Difference between revisions of "ERSAP Meeting Feb. 4, 2021"

From epsciwiki
Jump to navigation Jump to search
(Created page with "=== Connection Info: === This meeting scheduled via Zoom. Join link should be in your calendar. <!---------------------------------------------------------------------------...")
 
 
Line 34: Line 34:
 
#** Partial event building and processing with possible data reduction in a sub-detector streams was also discussed. This might be useful to achieve data transfer optimizations, as well as optimize a resource utilization.
 
#** Partial event building and processing with possible data reduction in a sub-detector streams was also discussed. This might be useful to achieve data transfer optimizations, as well as optimize a resource utilization.
 
#* TriDAS and ERSAP integration possible scenarios
 
#* TriDAS and ERSAP integration possible scenarios
#** The decision was made to perform test and benchmark measurements for TriDAS and ERSAP components on a more controlled testing environments (e.g. Hall-B and Indra lab), where we can configure data streams, and measure parameters such as event/data and lost stream-frame rates.  
+
#** The decision was made to perform tests and benchmark measurements for TriDAS and ERSAP components on a more controlled testing environments (e.g. Hall-B and Indra lab VTP setups), where we can configure data streams, and measure parameters, such as event/data and lost-frame rates.  
 
#** Later optimized components with their suggested compositions (e.g. HM, TCPU, or HM+TCPU, etc.) will be integrated into the ERSAP.
 
#** Later optimized components with their suggested compositions (e.g. HM, TCPU, or HM+TCPU, etc.) will be integrated into the ERSAP.
 
#* ERSAP-TriDAS template project proposal
 
#* ERSAP-TriDAS template project proposal
 
#** Vardan mentioned a new project that he started for the TriDAS integration (here is the git repository for the [https://github.com/JeffersonLab/ersap-tridas.git project]). Work is still in progress, and will be discussed in more details during the upcoming meetings.  
 
#** Vardan mentioned a new project that he started for the TriDAS integration (here is the git repository for the [https://github.com/JeffersonLab/ersap-tridas.git project]). Work is still in progress, and will be discussed in more details during the upcoming meetings.  
 
Vardan presented [https://userweb.jlab.org/~gurjyan/presentations/ersap_02.04.pptx slides] that was used during the discussion.
 
Vardan presented [https://userweb.jlab.org/~gurjyan/presentations/ersap_02.04.pptx slides] that was used during the discussion.

Latest revision as of 13:52, 11 February 2021

Connection Info:

This meeting scheduled via Zoom. Join link should be in your calendar.


Agenda:

  1. Announcements
  2. Streaming readout and streaming data processing design requirements
  3. TriDAS and ERSAP integration possible scenarios
  4. ERSAP-TriDAS template project proposal
  5. AOT

Useful References


Minutes:

Attendees: Tommaso C., Carmelo P., Benjamin R., Sergey B., David A., Carl T., Chris L., Vardan G.

    • Announcements
      • We all agree to go with biweekly meeting format, scheduled every other Thursday at 9AM Eastern Time (US and Canada).
    • Streaming readout and streaming data processing design requirements
      • An e-mail exchange between Carmelo and Vardan was mentioned, where they discussed possibility to deploy and operate HM and TCPU in a single run-time, thus preparing it to function as a component within the ERSAP ecosystem.
      • We discussed known back-end processing systems, namely batch processing, stream processing and micro-batch processing.
      • Lengthy discussion on the data handling requirements. Questions such as: Can we write/store data efficiently? Can we correlate streams and build events fast enough not to introduce blind spots in the stream?
      • Partial event building and processing with possible data reduction in a sub-detector streams was also discussed. This might be useful to achieve data transfer optimizations, as well as optimize a resource utilization.
    • TriDAS and ERSAP integration possible scenarios
      • The decision was made to perform tests and benchmark measurements for TriDAS and ERSAP components on a more controlled testing environments (e.g. Hall-B and Indra lab VTP setups), where we can configure data streams, and measure parameters, such as event/data and lost-frame rates.
      • Later optimized components with their suggested compositions (e.g. HM, TCPU, or HM+TCPU, etc.) will be integrated into the ERSAP.
    • ERSAP-TriDAS template project proposal
      • Vardan mentioned a new project that he started for the TriDAS integration (here is the git repository for the project). Work is still in progress, and will be discussed in more details during the upcoming meetings.

Vardan presented slides that was used during the discussion.