
I wrote my own fitter from scratch. It was the right thing to do, and I loved it. My colleagues just used an old one. You would have thought writing your own code would have given me street cred. Oh no. They disrespected me all the way, even as I blew through my analysis.
My analysis was unusual. It involved a 4-D rather than a 2-D fit like everyone else did. It took forever. As I neared my defense date we had to reschedule because my code took anywhere from a week to almost a month to run. And who got blamed? Me. Why not? It's what my Mom always did. Anyways, I tried to explain the slow down. Going from 2-D integration to 4-D integration was a big hit in terms of computability time. They didn't listen. My analysis only had 180 events. They fit 10,000 events and it didn't take as long. I asked the dimension of the fit. It was 1-D. I again explained the reason for the slow down was the 4-D integration. Another person responded with, "I fit 40,000 points and it didn't take this long." It of course was also 1-D. I even pulled out the documentation to show how the time to compute scaled with dimension. I am pretty sure the whole time they were thinking, "But I fit 10,000 points in less time."
I'm not done. In order to do my analysis I needed a lot of processors to do multiple jobs at a time. My experiment had such a system. I asked to do it and they told me there was a 2 hour limit on jobs. What the bloody hell was everyone else doing at this experiment?
(Title: Time Is on My Side, The Rolling Stones)

No comments:
Post a Comment