LIGO Document T1700206-v1

CWB Optimization for GW Signal Waveform Reconstruction Stage

Document #:
LIGO-T1700206-v1
Document type:
T - Technical notes
Other Versions:
Abstract:
2017 LIGO SURF project by Yangi Gu.

CWB identifies transient gravitational wave (GW) events, from a broad range of generic transient signals and determines which events could be considered significant above a certain level of standard deviation. Events are tagged with likelihood a value that ranks their probability of being a gravitational-wave signal. The significance of an event with given likelihood, is determined by a background rate—the rate at which detector noise produces events with a likelihood value equal to or higher than the candidate event. The bulk of compute time in the CWB pipeline is consumed by estimating these background rates, where time-frequency analysis and time-shift analysis over 10s of thousands of sky locations are performed.

The aim of this SURF project is to understand and address cause(s) for a section of CWB known to be computationally time consuming, a probable candidate for porting to execution on a GPU or Many Integrated Core (MIC) architecture.

Files in Document:
Other Files:
Keywords:
SURF17

DCC Version 3.4.2, contact Document Database Administrators