Software Benchmarking - PowerPoint PPT Presentation

1 / 6
About This Presentation
Title:

Software Benchmarking

Description:

Analysis WG. Apr 29-30, 2002. Software Benchmarking. Van Husson, HTSI. Erricos Pavlis, JCET ... Coordinate solutions are converging and some sources ... – PowerPoint PPT presentation

Number of Views:63
Avg rating:3.0/5.0
Slides: 7
Provided by: Van6
Category:

less

Transcript and Presenter's Notes

Title: Software Benchmarking


1
Software Benchmarking
  • Van Husson, HTSI
  • Erricos Pavlis, JCET
  • Maria Mareyen, BKG

2
Introduction
  • Coordinate solutions are converging and some
    sources
  • of variations have been discovered and
    eliminated
  • To expedite this process and to ensure our
    solutions
  • are of the highest quality, we need to
    benchmark our
  • orbit determination and solve software
  • Passing the benchmark will be a pre-requisite
  • requirement for ALL Analysis Centers

3
Benchmark Controls
  • Minimize sources of variations
  • Same eccentricities
  • Same input dataset (991010 LAGEOS-1)
  • 28-day period that we have used previously
  • High performance systems only
  • Bad normal points edited
  • ftp//cddisa.gsfc.nasa.gov/pub/slr/ilrsac/data/ben
    chmark/qldata.991010-991106.lageos1_sp.Z
  • Apriori positions and velocities (ITRF2000)
  • Same force models - TBD
  • Same data treatment - TBD
  • Same constraints - TBD

4
Benchmark Output Products
  • EOP and Station positions in Sinex
  • Text files of normal point observed minus
    calculated
  • (free format?)
  • Text files of orbital parameters (free format?)
  • - true of date satellite x, y, z xdot, ydot,
    zdot
  • - at a prescribed interval (2 minutes)
  • Corrections atmosphere, CoM, etc.
  • Orbits
  • Nominal Model OMCs Orb. Parameters
    corrections
  • Fixed EOP and Coordinates OMCs Orb.
    Parameters
  • corrections
  • Final Orbit EOP POS OMCs Orb. Parameters
  • corrections

5
Benchmark Metrics
  • Inputs
  • Eccs, force models, constraints, data treatment,
    aprioi positions
  • Outputs
  • RMS of arc
  • EOP Site position solution estimates (e.g.
    baseline lengths, heights, etc.)
  • Orbital parameters
  • Site residuals
  • Anything else?

6
Next Steps
  • Document models to be used by May 15, 2002
  • Invitation by May 17, 2002
  • 1st Iteration of Solutions by July 1, 2002
  • Need volunteers to perform the comparisons
  • Provide initial feedback on the results by Aug 1,
    2002
  • Report on the results at the next AWG meeting
    (October 3-4 2002, Washington DC)
  • Expand the dataset (e.g. expand the time series,
    add LAGEOS-2, )
Write a Comment
User Comments (0)
About PowerShow.com