the manager and Galen's modules.
    • (04-Feb) Post GC Obs
12. Add software version information into FITS file
    • (11-Feb) After GC obs

Other items: