June 27, 2018 online meeting minutes: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
present: Sergey Boyarinov, | present: Sergey Boyarinov, Stepan, Ben, Hai, Ed, Graham | ||
We discussed a progress in DAQ/Trigger upgrades for upcoming run. | We discussed a progress in DAQ/Trigger upgrades for upcoming run. | ||
FADC bit packing: on last meeting we planed that in the end of June we may have something to test, and will try to get one FADC250 version for everybody | * FADC bit packing: on last meeting we planed that in the end of June we may have something to test, and will try to get one FADC250 version for everybody; Hai finished simulation and passed to Ed; Ed works with Ben and expect first version for testing by July 3; FADC busy may not be there in first version, will be implemented soon after; Bryan Moffit has to change library to handle a couple of extra control bits; Sergey has to change rol2 to understand packed data; have to tell Gagik to put decoding in | ||
DC road finder: | * DC road finder: last time we discussed to have first version mid of July; | ||
TI/TD/TS firmwares and libraries: Hall D is finished testing, observed significant improvements in stability, some issues may still exist in fiber length measurements; | * TI/TD/TS firmwares and libraries: Hall D is finished testing, observed significant improvements in stability, some issues may still exist in fiber length measurements; | ||
Procurements: TD boards, VTP repairs approved; VXS crates, fileserver+3 IOCs, CPUs, DSCs, DT->TI to go | * Procurements: TD boards, VTP repairs approved; VXS crates, fileserver+3 IOCs, CPUs, DSCs, DT->TI to go | ||
* Online farm: | |||
CLAS12 Event Recorder testing with different raid configurations, number of partitions and number of streams on clondaq6 with Paul Letta; found that raid0 single partition has highest performance, raid5 single partition works as good as raid0s; multi-partition does not give any advantages and requires more complex logic; 4 streams seems optimal; inserting data into ET grand station data rate about 2.5GB/sec was achieved on empty disk and free clondaq6 [https://clonwiki0.jlab.org/wiki/clondocs/Docs/raid5_4stream_1partition_nextday.png (runcontrol screenshot)]. | * CAEN TDC calibration: | ||
* CLAS12 Event Recorder testing with different raid configurations, number of partitions and number of streams on clondaq6 with Paul Letta; found that raid0 single partition has highest performance, raid5 single partition works as good as raid0s; multi-partition does not give any advantages and requires more complex logic; 4 streams seems optimal; inserting data into ET grand station data rate about 2.5GB/sec was achieved on empty disk and free clondaq6 [https://clonwiki0.jlab.org/wiki/clondocs/Docs/raid5_4stream_1partition_nextday.png (runcontrol screenshot)]. | |||
Revision as of 09:21, 27 June 2018
present: Sergey Boyarinov, Stepan, Ben, Hai, Ed, Graham
We discussed a progress in DAQ/Trigger upgrades for upcoming run.
- FADC bit packing: on last meeting we planed that in the end of June we may have something to test, and will try to get one FADC250 version for everybody; Hai finished simulation and passed to Ed; Ed works with Ben and expect first version for testing by July 3; FADC busy may not be there in first version, will be implemented soon after; Bryan Moffit has to change library to handle a couple of extra control bits; Sergey has to change rol2 to understand packed data; have to tell Gagik to put decoding in
- DC road finder: last time we discussed to have first version mid of July;
- TI/TD/TS firmwares and libraries: Hall D is finished testing, observed significant improvements in stability, some issues may still exist in fiber length measurements;
- Procurements: TD boards, VTP repairs approved; VXS crates, fileserver+3 IOCs, CPUs, DSCs, DT->TI to go
- Online farm:
- CAEN TDC calibration:
- CLAS12 Event Recorder testing with different raid configurations, number of partitions and number of streams on clondaq6 with Paul Letta; found that raid0 single partition has highest performance, raid5 single partition works as good as raid0s; multi-partition does not give any advantages and requires more complex logic; 4 streams seems optimal; inserting data into ET grand station data rate about 2.5GB/sec was achieved on empty disk and free clondaq6 (runcontrol screenshot).
