Presentation is loading. Please wait.

Presentation is loading. Please wait.

Emanuele Leonardi PADME Weekly Meeting - LNF February 9th, 2018

Similar presentations


Presentation on theme: "Emanuele Leonardi PADME Weekly Meeting - LNF February 9th, 2018"— Presentation transcript:

1 Emanuele Leonardi PADME Weekly Meeting - LNF February 9th, 2018
Computing Emanuele Leonardi PADME Weekly Meeting - LNF February 9th, 2018

2 Optical link between IT and BTF
8-12 monomodal fibers 10Gbps LR (up to 3km) Rodent-proof 2 x 10Gbps LAT-IT connections available for PADME LAT CR 10Gbps connection between LAT and CR BTF T2 IT Available now Order underway 15/12/2017 E. Leonardi - WM Networking

3 E. Leonardi - WM 2018-02-09 - Computing
PADME central switch Server per L1 e disk storage Server per RC, DB, DCS e Online Monitor Trigger system DAQ node (l0padme4) DAQ VME crates 9/2/2018 E. Leonardi - WM Computing

4 10Gbps connectivity test
l1padme1 and l1padme2 (disk buffer nodes) are now connected with optical fiber at 10Gbps to the PADME central switch Network speed test (TCP) confirms about 9.4Gbps link between the two nodes Data transfer disk to disk via scp (not very efficient): l1padme1 ↔ l1padme2 180MB/s l0padme4 ↔ l1padme MB/s (l0padme4 on 1Gbps link) 9/2/2018 E. Leonardi - WM Computing

5 Event Format (DAQ & RAW)
The DAQ and RAW event formats currently in use were defined in 2015 and were used in all testbeams since. Well suited for testbeam needs but lack some general information (e.g. global timing) In order to support a full data acquisition some revamping of the formats is needed. 9/2/2018 E. Leonardi - WM Computing

6 DAQ data format Added board serial number to file header
Enlarged event status bit pattern Added code of 0-suppression algorithm used Status pattern b0: [0=event is empty|1=event has data] b1: [DRS4 corrections 0=not applied|1=applied] (b2,b3): [0-suppression 00=off, 01=flagging, 10=rejection, 11=reserved]) b4: board fail (BF) flag (b5-b9): reserved 9/2/2018 E. Leonardi - WM Computing

7 RAW data format TADCBoard UChar_t fBoardId UInt_t fBoardSN UShort_t fLVDSPattern fBoardStatus fStatus (obsolete) fGroupMask fEventCounter fEventTimeTag f0SuppAlgrtm fActiveChannelMask fAcceptedChannelMask fNADCChannels TClonesArray* fADCChannels fNADCTriggers fADCTriggers TADCChannel UChar_t fChannelNumber Short_t fSamples[1024] TADCTrigger UChar_t fGroupNumber UShort_t fStartIndexCell fFrequency Bool_t fTriggerSignal UInt_t fTriggerTimeTag Short_t fSamples[1024] TRawEvent UInt_t fRunNumber fEventNumber TTimeStamp fEventAbsTime fEventRunTime fEventTrigMask fEventStatus UChar_t fNADCBoards TClonesArray* fADCBoards fBoardSN, fBoardStatus, and f0SuppAlgrtm are copyed from the DAQ data format N.B. fStatus is kept for downward compatibility. Will disappear in the future. fEventAbsTime is the global UTC time on the L1 server when the event is merged. Servers use NTP so timing should be good down to the ms level. fEventRunTime is the time from the start of the run. It is computed taking into account the timing information of all the boards (which should in principle be well synchronized). fEventTrigMask and fEventStatus are 32 bits patterns reserved for future use. As the new format is a superset of the previous format and the old class interfaces are respected, all code currently in use for analysis should in principle work with files in the new format (need to recompile using the new libraries, of course). I am now studying how to create code that can use both old and new format (possible according to ROOT).


Download ppt "Emanuele Leonardi PADME Weekly Meeting - LNF February 9th, 2018"

Similar presentations


Ads by Google