 |
 |
There will be 5 unrecognized channels every event for the data that is not assigned to crystals but is in the same bank. If there are more than this, you should check the log to make sure the run number is being used correctly. Monte Carlo data all has run number zero so it is not good for testing this. I'm pretty sure the cosmic runs should work fine here, or at least last I checked it this was working fine. > On Nov 17, 2014, at 7:33 PM, Nathan Baltzell <[log in to unmask]> wrote: > > To be sure, I did a rigorous test of all 442 channels from the conditions DB at jlab for table "ecal_channels". > And the match is perfect regarding (x,y) <-> (crate,slot,channel) with respect to the working table I have that > we believe to be physically correct. This was done on collection_id=2, which is currently the active set for > 2000<=run<=10000. > > -Nathan > > >> On Nov 17, 2014, at 9:23 PM, Nathan Baltzell <[log in to unmask]> wrote: >> >> Hi All, >> >>> Was there more that was needed for the mapping of channels between the detector geometry and the DAQ? >> No, (crate,slot,channel) uniquely determines the calorimeter channel (x,y). >> >> I just now checked by eye a dozen random channels in the ecal_channels tables in jlab's conditions database, >> and they are as I expect regarding (crate,slot/channel)<->(x,y). I did not do a full 442 test, but could. >> How many channels are not recognized? >> >> -Nathan >> >> >> >> >> >> >>> On Nov 17, 2014, at 5:21 PM, "McCormick, Jeremy I." <[log in to unmask]> wrote: >>> >>> Hi, >>> >>> I'm not sure about this, but doesn't the set of EcalChannel conditions replace what we used to use for the DAQ map? >>> >>> It has a mapping of geometric IDs (ix and iy) to the crate, slot and channel. >>> >>> Was there more that was needed for the mapping of channels between the detector geometry and the DAQ? >>> >>> If there is something we left out then let me know and we can add any additional classes needed by the ECAL Drivers, >>> but it seems to me that EcalChannel has all the information you should need. >>> >>> --Jeremy >>> >>> -----Original Message----- >>> From: Andrea Celentano [mailto:[log in to unmask]] >>> Sent: Monday, November 17, 2014 2:14 PM >>> To: McCormick, Jeremy I. >>> Cc: Nathan Baltzell >>> Subject: Re: aida.IPlotter >>> >>> This can explain the errors I got with the ECAL DAQ monitor driver (some channels not recognized) However, the ECAL reconstruction seems to work properly, so it is loading the ecal DAQ map from "somewhere".. >>> Wasn't the file I sent around with conditions loaded on the database? >>>> On 11/17/2014 11:10 PM, McCormick, Jeremy I. wrote: >>>> Hi, >>>> >>>> It looks to me like we never ported the ECAL DAQ map to the new conditions system... >>>> >>>> --Jeremy >>>> >>>> -----Original Message----- >>>> From: Andrea Celentano [mailto:[log in to unmask]] >>>> Sent: Sunday, November 16, 2014 3:02 PM >>>> To: McCormick, Jeremy I.; Nathan Baltzell >>>> Subject: Re: aida.IPlotter >>>> >>>> Hi, >>>> currently on the trunk: >>>> >>>> - All the drivers that are involved in EcalLedCommissioning.lcsim steering file use the new condition system. This steering file works with the monitoring app >>>> >>>> - I am working on the drivers involved in the EcalMonitoringFinal.lcsim steering file, but I have problems. >>>> In particular, I am not sure that EcalDaqPlots is using the new condition system, since the DAQ map that I get seems to me the old one. >>>> Jeremy, can you please have a look at it? In particular, can you check the "detectorChanged" method? >>>> >>>> In this method, I want to determine which DAQ slots are used for ECAL top and ECAL bottom. I do this by performing a loop on all crystals, and using the conditions system to determine the corresponding crate/slot/channel. Then, I determine the slots that are used for the two halves. >>>> >>>> Andrea >>>> >>>>> On 11/16/2014 09:49 