Nuclear Shell Model Codes

Home of NuShellX & NuShell & MuShellX
Home
About Us
NuShellX
NuShellX for Linux and Unix
Terms of use
Contact Us
Site Map
News
Anniversary
MuShellX
NuShellX & NuShell
In 1974 I started to write a nuclear shell model code. Its first real useable form was called the "Oxford Shell Model Code". It was a code for calculating nuclear energy levels. It was written in collaboratiion with graduate students at Oxford, Colin Zimmerman and Nigel Godwin. As it developed Alex Brown added functionality with other students at Oxford and it grew into OXBASH. Many other authors added more enhancements. I have now started again from scratch to write NuShell. NuShell is a completely new code written in Fortran 95 using OpenMP multiprocessing. This time I have written the whole package myself, but have had help debugging it from Alex Brown and Dan Judson. MultiShell was an extension of the Oxford Shell Model code. I have also written an eXtension to NuShell, called NuShellX inspired MultiShell and by emails exchanged with Jussi Toivanen, from Finland. Bill Rae, Garsington, Oxford 2008. ( shellmodel@garsington.eclipse.co.uk ) See the new page on MuShellx.
What's new?
 
 My web company deleted all my files. Please go to the NEW site at knollhouse.eufor current files. 
 
See the new News page for latest tips, and bugs
 
The new NuShell code is easier to maintain, faster and hopefully more portable. It is in many ways similar to OXBASH but it has been improved. The angular momentum projection routine has been completely redeveloped  It now handles high j orbitals such as h11/2, i13/2 up to j=17/2 and many particle cases ( >10) where OXBASH failed in some cases and it is
much faster for many calculations -sometimes a 10 fold increase in speed is obtained.
 
On multiprocessor systems eg Intel Duo or Quad processors performace is further enhanced. But it does require at least 1GB memory to run and 2GB or more is preferred. The sizes of arrays can be set via environment variables on all platforms so the size of calculation is only limited by memory available.
 
The closest competitors to NuShellX are Nathan from the Strasboug group and Jussi Toivanen's EICODE neither of which are in the public domain. 
 
What do people think?
Alex Brown has used the NuShell code to create his NuShell @ MSU version of NuShell with an OXBASH look and feel and standard OXBASH input output. Alison Bruce has said "it is much faster and its good to have a new code available" and in collaboration with the Brighton group a version for Solaris on x86 or SPARC platforms has been developed. Contact Alex Brown.( brown@nscl.msu.edu ). His version of NuShell will effectively replace the original OXBASH.
 
NuShellX has already been acclaimed as a "nice Xmas present"   by Alex Brown  and I think it has a promising future. Alex will produce a version NuShellX @ MSU.
 
Mihai Horio has reported about the latest version (4.2) of NuShellX - "This one is working well. It took 3h and 57 min for five 0+ states in 56Ni. That's probably a world record."
 NuShellX pages were viewed over 1400 times during the period of Jan 2008 - Mid Feb 2008. There have been a constant 1000 pages read per month since then.
 
 First publications using NushellX are now starting to appear in print. Versions have been running efficiently on 100 cores. This proves it is very scalable.
 
Its possibly one of  the best ever codes.