Tutorial 2: Using your recently acquired skills

Introduction

This tutorial is more of the same as tutorial 1. The difference is that you will get to use slightly larger configurations and more of them.

Configurations

Followig this link will take you to a directory containing

5 Gauge Configurations, named dumm_run_cfg_XXX.lime. Here XXX stants for the HMC sweep number and can take the values 220, 240,260, 280 and 300. The configs were produced with Wilson Gauge action and 2 flavours of dynamical Wilson fermions. The physics params were:
- beta = 5.4
- kappa= 0.11
The configurations have volume=8x8x8x16, and are saved in the SZINQIO format.
One file for chroma to compute spectroscopy on config 220
An XMGR parameter file for plotting (see later)

Your task, should you choose to accept it, is to compute effective masses from the spectroscopy carried out on these files

Doing the first of 5

Make a directory called tut2. Enter into it with

cd tut2

Copy the configuration dummy_run_cfg_220.lime and the example parameter file example_220.ini.xml into your working directory.

Take a copy of chroma into the working directory. You can reuse the one from the first tutorial or just download it again from here (if you have Fedora Core 3 or compatible linux). Alternatively you can get it from the installation directory /usr/local/chroma/scalar/bin/chroma if you are attending this tutorial in person.

Run chroma with the command

chroma -i ./example_220.ini.xml -o ./output_220.xml

Chroma should run. You should find an output file spectrum.out.xml for the spectroscopy. I have turned off hybrid mesons and baryons for this test to make life faster. Also the propagator is no longer being saved.

Do the rest

At this point, you can carry on and do the other 4 configs. You can get the configs from this link. You should also get the input XML files for config 220. In order to finish this exercise quickly I have the following hints:

Using Sed

You can perform search and replace on a given XML (or indeed any text) file using sed. You can use this for example to change the config number from 220 to any other config number from an input XML file to an output one:

sed -e "s/220/240/g" example_220.ini.xml > example_240.ini.xml

will do a search and replace on example_220.ini.xml and create a file example_240.ini.xml suitable for use with config 240. Have a care tho. This will replace every occurrance of the string "220" with "240". So if you were in a situation where you'd have say kappa=0.1220 it would also be changed to kappa=0.1240. However in this exercise you should be in no danger of that. If you want more selective XML processing, find out about XSLT transformations.

Stripping

Take a copy of the strip_spectrum utility you used in the last exercise. Make a directory called strip and strip the files into it as before. A good way is to go into the directory and do the stripping there ie:

cd strip cp ../strip_spectrum . ./strip_spectrum ../spectrum.out.220.xml ../spectrum.out.240.xml ../spectrum.out.260.out.xml ../spectrum.out.280.xml ../spectrum.out.300.xml

Effective Mass Utility

The adat analysis package also contains a utility to look at quick and dirty effective masses. Recall that you built strip_spectrum. Let us now build the effective mass utility.

Go into the directory where you have checked out adat. Instead of going into the main/strippers directory as before, you should now go into the main/ensem directory, and type

make meff

You should build a utility called meff. As usual if you have FedoraCore or a compatible system you can circumvent the building of it by simply getting it from here.

Now copy this to the directory where you stripped the spectrum results (the one we called tut2/strip). cd back to the tut2/strip directory (where you have just copied meff to) and type

./meff ./meff pion.D5455.P_1.P_1.PP

You should see the output:

0 3.39003 0.0064513 1 2.72824 0.0150325 2 2.65792 0.0144408 3 2.62275 0.0307213 4 2.65234 0.0374466 5 2.62383 0.0316489 6 2.563 0.0329779 7 1.98347 0.0595311 8 -1.91219 0.076478 9 -2.63864 0.0261858 10 -2.6253 0.0235725 11 -2.62692 0.0333184 12 -2.65712 0.0177053 13 -2.65389 0.0277146 14 -2.69619 0.0105981 15 -3.41136 0.012748

The first column is the timeslice number, the second is the effective mass, and the third is the jackknife error. This correlator is symmetric and has not been folded in the middle. Hence at the midpoint the effective mass turns and becomes negarive. I did say it was rough and ready.

Plotting it

You can of course save the above information by redirecting the output of meff to some file eg:

./meff ./meff pion.D5455.P_1.P_1.PP > eff_mass.dat

which you can then plot with your favourite plotting program. I like xmgrace. This XMGR Parameter File should make the plot more pretty. You can use it as follows:

Download it into a local file called xmgr_params in your working directory
Run XMGR with the command:
xmgr -param xmgr_params pion_fold_eff_mass.dat

Finally: Fitting Correlators

I choose conciously not to discuss this aspect of the work here. Everyone has their own favourite fitter to which they stick to through thick and thin. Consult your local fitting expert for what is best for you (them).

And you're done

Congratulations. You should now know how to do spectroscopy with Chroma.