PM, McCormick, Jeremy I. wrote: >>>>> Hi, >>>>> >>>>> It may be that the threads aren't killed properly for other plots too, but it is just noticeable here because you click the event display hundreds of times. >>>>> I don't think this should be too terribly hard to fix. The previous >>>>> plot state just needs to be cleared up or the histograms should be >>>>> cleared and reused rather than recreated. I'll can look at it Monday >>>>> but... >>>>> >>>>> We need to update the LED Driver to use the database conditions, and >>>>> we need to get current, valid ecal_led conditions into the database. I don't think we should work on the ecal-commissioning branch any longer. When will this get done? It shouldn't be a lot of work, but someone just needs to do it. >>>>> >>>>> -Jeremy >>>>> >>>>>> On Nov 16, 2014, at 4:57 AM, Nathan Baltzell <[log in to unmask]> wrote: >>>>>> >>>>>> It only happens in the drivers from Andrea that working interactively >>>>>> by clicking on a crystal in Kyle's gui and then updating the histos. >>>>>> The neverending thread(s) is spawned in the crystalClicked method by >>>>>> the call to hep.aida.IPlotterRegion.plot (exactly one thread per >>>>>> plot call) at line 377 of >>>>>> org.hps.monitoring.ecal.plots.EcalEventDisplay >>>>>> and >>>>>> line 325 of org.hps.monitoring.ecal.EcalLedCommissioning >>>>>> and the lockup is definitely caused by too many threads. Steering >>>>>> file ECalLedCommissioning. Maybe Andrea's code is misusing the plot command. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>>> On Nov 16, 2014, at 7:06 AM, "McCormick, Jeremy I." <[log in to unmask]> wrote: >>>>>>> >>>>>>> Can you tell me exactly how I can reproduce this issue? I have not seen this while testing. I would need to know the steering file you are running and what you are doing during the session to cause the machine to lock up. Is this something that happens only with the interaction of the event display and the plotting? Or does it happen with simpler noninteractive usage? >>>>>>> >>>>>>>> On Nov 16, 2014, at 3:11 AM, Nathan Baltzell <[log in to unmask]> wrote: >>>>>>>> >>>>>>>> It doesn't technically crash, just eventually bricks the system by >>>>>>>> having spawned so many neverending threads that the system cannot >>>>>>>> do anything. Every time plot is called, we get another thread. >>>>>>>> -nathan >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>>> On Nov 16, 2014, at 5:56 AM, "McCormick, Jeremy I." <[log in to unmask]> wrote: >>>>>>>>> >>>>>>>>> Hi, >>>>>>>>> It could be a bug. Can you give me more info? What is the error traceback? >>>>>>>>> --Jeremy >>>>>>>>> ________________________________________ >>>>>>>>> From: Nathan Baltzell <[log in to unmask]> >>>>>>>>> Sent: Sunday, November 16, 2014 1:33 AM >>>>>>>>> To: McCormick, Jeremy I. >>>>>>>>> Cc: Kyle McCarty >>>>>>>>> Subject: aida.IPlotter >>>>>>>>> >>>>>>>>> Hi Jeremy, >>>>>>>>> >>>>>>>>> While on a beamless shift, Kyle and I were investigating why the >>>>>>>>> montioring app sometimes crashes. It seems to be because >>>>>>>>> hep.aida.IPlotterRegion.plot forks a new thread that never dies everytime it's called. In hps-java, this is called from org.hps.monitoring.ecal.plots. >>>>>>>>> Any ideas? >>>>>>>>> >>>>>>>>> -Nathan >>> >>> ######################################################################## >>> Use REPLY-ALL to reply to list >>> >>> To unsubscribe from the HPS-SOFTWARE list, click the following link: >>> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1 >> >> ######################################################################## >> Use REPLY-ALL to reply to list >> >> To unsubscribe from the HPS-SOFTWARE list, click the following link: >> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1 > ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the HPS-SOFTWARE list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=HPS-SOFTWARE&A=1