L2jet Investigation: Ross's Code
Here I note some discrepencies between my 5x5 JP jet code using L2VirtualAlgo2006 and my 5x5 JP jet code using L2VirtualAlgo2009.
The diagnostic plots from VirtAlgo2006 are found here and the plots from VirtAlgo2009 are here.
In particular the bottom left plot on page 5 seems troubling.
Solution
I believe that the problem was with the parts of the code which calculated the energy weighted eta and phi positions of the jet patches. To garud against dividing by zero while doing the averaging, certain variables were initiated to a small value, 1.0. This value was small compared to the raw adc values used in VirtAlgo2006 but were sizable compared to the calabrated energy gotten from the event stream in VirtAlgo2009. These initial values were skewing the averages. To rectify this, I have changed it so that these variables are initialized to 0.001 which should be small enough not to significantly impact the weighted positions. Details of the changes follow on my next blog page.
- pagebs's blog
- Login or register to post comments