Project

General

Profile

27-Sep-2018

Tomo -- New benchmark fully utilizing node .. only 30% worse. One thread had a substantial difference on the better nodes. He noticed, the report is per thread,.. And sometimes only reports 15 threads (Tomo is using fcl file to configure a certain number of threads).. Also distribution of events to threads is uneven. First event takes a lot longer, on each thread,.. Like a second. Looking into running the profiler. Eric and Ron.. say art is known for slow opening of output file (ROOT is doing a sync).

Bertrand -- Starting to look at trigger menus.

Glenn -- Communicating with rack monitor box. Reliably gives data back. Has been exercising mu2edcsdev01. Let’s wait a week, and see if rack monitor box is still running. Matthew Jones has reached out.

Eric -- working with Rick a bit to get CFO utility going (that can handle run plan inputs). New version of artdaq out today. Will try to update mu2e_artdaq.

Rick -- working on CRV version of DTC code (fanning out punched clock) -- two different versions of firmware, there is a #define switch before place and route. Tested the v2 timing card a bit further, except RJ45 connector. Two weeks to get communication back for “open” DTC. Next is working on automatic resest sequence.. Then back to CFO to enable the run plan.

Greg -- Working on sync demo between CFO and DTC (not ROC yet). Having issues: (1) is the kernel driver that ots accesses the same as the chants kernel driver? (2) can not reconfigure and get loopback working.. The first time after bootup, the loopback works repetitively.

Jose -- Working on ROC firmware for sync test.. Working to fix fixed latency delay. Trying to add ROC monitor so reset sequence can be driven by software. Having place and route timing issues. 1% utilization. Multipasses might indicate luck involved.

Ron -- Will be testing new artdaq

Ryan -- working with Greg and Jose