Message boards :
Sixtrack Application :
10^5 in name, 10^6 in nature
Message board moderation
Author | Message |
---|---|
Send message Joined: 29 Sep 04 Posts: 281 Credit: 11,859,285 RAC: 0 |
I would normally expect 100,000 turn tasks to run for a little over an hour and million turn tasks to be 10 hours or more. However, much of this recent batch of work has seen tasks declared as 105 taking as long as a typical 106 job eg. LHC_2015_LHC_2015_234_BOINC_errors__15__s__62.31_60.32__4.9_5.0__5__55.5_1_sixvf_boinc51308_0 which ran for 13 hours. Has the naming convention been dropped or are these jobs looking more in depth to account for the longer runtime? |
Send message Joined: 29 Feb 16 Posts: 157 Credit: 2,659,975 RAC: 0 |
Hello Ray, thanks for the acute question. No, those WUs are truly 10^5 turns. Contrary to many other cases with 10^5 turns, these tasks last longer, as they include the physics of beam-beam effects, i.e. non-linear forces between bunches at the interaction points (including parasitic encounters), which are CPU-expensive. The usual 10^5 turns jobs you are used to take into account only magnetic errors (which are treated at initialisation, but then they add 0 CPU extra time). Hope it helps! Cheers, A. |
Send message Joined: 29 Sep 04 Posts: 281 Credit: 11,859,285 RAC: 0 |
Thanks for that explanation, Alessio. I was just curious. Sixtrack was one of my first Boinc projects so I try to catch as many as I can, when they're available. |
Send message Joined: 29 Feb 16 Posts: 157 Credit: 2,659,975 RAC: 0 |
Thanks a lot for your care, Ray, and for being generous with us! All my colleagues and I are very grateful! Happy crunching! |
©2024 CERN