Research

While I'm waiting for one of my collegues to come and pick me and my stuff up, I thought I'd spend some time explaining what I've been doing here, research wise. I actually spend most of my time here in front of my computer writing, running or debugging code, despite what it seems like from the previous entries, when I'm already going out and doing stuff.

So for the past couple of months, I've been working on the Level 2 ATLAS Electron Trigger. Now I know that for most people, this makes no sense whatsoever, so I'll break it down.

I recall that in a previous post, I gave an introduction about ATLAS, a general purpose precision particle physics detector located at one of the collision points on the LHC ring. Operating at design specifications, the detector will see one collision every 25 nanoseconds, and with each event generating a megabyte of data, it is physically impossible to record all this information to disk within a reasonable budget. Not to mention, not all of these collision events will be interesting from a physics point of view.

So ATLAS employs what we call a trigger system, to select which events look promising and store them for offline processing and analysis. ATLAS has a three level trigger, the first implemented in onboard electronics and the other two using specialised software. I work on the first of the software systems. In brief, I am modifying a cluster seeding algorithm so that it doesn't lose a certain type of physics event from a specific particle. The algorithm I am implementing is a two-dimensional sliding window algorithm, performing energy sums to find the highest three by three area within a larger grid to seed cluster formation in subsequent algorithms.

Okay, that paragraph is very technical and I don't know where to start explaining everything. I keep using the word cluster, as the signature of an electron in the ATLAS detector is a curved track in the Inner Detector and some energy deposited in the Calorimeter. Now the way an electron deposits energy only depends on the initial energy of the electron and the material. The energy deposit is called a cluster because the Calorimeter is made out of cells, and when an electron goes through, a group or cluster, of these register hits.

I am finding it quite hard to explain this stuff coherently without feedback from somebody listening, to ask questions and to gauge reaction. Anyway, this work has been quite slow, firstly trying to understand the ATLAS trigger software, and the electron slice in particular. Then figuring out what information was accessible to me for use in the algorithm, write the code, then test its physics and timing performance, as there are fairly strict timing budgets for each of the ATLAS trigger levels. And that is kind of where I am now. Testing the simpler version of the algorithm and implementing a more complicated version. The simpler version involves just a straight out substitution of the old code for my code, and the new one involves changing the interface of the class.

So that is one of the things I am doing at the moment, I think I'll leave the description of the other things for another day...