GENIE:GENIE Versions

GENIE Versions
This page records the details and meanings of tags in the GENIE code repository.

The GENIE code base is tagged at various junctures with the following naming convention:

rel---

HEAD
Note that you can think of the repository's HEAD version as a kind of model beta version. Any code commited to the repository should have been tested against the suite of short tests. These tests are designed to run easily and quickly--to encourage people to use them. They also aim to exercise as much of the model code in the short time that they take to run. However, you will appreciate that there is an inherent tradeof. To supplement the short tests, a longer suite of tests is run as part of the nightly build.

Released versions are scrutinised a good deal more than the automated tests can do. For example, the model is typically subjected to a large number of tuning runs, exercising the model across large portions of the space of tunable parameters. In addition to the elevated quality control, you will also have a set of optimised parameters to use in conjunction with a stable version of the model. (The is no guarantee that such a parameter set will be optimal for another version of the model's source code.)

So, in summary, it is far better to use a released version of the model than the HEAD.

rel-2-4-2

 * Stable release for UoB MSc course.

rel-2-4-1

 * Bug fix after development on trunk broke model.

rel-2-4-0

 * ENTS has been geniefied!

cup-1-0-0

 * Version 1.0 of cupcake. Merges to/from trunk to follow.

rel-2-3-0

 * Tag immediatly after switching to XML.

rel-2-2-1

 * Last tag before switching to XML format config files.

rel-2-2-0

 * Tag created for UEA workshop.

rel-2-1-0

 * GOLDSTEIN equation of state corrected
 * Corrections to embm, goldstein and seaice modules to align eb_go_gs with genie-cgoldstein

rel-2-0-0

 * Formerly tagged XmasGC
 * Version used for the "Grand Challenge" experiments conducted between Dec. 2005 - Mar. 2006

GENIE Branches
Un-merged branches in the CVS repository for specialised study

ipcc-ar4-2005-rel-1-2

 * Corrected equation of state
 * Version used for final submission (Oct. 2005)

ipcc-ar4-2005-rel-1-1

 * Minor correction to inm.f
 * Version used for initial submission (Mar. 2005)

ipcc-ar4-2005-rel-1-0

 * Initial tag intended for production runs in the ipcc-ar4-2005 branch

Model Comparisons
On the 14/12/06 Gethin had the following final values for average SAT (reported to STDOUT):


 * {| border="1" cellpadding="3"

!Model !command !value
 * cgoldstein
 * "goldstein < goin.test")
 * 10.5088016763012
 * genie_eb_go_gs
 * "genie_example.job -f configs/genie_eb_go_gs.config"
 * 10.5088016337715
 * }
 * 10.5088016337715
 * }
 * }

On the 12/02/07 Gethin had the following final values for average SAT, after doubling the run length:


 * {| border="1" cellpadding="3"

!Model !command !value
 * cgoldstein
 * "goldstein < goin.test")
 * 11.3269478462944
 * genie_eb_go_gs
 * "genie_example.job -f configs/genie_eb_go_gs.config"
 * 11.3269577445993
 * }
 * 11.3269577445993
 * }
 * }