Dear IRR Review Committee:

 The Pulsar group would like to thank the committee for taking time
to review the project and assist us in the transition process.  We
have carefully reviewed the recommendations.  I will briefly update
the Pulsar status here.  A more detailed report will be given in the
Trigger Hardware Meeting this Friday.

 Most of the items in category 1 (tests and debugging in preparation
for operations) have been completed or nearly so.  Below I classified
the items into three categories:  completed, in progress, or on
to-do-list.

Completed Items:

1.) Correctness and robustness of decisions with silicon at
   various luminosities.
     [ We've been checking decision bits between Alpha and Pulsar
     for close to a month now in symbiotic running.  Pulsar and
     Alpha decisions agree well.  The discrepancies are understood
     (only in rate limiting, isolation and showerMax triggers). ]

2.) Demonstrate the readout list calculation and transmission to the TS

   [ We have checked that the ROL from Alpha agree with
     Pulsar.  In a test run with Pulsar driving alone, we have
     checked that the Pulsar ROL agree with what TS thinks
     (from TS1D bank)]

3.) Add a bit to TL2D bank indicating whether it is from Pulsar or Alpha.
     [ We have added a hex string "2B" in the unused portion of the L1
     Trigger card to indicate that the block is from Pulsar. ]


4.) Add Alpha/Pulsar indications to the ScalerMon window

   [ Thanks to Bill Badgett, the ScalerMon GUI now shows either the
     Alpha Logo or the Crab Nebula picture and some text as well to
     indicate which scaler it's showing. ]

5.) Develop a L3 filter that flags events where TL2D from the Alpha
   does not match TL2D from the Pulsar.

   [ The first version of the filter is in the hands of the Level-3
     group.  This version only compares the trigger decision bits.      We'll add the remaining TL2D data check in the next round of update.]

6.) Operational issues:
      -  Identify all code that needs to be in CVS.

         [ All Pulsar monitoring code and trigger algorithm code
           are in CVS. ]

      -  Fix known run control operational efficiency problems

         [HRR transition time has been optimized.  We are limiting
          diagnostics to only the ones essential for operation. ]

      -  Trigger GUI

         [ TDWG now build Alpha and Pulsar tables as default.  Veronica
           has rebuilt all trigger tables with Pulsar and Alpha ]


In Progress or nearly completed:

1.)  Demonstrate that dynamic prescaling is working

    [ Dynamic prescale change was checked before.  However, we recently
      moved the L2TS board (for reading out TL2D) into the lower alpha.
      That change has affected the DPS chain and the code is being
      updated now.  We'll be checking it shortly]

2.)  Demonstrate that Pulsar is pipelining the assembly and the
    algorithm processing of events

    [ We will take some more data at different luminosity/L1A rate to
      study this ]

3.)  Head-to-head comparisons of the Pulsar alone vs Alpha alone
    L2 latency (from TS) and L2 deadtime at reasonably high lumi.

    [ We are ready to do this test. We will coordinate with the
      Ops managers for a special beam run (lumi >100E30) as
      soon as possible. ]

4.)  Provide detail monitoring plan

    [ We are working towards a version of PulsarMon that's
      CO friendly.  We expect the code to be ready to be integrated into
      TrigMon by the end of this week. In addition to online monitor, we
      will also run over all the data offline to validate the trigger
      performance. ]

5.)  Demonstrate sufficient hardware pool is available for maintenance.

    [ We are in the process of setting aside a pool of fully
      loaded Pulsar boards with mezzanine cards and AUX cards
      as hot spares. We have sufficient hotlink/taxi/slink mezzanine
      cards, AUX cards and Pulsar boards to put together a copy of
      the system.]


To-do-list:

1.)  Automated Pulsar firmware version tracking relative to trigger
    version and provisions to prevent mismatches.

    [ We could envision at CONFIG time to check Pulsar firmware
      version.  However, the implementation will not be ready by Mar 1st.]

2.)  L2 bit space: implementing and testing Pulsar code to handle the
    increased L2 trigger bit space.

    [ The code change required on the Pulsar side is trivial (few hours).
      We'll contact the TDWG to generate a trigger table with more
      than 128 trigger bits.  However, I just want to remind the
      committee that the main issue with going to more than 128
      trigger bits is not on the Pulsar side, it's all the infrastructure
      downstream of Pulsar that need to be updated (eg.  ScalerMon,
      front-end DAQ code, Prereq modules, Consumers monitors, etc...) ]

3.)  MOU is one of the important concerns from the review committee.  The RunIIb
    management and the spokes are working on this issue.

4.)  Recently there has been some ShowerMax problem (corrupted calibration
    constant) upstream, which only happens once in a while, and Pulsar ended
    up being the early-warning system for the ShowerMax problem. We will work
    with the ShowerMax experts to fix the upstream problems and also try to
    understand more on the effect of upstream data corruption on the Pulsar
    system and improve the robustness of the overall system.

5.)  Update Data format CDFNote-4152 to document the additional bit "2B"
    in the TL2D bank for identifying whether it comes from Pulsar or Alpha.
    We will also need to upadte the note to include a section for the Pulsar
    diagnostic bank (the note has been written, just need to be included in
    the official CDFNote).


 In addition to the items listed above, we have spent quite a bit
of efforts since the review to finalize the DAQ readout code.  That was
one area that we had great concerns before.  The DAQ readout code
with the dual TL2D readout and scalerMon implementation has been
declared default.  The offline farm has processed some of those runs
already and we are running some physics analysis jobs to make sure the
data is okay.  Furthermore, to improve the readout time, we have
changed the readout configuration so that the TL2D bank is readout from
a separate crate (lower alpha).  In the past, we were reading out
Pulsar diagnostic banks and TL2D bank from the same crate.  We will
still need to do more testing but so far the readout code and
readout time appears to be no longer a major issue.


 To ease the overhead of the transition, we would like to urge
TDWG to start freezing (at the very least minimize) trigger algorithm
changes.  It's an additional overhead that is quite disruptive
to the commissioning effort.  We also think that now is the
time to move to the "overlap" period where we required alpha
and Pulsar build to be successful.  We want to make sure that
Pulsar can run all the trigger tables that alpha can.


 The default plan of the Pulsar group is still to begin on March 1st
the 1 week run with Alpha driving, Pulsar in auto-reject and Level-3
filters running to compare Pulsar and Alpha decisions.  Any
discrepancies will be filtered to D stream for offline analysis. The 2nd and 3rd of week, we will have Pulsar driving L2
and Alpha in auto-reject mode.  Depending on what we learn, we
may request from time to time some other special test runs during
the first three weeks of trial run.


Sincerely,



The L2 Pulsar Upgrade Group