90 likes | 221 Views
Computation Time Analysis - Climate Reanalysis Data. Dipanwita Dasgupta Graduate Operating Systems. Overview. Problem Description Dataset Results Proposed course of work. Problem. Cannot increase the number of parameters Code in R script Code using direct computational techniques
E N D
Computation Time Analysis - Climate ReanalysisData DipanwitaDasgupta Graduate Operating Systems
Overview • Problem Description • Dataset • Results • Proposed course of work
Problem • Cannot increase the number of parameters • Code in R script • Code using direct computational techniques • Data to be saved in the machine • Data has to be loaded in chunks in terms of longitude to get a better computation time • Very high computation time • No automated technique for computation
Data • NCEP Reanalysis Data • Composed of data at 17 pressure levels • Total of approximately 10000 grid points • Available at ftp://ftp.cdc.noaa.gov/Datasets/ncep.reanalysis.derived/pressure/hgt.mon.mean.nc
Results • Time increases with number oflevels • Represents time per block of data • Only one parameter used • Only one measure of dispersion
Course of Work • Transfer the code to Matlab • Use distributed computing to reduce computation time • Compute for each level separately • Combine the results from each level • Increase the number of parameters