Bonjour Tout le Monde!
It is getting hazy, hot, and humid in Geneva. You can't see Mont Blanc at all. She's been gone for three days, but today she is non-existent. It will hit 100 tomorrow and we are officially at the peak of the heat wave. 
The temperature gets hotter but the physics gets cooler! Today we walked from Switzerland to France in the CERN complex and went to learn about computing at CERN from Marcus Joos.
Marcus has been with CERN since before the Higgs was found and has been writing program since he was in high school. Now Marcus' big takeaway for us was to realize that when he looks at CV's from students all over the world, he isn't looking for all A's because he can't compare one school system to another or one school district to another, he looks for passion. That was the big take away. Why do you use computing? How do you use it when there isn't a teacher making you? Some interesting things to think about are that CERN, despite needing Petabytes to calculate its data from detectors uses the least amount out of Facebook, CERN and Amazon. CERN uses 450,000 computer cores, independently running processing units. Each experiment runs 50,000 each (ATLAS, CMS, LHCb and ALICE). There are about 1,000,000 other units used world wide in the computing grid.
Marcus has been with CERN since before the Higgs was found and has been writing program since he was in high school. Now Marcus' big takeaway for us was to realize that when he looks at CV's from students all over the world, he isn't looking for all A's because he can't compare one school system to another or one school district to another, he looks for passion. That was the big take away. Why do you use computing? How do you use it when there isn't a teacher making you? Some interesting things to think about are that CERN, despite needing Petabytes to calculate its data from detectors uses the least amount out of Facebook, CERN and Amazon. CERN uses 450,000 computer cores, independently running processing units. Each experiment runs 50,000 each (ATLAS, CMS, LHCb and ALICE). There are about 1,000,000 other units used world wide in the computing grid.
Some things we talked about were the operating system, Linux, the primary coding language, C/C++, and the need to remain free from the thumb of Microsoft, et al. who try to raise prices and really hold you under threat. To the question of "should I learn to code in school?," he responds, "Do you want to be the driver in the car or the passenger in the bus?" How much freedom you have depends on the amount of skill you possess.
We were then able to visit the DATA CENTER on campus which is basically in charge of the day to day things as well as the web connection. Most of the number crunching from experiments from after trigger 2 happens on the other campus in France. Trigger 1 happens on premises of each experiment and drops the 100,000,000 collisions/second to 100 high interest collisions per second. Then the secondary trigger drops those down to about 10 collisions/ second that are sent off to the world to be processed. Another interesting point to note, this room was once the ONLY direct connection for the internet to reach from Europe to Chicago, IL where Fermilab is located. It also manages the connections for the EU and the UN.
How long would it have take in each of these years to process the information obtained by 1 year of operation at CERN?
back left is in charge of the EU and UN and other stuff
Free look at what is happening. Everything at CERN is open source
We took some time as well to inspect mystery boxes and think about how to implement them in our classes. I will be making some interesting boxes for day 1! Get ready, students. 
My partner was my colleague from Italy
Lastly, we heard a lecture from Dr. Flavia de Almeida Dias, who talked to us about data analysis. She worked on CMS before the Higgs and now works for ATLAS. She is interested in finding new physics in the data from the LHC experiments, so she has to investigate the data that makes in through the triggers to look for interesting data that pops up above the background data at a level that would be unthinkable otherwise. For example, if the peak at 125 GeV in the 2012 paper was not the Higgs and was instead random background, that would be a 0.000035% chance. Statistically improbable, and having now found several Higgs candidates, we can say pretty positively that it is likely the real deal.
We ended the night with some UNO....I did not win. Pictured colleagues from (front: Portugal, rt to left: Italy, Italy again, Canada, Hungary, Serbia, Estonia, me, US, Lebanon, Australia)
No comments:
Post a Comment