Trigger Database and Trigger Configurations and Trigger Issues
Trigger Database and Trigger Configurations and Trigger Issues Elizabeth Gallas, Jeremy Simmons (Fermilab - Computing Division) See my Trigger Working Page http: //www-d 0. fnal. gov/~gallas/trigger. html Database meeting January 07, 2002 Jan 7, 2002 E. Gallas/Trigger Db 1
Recent Success • Trig. Db generated Online configurations • Rate Tests: – ratetest_L 1 Cal-1. 00 - 31 triggers – ratetest_L 1 Cal. Muon-1. 00 - 39 triggers • first time more than 32 trigger bits were used online – ratetest_L 1 Cal_3 and 4 jet-1. 00 - 10 triggers – ratetest_L 1 Cal. Muon-1. 10 - 53 triggers • largest number of triggers used online to date • Physics datataking: – global_Cal. Muon-3. 00 - 33 triggers – global_Cal. Muon-3. 10 - 48 triggers • created by Levan while I was away. Only one minor problem with xml generation because new feature was being used. Quickly identified, fixed. • GOOD: No changes in trigger db server – and no anticipated changes until a schema change is required such as that needed to include release or stream or rate info • minimal impact from downtime/restarts – server or database on production machine Focus remains on client code, entering needed data, reporting, ‘xml’ output Jan 7, 2002 E. Gallas/Trigger Db 2
High Priority Activity (EG) • Need to get Muon V 2. 00 terms in the database and update current TL to use it – high priority because more triggers exist than what we are using now. • Prescale generator (need 3 command line ‘programs’ to run online) – Needs a list of trigger names for a TL and enter their prescale class (psc) assignments • fixed rate, no prescale, usually no prescale, hopefully no prescale, always some prescale, always high prescale. . . – Generates normalized to min_bias rate for the dynamic comparison of rates to previous measurements. – Read generated/measured rates to create a prescale file for use online that will allocate bandwidth according to the psc assignment. Jan 7, 2002 E. Gallas/Trigger Db 3
Also working on. . . • XMLGENerator – modifications, additions of features • document these modes in the help feature so I don’t forget that they exist, also provides some doc • Worried about differences in online and simulation xml at Level 3 (and perhaps at Level 2) as long as we are generating the ‘xml’ file. • Client ENTRY interface – many bug fixes, improvements • discovered as new triggers are entered - require minimal changes in code (they just take time to find) – new: Every valid trigger list uses L 1 NEOTERMS from one Nty. Version for each subdetector. • need program to copy/create a new trigger list using one Nty. Version to another Nty. Versions (names generally stay constant). • Current example is Muon V 2. 00, though in that case all the names have changed (not typical). • Client REPORT interface – many known changes need to be made to make this fully functional – start a list of additional reports that would be useful (but would not fall under this interface) that might be written by someone other than EG or JS. Jan 7, 2002 E. Gallas/Trigger Db 4
Other thing on TODO list Proposals requiring schema, server and client changes – Ideal trigger rates from lum for prescale gen • prescale generator mentioned - L 1 only • with L 2 and L 3: may ultimately factor into load balancing of streams into file families – Tying TL to L 2, L 3 Releases – Streams DATABASE (same as last time) – Entering backlog of data changes, some coordination with xml changes – finding/removing junk records from database (affect status changes) – stat. Used triggers - some not quite right, figure out/fix xml generator – Level 2 configuration may come online soon! Need to be prepared to include that when it is available. – new proposal from COOR, L 2, L 3 - should simplify - no agreement at this point. Jan 7, 2002 E. Gallas/Trigger Db 5
Phase 1: last items still remain • Jeremy: – STAT_CURR (2 -3 days) • select records by PK • allow a number of changes at one time. – L 1 Dialogs entry (3 days) • Entry of data into one table • using EG’s template of entry programs, suggest ways to improve design of client programs, implement user actions (phase 2), Java. Script. . . • Elizabeth – NEOTERMS (3 days) • working but in debug phase • debugging using entry of Muon v 2. 00 – Forgot generation of ‘tools, filter. rcp’ (4 d? ) • Rich wrote server methods last summer • needs easier setting of STAT_CURR • needs client interface (web or com line? ) • need L 3 contact for long term plan (rel? ) Jan 7, 2002 E. Gallas/Trigger Db 6
TO DO Summary • Phase 1 (in progress) - Feb 14, 2002 • Phase 2 (client code) - April 2002 – Implement User dependent actions • give different roles different buttons • allow users to update, delete their ‘own’ elements – L 2 Preprocessor • association with L 2 global terms – Devices, Device Groups for Exposure and TL • Phase 2. 5 (Schema, server, client changes) July, Aug 02 – Ideal Trigger rates for prescale calculations – Streams – Releases • Phase 3 (more client methods) – NTY Simulations – Prescale Sets • Phase 4 (design is incomplete) – L 1 Simulation to Firmware – L 1 board registry (no server methods exist) Jan 7, 2002 E. Gallas/Trigger Db 7
- Slides: 7