Optimization of Preliminary Data Analysis for PINGU

The IceCube Neutrino Observatory (INO), a 1 km$^3$-sized detector at the South Pole, collects Cherenkov light from neutrino interactions in the ice. The light is detected by Photomultiplier tubes (PMT), which are contained inside the Digital Optical Modules (DOM) along with an FPGA board with an ARM CPU meant to process and analyze the incoming signal. The analysis includes the Wavedeform routine, which uses a Non-Negative Least Squares algorithm to unfold the signal into a series of separate pulses. The Precision IceCube Next Generation Upgrade (PINGU) is a proposed low energy infill extension to the INO, which will require the deployment of many additional DOMs. The new DOMs can be optimized by running a routine like Wavedefom on the FPGA fabric or CPU. Because the output of Wavedeform is much more compact than the original waveforms, the amount of data transmitted from the DOM would be greatly reduced. As it is, Wavedeform is a complicated procedure that depends on several libraries and requires a lot of processor power to run. It needs to be optimized or substituted by a faster algorithm. The performance of other algorithms for identification of single photoelectron pulses (SPE) is evaluated and the received parameters of SPE pulses are compared to those by Wavedeform.