EPSCI SRO Meeting May 4, 2020
Jump to navigation
Jump to search
Connection Info:
This meeting scheduled via MS Teams. Join link should be in your calendar.
Agenda:
- Previous Meeting
- Announcements
- Workshop May 13-15
- Talks on JANA2(David L.), CLARA(Vardan), data transport (Carl), CODA->SRO (David A.)
- include Graham et al's work on data format/content?
- https://www.jlab.org/indico/event/378/
- Talks on JANA2(David L.), CLARA(Vardan), data transport (Carl), CODA->SRO (David A.)
- EIC YR meeting last Wed.
- Status of stand-alone TriDAS
- Status of ADIOS Testing
- Use of Hall-B testbeds
- Status of WMS design
- AOT
Useful References
Minutes:
Attendees: David L., Carl T., Jie C., Nathan B., Vardan G., Chris L., Graham H., David A.
SRO Workshop
- 5 SRO software related talks from JLab
- Carl will include some data formatting from Graham's secret document
Hall-B Testbed
- VTP can output 4 x 10Gbps streams (must be >=4 streams)
- Currently only 3 streams active in second testbed that Vardan is using
- Dave A.: Need some type of sync event/time frame
TriDAS
- Chris compiled on desktop and started investigating code
- Design drops frames if problem keeping up with rate
- Simulator crashes (not sure why yet)
- Uses BOOST heavily
- Still unclear extent of adaptability
- parallelize single task?
- System design has "Towers", "floors", and "sectors" integrated throughout design.
- Uses ssh to start remote processes
ADIOS Testing
- Indra seemed to limit to 1 Gbps
- Tried "blaster" (28 yo code) - same rate
- played with TCP buffer size and got up to 4 Gbps (though not stable)
- Graham's update software got ~4 Gbps per link
- Limit on indra machines buffer size
- Adjusted Linux settings and got up to 5 Gbps
- Full bandwidth with multiple links
- (What are gluons set to?)
- Dave A. will talk to Bryan H. about setting switches to support jumbo frames
Specs Document
- Need to start a specifications document for what JLab SRO projects will require
- Vardan will create shared document so all can start to contribute