Follow this link for the more recent analysis on the topic.

CORSIKA INPUTS file:  INPUTS (generated by handle.c)

CORSIKA was run on 16 PII-300 and 7 PII-400x2 computers in the background for approximately 2 months. On each processor a separate thread was run with a different random number generator (RNG) sequence. To minimize possible loss of data from interrupted runs, in each thread a data file was written to disk every 1000000 events, together with an additional file which contained RNG SEED information, necessary to continue the calculation where it was left off. A total of 2339e+6 primaries were analized, and the output was saved in the F2000 format (with a couple of modifications, listed below). The output consists only of the showers that have secondaries.

The .car file with all our modifications is available here:  corsika562.car
The program, that managed the calculation is  handle.c
The program, that we used to unify all data files into one gen data arhive is  un.c
A couple of programs, that we used to read the data file and build hisograms:  rr.css.c  (together with  user.c )

The gen data file is  datafile.f2k.gz . Beware: this file is 940549692 bytes long, and will take a lot more time to download, than the files listed above.

THE NEW FILE is  newfile.f2k.gz . It is generated by the program  ucr.c  from the above file. This fixes:
1.  z coordinate of the level of observation
2.  micoseconds changed to nanoseconds in the times of arrival at the observation level
3.  track numbers are resequenced so that they are unique
4.  Particles that do not reach the observation level (primaries) are removed
5.  USER block is removed from each event, even though it is still defined
6.  x and y of all particles are properly randomized according to the cylindrical geometry of the problem
7.  perhaps something else is changed and not listed here.
Total length of the file is 629626084 bytes.
 

Modifications to the F2000 file format:
In order to save the NKG information about the EM component of the showers, we defined a user block:
USER_DEF EVT_STAT EL_NUM AGE_PAR WN_HADR WN_MUON WNUM_TOT

    EL_NUM is the average number of electrons on the observation level
    AGE_PAR is an age parameter, used in the NKG formular
    WN_HADR is a weighted number of hadrons in the shower
    WN_MUON is a weighted number of muons in the shower
    WN_TOT is a total number of all (non-EM) particles in the shower

The parent ID was a question mark for all particles, except muons. It was 1 for muons, resulting from pion decay, and 2 for muons, resulting from kaon decay.

In our preliminary analysis we have obtained a value of -2.7731+-0.0008 for the muon spectral index (with Chi^2/DF = 1.096), using only protons as primaries, using procedure, as described in our paper (ICRC 26 HE 3.1.07)