The configuration files

• mtpermit.txt: Is located in the main area (ie /opt/mtdata/bin..), it is necessary to access the databases. If it ever happens that this has to be modified, contact the NPL.

• mtconfig.txt: This file need to be present in the directory in which you are when you launch MTDATA. It tells it where to find the databases. You can edit it if you want to add your own databases. If you do so, you will see that the databases you most often use are located with

  plus=mt-sgsplus plus is the name that you will be able to use in MTDATA to set this database as a source.
  mt refers to the main area again, and therefore MTDATA looks for this database in /big/hugh/... You can also give a path: mydata=/huge/users/thomas/mtdata/ownbase which will make the data in ownbase available under the name mydata.

• mtsignon.txt

In the same way as for mtconfig, this file is first looked for in the directory in which you are when you launch MTDATA, if there's no such file, the default one located in the main area is used; this is currently the case.
This file is used to set the configuration variables. An important variable is HCPY which defines the format for graphical output. The default is POSTSCRIPT, but if you use LaTeX, you will be happy to know that you can set it to give you an EPS file during a session by typing: [HCPY=EPS

The Configuration Variables

Note that you have to press twice Return to get back to the prompt.

One you may want to know about is STAGE_1. You can set it to 'new' ( [STAGE_1=new ) if you want to use the most recent version of this algorithm.
It is a bit more accurate (however, the precision concerned is far above the one we are concerned with, which is not the case of people working with gaseous systems), but also more robust. I actually found a case in which a calculation producing really weird results with the 'old' algorithm gave more sensible ones with the new one.
If ever you get interested in using the other configuration variables, you will find their names and possible values in the Handbook.

Info about MTDATA on obsolete SUNs, see Linux

There was a time when MTDATA would only run on our SUN machines, these cost about 3-4 times the price of a PC having superior performances. Needless to say the port of MTDATA to Linux was very welcome !

If for compatibility reasons you were to want to use MTDATA on one of our surviving SUNs, you'd need to know that:

The user

MTDATA is a user in our group, with its home directory, its .cshrc file, etc.

The latest version of MTDATA's (4.71) executable files are in /big/hugh/mt471/bin, together with the databases. You may find others copies (databases or program) around, but they are just there to take some hard drive space.

The home directory of MTDATA is /home4/users/guest, this is where the different files generated during a session will be stored, for good as you can guess when observing how full is this directory.

Things to do before using MTDATA on a SUN:

To avoid saturation of MTDATA's home space, each user should create a subdirectory with its name in /home4/users/guest.
The first time, as MT_DATA user, do:

• mkdir thomas
• cd thomas
• start_script

• cd thomas
• bash (optional, but the prompt will be nicer and you will have tab complettion etc..)
• mtdata

The start_script will create a link to mtconfig.txt, which tells MTDATA where to find the databases.

The advantage is that modifications only have to be made on the real file. If you wish to have your own settings, just copy the file and modify the copy.

Although it is not impossible, I have made it difficult to run MTDATA if you are in /home4/users/guest (you will be told to create your own directory).

So what ?

All the files generously spread by MTDATA will now fill your own directory, and annoy you and only you !

• .gph, .gky generated when you do stepped calculations
• .log is a complete log of your session,
• .out appears when plot a graph,

You can get rid of these in most cases. For more clarity, you can create separate subdirectories for your mpi files, results etc..

New Databases

The new databases should be available on all the machines except ptlin5 and ptlin1. (These use the older permit file because of peculiarities in the different versions of linux).