It’s been quite a week… Spectacular northern lights for hours on Thursday night. A great comet in the evening skies (though so far I’ve have only caught glimpses, thanks to atrocious viewing conditions.) And now, I’m at CERN (the pan-European particle physics laboratory) for the first time since the pandemic began. I’ll be giving a talk at a conference of CMS experimenters. (CMS and ATLAS are the two general purpose experiments at the Large Hadron Collider [LHC].)
The topic of the workshop is a novel technique called “Level-1 Scouting” — though it isn’t really about “scouting” for anything. It has to do with evading the strait-jacket of the trigger, an essential feature of data gathering at each of the LHC experiments. With tens of millions of collisions per second, the data flood at CMS is too great, and only a tiny fraction of these collisions can be stored. The trigger decides real-time which ones to keep and which ones to discard forever. That’s been the basic rule since the LHC began running.
But this rule no longer applies, thanks to new technology and human ingenuity. CMS now uses level-1 scouting to record sketchy information about every single collision that happens in their detector. LHCb, with a smaller detector, was the first to try something along these lines. ATLAS is on a parallel track. These developments have the potential, looking ahead, to substantially enhance the capability of these detectors. More about this after I’ve given my talk.
2 Responses
Very cool. I’ve long wondered what might be getting lost in all that data.
Guess we’ll find out soon!
To be fair, though, we mostly know the most common things that are “in all that data”, because there are “prescaled triggers” that take a small fraction (chosen at random) of events that would otherwise be discarded. That’s not good for discovering a relatively rare process that might show a breakdown of known physics, but it is a way of measuring details of very common phenomena that are hard or impossible to calculate, and to check whether the detailed model of the experimental detector might be missing small but important effects.