L1 Calo Software meeting at RAL on 30 January 2001.
---------------------------------------------------

Present: Bill, Bruce, Eric, Gilles, Murrough, Norman, Steve.


Actions from previous meeting:
------------------------------

Norman:
- has started assembling all Glink and Slink data formats into and out
  of our RODs into a single document. To be continued...
- has also started on a document about the VME-- spec and VME address
  space of our modules. To be continued...
- and has ordered a new PC for the lab
- but has not yet ordered new VME CPUs. In discussion we agreed they
  should have new SCSI disks
  
Murrough:
- has produced a new version of the TTC document after consultation
  with Norman. However it doesnt contain the existing DSS broadcast
  commands which should be added. Bruce isnt happy with the proposed
  addressing scheme and wants to try to think of a better one...

Steve:
- has done a crate and disk audit. There is one suitable crate at Bham
  and five at RAL. The LynxOS systems have small, probably unsuitable
  disks.
  
Bill & Bruce:
- have exchanged information about Bills test vector generators. Some
  bit swapping has been identified and will be fixed soon.
  
  
Work on Run Controllers
-----------------------

Murrough has been starting development with the Online software and has
created a hardware configuration database with an approximate slice test
configuration of crates, modules and run controllers. Some bugs in the
latest release (0.0.12) have been reported. Tests have been made with
run controllers distributed several across QMW Linux nodes.

Future work includes more on the database (calibrations, trigger menu),
information service data blocks, IGUI panels to display/set them and
implementation of hardware actions using a layer of HDMC classes. The
latter needs more discussion with eg Bruce.


Work on the ROD
---------------

Bruce has been using both internally generated Glink packets from the DSS
and data generated via BIlls program and has successfully seen packets from
the ROD, triggered by the TTC system. As mentioned above, some incorrectly
set bits are apparent in the test vectors.

Now that the ROD is shown to work for single events, more exhaustive tests
are planned. Also the ROI firmware needs to be checked as well. At some
point we should contact the ROIB people to set a suitable date (and plan)
for the joint test at CERN.

It would be convenient if the two FPGAs on the ROD could be loaded with
the two different algorithms. Then both the ROD in our lab and that in the
electronics lab could test both without intervention to reload EEPROMs.



Discussion on Test Vectors
--------------------------

This was a long session, spanning lunch. In the morning we covered a number
of functional tests related to the ROD and the CPM. We need to come up with
tests for each of the functional blocks or links in the system. For the CPM
this means the LVDS links, the BC Demuxing, the Serialiser and its readout,
the backplane data fanout, the CP chip algorithm and its ROI readout, the
hits sent to the CMM. Similar details are required for the other modules.

After lunch we discussed a wide ranging list from Eric of areas through the
whole system which need covering.

We decided it would be useful to try to capture and develop this discussion
by producing summary "test/explotation plan" type documents for each module
and for the whole system. Birmingham (Steve plus Bill and Gilles) will do
this for the CPM, Bruce for the ROD, Norman for the CMM and a Rutherford
based group (Norman, Eric, Bruce) will look at the whole system.

The aim is to produce initial versions for our next software meeting which
can then be circulated before the joint meeting where they can be presented
(at either software or main meeting) and discussed.


Other Items
-----------

- HDMC support: we are concerned about the future of HDMC support. We hoped
  that Heidelberg will continue to be able to offer support in future.
  
- Software organisation: Murrough has produced a motherhood and apple pie
  document with a few suggestions. No brickbats received so far...
  We discussed the possibility of converging/integrating web and other
  documents without reaching a conclusion.
  
- Calibration: we still arent paying much attention to this. But Murrough
  has contacted Lorenzo Moneta who is developing the ROD crate controller
  for the LAr system and has suggested a chat about calibration issues at
  some point during the forthcoming ATLAS week.
  
- Birmingham meeting. There is a _very_ provisional agenda for the software
  meeting in Birmingham. Suggestions welcome!
  See here.
  
  
AOB
---

- Steve has looked at Paul Bright Thomas draft note on numbering schemes
  and would like to make it fixed length for easier parsing by software.
  The proposed convention is not very obvious for crates and modules in
  slots - we will want a conversion.
  The topic of numbering schemes should be raised again at Birmingham.
  
- Norman: after the slice tests we should have a full scale review of our
  software. Are we doing the right thing in all areas from HDMC to DAQ.

  
Next meeting
------------

The next meeting will be at RAL on Monday 12 February. We should invite
some of the electronics group to identify areas of module testing that
they already cover well so we can concentrate our efforts on the rest.


Summary of Actions
------------------

Norman: finish data format, VME, slice test config documents
Norman: buy VME CPUs, disks, etc
Norman: produce CMM test/exploitation plan
Steve/Bham: produce CPM test/exploitation plan
Bruce: produce ROD test/exploitation plan
"RAL": produce overal slice test/exploitation plan