Princeton Student Finds Bug In LHC Experiment

An anonymous reader writes "A Princeton senior has found a bug in the hardware design for the Compact Muon Solenoid (CMS) experiment of the Large Hadron Collider (LHC). In the hardware used to record and capture events in the LHC, she discovered errors that were leading to the appearances of double images because of particle streams known as jets. 'Xiaohang Quan '09 was working on her senior thesis when she found a miscalculation in the hardware of the world's largest particle accelerator. Quan, a physics concentrator, traveled to Geneva, Switzerland, last week with physics professors Christopher Tully GS '98, Jim Olsen and Daniel Marlow for the annual meeting of the European Organization for Nuclear Research (CERN). This year, however, they also came to discuss Quan's discovery with the designers of the hardware for the Compact Muon Solenoid (CMS) experiment, which, as part of the Large Hadron Collider, has the potential to revolutionize particle physics.'"

Re:wha?

[scheme]

Her last name is "09" and she is a "concentrator?" Who wrote this?

It's from a student newspaper. Hence the 09 which refers to her graduation year. Also the concentrator part means that she's concentrating on physics. Some universities call it concentrating on a subject rather than majoring.

Re:wha?

[Quothz]

Her last name is "09" and she is a "concentrator?"

That threw me, too. The '09 appears to be standard form for the Princetonian, representing her (expected) graduation year.

Who wrote this?

Tasnim Shamma

Personal Info

* Degree: A.B. in English, IPS in Journalism

* Hometown: Jamaica, NY

* Contact Email: [e-mail]

Personal Bio

Princeton '11, Brooklyn Technical High School '07, Daily Princetonian news/blog/multimedia staff, Orange Key tour guide, Daily Princetonian Class of 2001 Summer Journalism Program Alum'06/ Program Staff Associate '08 (www.princeton.edu/sjp), Aspiring Reporter (if there are jobs left when I graduate) ;)

Off topic: Miss Quan is cute.

Not a hardware bug

[rminsk]

A Princeton senior has found a bug in the hardware design for the Compact Muon Solenoid (CMS) experiment of the Large Hadron Collider (LHC).

The bug was in the algorithm analyzing at the data from the CMS and not the hardware.

Re:Public Spin

[JoeBuck]

Once again: if there were any chance that LHC could produce earth-swallowing black holes, we'd be dead long ago, because Earth is regularly hit by much more powerful events than anything the LHC will be able to produce.

Re:they should not turn it on

[JoeBuck]

You are ignorant; the universe is already conducting high-energy physics experiments. They are called cosmic rays, and some of them are billions of times more powerful than the LHC. Yet the earth is still here. And your notion that we delay until we completely understand the laws of physics is comical. What do you think the LHC is for? It's to help us understand the laws of physics! You don't discover laws of physics by just thinking deeply. You discover them by experimentation.

Re:Great story.

[Gromius]

Its not actually amazingly impressive, its made to sound a lot more impressive than it actually is. One the meeting in question was "CMS week", one of several weeks a year we get all our collaborators together at CERN not CERNs annual meeting. She's basically improved our jet algorithm (as far as I can tell, the article is woefully lacking in details), a decent job for an undergraduate and will certainly help her walk into a PhD place as a shes clearly good enough but she's certainly not the only undergradute to have made a contribution such as this.

Re:wha?

[edittard]

Pity the "anonymous reader" didn't go the extra nine yards and become an invisible writer. Then we'd never have known what a waste of oxygen he was.

Re:they should not turn it on

[Gromius]

its not really a design flaw. Basically its a minor bug in the algorithms (if I'm reading it right, the article is very confused to say the least) which allow physicists to reconstruct the energies of hadronised partons. Now we can do it a bit better and make slightly better measurements. This software we know isnt optimal, it requires a great deal of knowledge to write and to be honest a major part of the effort in the earily days of an experiment is improving the reconstruction software with fixes such as this. And there will be many more such improvements. Bugs here do not pose any danger because the software is run *after* the event has occured so it cant effect the event, just our understand of what actually happened.

Also just to make clear that the LHC and CMS are very different things. The LHC is the accelerator and its what makes the particles go very fast. CMS is a detector, it just sits there and records what happens in the collision. CMS is built and designed by a completely different set of people to the LHC. CMS doesnt need the LHC to function and the LHC doesnt need CMS to function but they are a bit pointless without the other.

Re:Not a hardware bug

[Anonymous Coward]

Actually, not a bug at all. It's a design choice.

The CMS Global Calorimeter Trigger hardware uses a 3*3 sliding window algorithm to find local maxima (jets) in the calorimeter regions. These 3*3 windows can partially overlap, meaning some energy is double-counted. Having a small amount of double-counted energy has no real consequence on the validity of the triggering, but does greatly simplify the firmware.

Re:they should not turn it on

[hobbit]

Yep, the universe is conducting huge physics experiments, but very VERY far away from where we are, not in the middle of Europe.

The earth is constantly being bombarded with high-energy radiation. Some of that bombardment happens above Europe. The ionosphere is about 50km-1000km away from the earth's surface. What exactly do you mean by very VERY far away?

Re:they should not turn it on

[hobbit]

Maybe it doesn't turn into a big black hole, but it can explode and/or cause any other tragedy.

More likely is that any power plant you name could explode and cause a tragedy.

I know, let's shut down every single power station in existence!

At least that would have the side effect of downing the internet and we wouldn't have to listen to any more of your nonsensical tripe.

Re:Great story.

[Anonymous Coward]

That's right. Outside of the research community people might not know how much of the grunt work like this is done by students. She certainly has made a real contribution, but it isn't necessarily extraordinary. I'd bet that even had nobody seen it yet, as soon as collisions start it would be obvious that there were more muons than there should be and the problem would quickly be rectified.

Re:Not a hardware bug

[Anonymous Coward]

THIS. TFA is mistaken when it talks about hardware - she found a bug in the SOFTWARE that's used to interpret the decoder data. (Disclosure: I work on ATLAS, the competitor experiment to CMS). How important this bug is, I can't know offhand, but I'm pretty sure if it were significant it would have been spotted by now. These reconstruction algorithms (our software lets you choose from 4+ major different algorithms with a lot of customizable options) are being and have been grilled to optimize performance using simulated events. It is *somewhat* impressive that an undergrad found this, but this work is not as hard as people might think it is. In other words, business as usual. Profit: clap on the back and probably another 2-5k/year from the grad school of her choice.

Re:Great story.

[Gromius]

These undergradutes are members of the collaboration working under supervision of experienced physicists so they have full access to everything. Anyway consulting the technical design reports [cms.cern.ch] will give you some hardware info on CMS but its not really presented in a way accessable to a non physicist.

young physicists

[Goldsmith]

"If you haven't done anything in physics by the time you're 21, you never will."

I've been told this quote comes from Heisenberg, and at the time I heard it, I thought it was a load of crap. However, the idea is correct. If you want to be a physicist, you have to be able and willing to jump into research right at the beginning (as an undergraduate), or you'll probably never do real research. Of course, most undergraduates don't end up finding bugs in code which has been checked by dozens of postdocs and grad students.