MantisBT - Hall D Online
View Issue Details
0000092Hall D OnlineDAQpublic2011-06-17 15:572014-11-30 08:40
wolin 
davidl 
normalfeatureN/A
resolvedfixed 
Yes
8
0
0
0000092: Design disentangling architecture and develop prototype event disentangler
Design architecture appropriate for event disentangling in the Hall D DAQ system. Design a CODA event building component that disentangles events prior to the farm system and event recorder such that event data is contiguous in memory. Software should include working example that disentangles data from F1TDC, fADC125 and fADC250 modules that will be used in Hall-D. Develop and test prototype in realistic environment. Deliver report that summarizes results, recommends a production architecture and includes an outline of steps needed to create and commission production disentangler with estimates of required time and manpower.
Consult as required with the Hall D Online group concerning the specific requirements and constraints of the Hall D computing system. Major architectural decisions must be reviewed by the Hall D Online group to ensure compatibility with the experiment online environment.
No tags attached.
Issue History
2011-06-17 15:57wolinNew Issue
2011-06-17 15:57wolinWork by outside group for Hall D => Yes
2011-06-17 15:57wolinEstimated man-weeks => 8
2011-06-17 15:57wolinPercentage complete => 0
2011-06-17 15:57wolinActual man-weeks => 0
2011-06-17 16:23davidlDescription Updated
2011-06-20 10:12wolinStatusnew => assigned
2011-06-20 10:12wolinAssigned To => heyes
2011-06-20 10:16wolinNote Added: 0000120
2011-06-20 10:17wolinNote Added: 0000121
2011-06-20 10:17wolinStatusassigned => feedback
2011-08-31 07:28heyesNote Added: 0000148
2012-01-23 10:04wolinNote Added: 0000251
2013-03-19 11:56wolinNote Added: 0000478
2013-03-21 14:04wolinAssigned Toheyes =>
2013-03-21 14:04wolinNote Added: 0000503
2013-03-21 14:04wolinStatusfeedback => new
2013-07-18 10:12wolinNote Added: 0000556
2014-11-30 08:40davidlNote Added: 0000600
2014-11-30 08:40davidlStatusnew => resolved
2014-11-30 08:40davidlResolutionopen => fixed
2014-11-30 08:40davidlAssigned To => davidl

Notes
(0000120)
wolin   
2011-06-20 10:16   
Graham...kindly review the task description and modify as required, especially the estimated man-weeks of effort needed. We'll iterate as needed.
(0000121)
wolin   
2011-06-20 10:17   
Please update task description as needed
(0000148)
heyes   
2011-08-31 07:28   
"Software should include working example that disentangles data from F1TDC, fADC125 and fADC250 modules that will be used in Hall-D".

It isn't clear that this is possible at this time since the data format from the various modules isn't stable yet. A prototype "proof of principle" is possible.
(0000251)
wolin   
2012-01-23 10:04   
Probably can disentangle in quad-core or better ROC or in EMU stage. DAQ group will determine how much CPU is available in ROC at high-rates.
(0000478)
wolin   
2013-03-19 11:56   
I believe Carl has made progress on this. Different strategies for where to run the disentangler (ROC, EB, both?) will be investigated in the Hall D DAQ test setup by the DAQ group.
(0000503)
wolin   
2013-03-21 14:04   
Not sure who will do this.
(0000556)
wolin   
2013-07-18 10:12   
Dave and I now realize the two best places to do disentangling are on the ROC and in the L3 farm. I.e. a separate disentangling EB stage may not be the best choice. We need to further investigate these two strategies.
(0000600)
davidl   
2014-11-30 08:40   
Event disentangling is currently being handled in the DAQ plugin in the Hall-D software. It will likely stay there since it allows the disentangling job to be distributed to multiple computers rather than be done on a single computer if it were handled in the event builder.