As of v. 1.2.1 and 1.2.2:

Minor changes which should not affect use within the AMANDA software chain.
Option -maxbuf is no longer necessary. ABB formula was corrected for taus.


As of v. 1.2:

Some changes in the mediadef file definition. Air medium is added. It is now
possible to vary density of the propagation media.

A bug in the e+- pair production cross section was corrected resulting in a 5%
higher cross section. Option -sh=2 enables newer nuclear structure function.
For AMANDA MC cross sections are selected with -lpm -bs=1 -ph=3 -bb=2 -sh=2.

To propagate muons originating in ice, underlying rock, or air just above the
ice surface, run mmc with the following options:

ammc -r -a -romb=5 -raw -user -sdec -time -cont -lpm -bs=1 -ph=3 -bb=2 -sh=2
-scat -amasim -rdmc -mediadef=mediadef

where mediadef file contains the following:

	# media definition file
	# include with -mediadef=[this file]
	# use the following syntax:
	# all        det vcut ecut conti rho medium
	# sphere z r det vcut ecut conti rho medium
	all                  0 0.05 5.e2 1 0.673 Air
	sphere  1730 6400000 0 0.05 5.e2 1 0.825 Ice
	sphere  1530 6399800 1 0.05 5.e2 1 1.0   Ice
	sphere -1080 6397190 0 0.05 5.e2 1 1.0   Standard Rock

Of course, if this setup is used, -DCORR=35 option should not be used with
ucr. MMC run in this way will generate and use more tables and takes more
memory at the run time.


As of v. 1.1:

Default compilation with gcj now proceeds with the "-O3" optimization option,
which appears to work on both Linux and OSF1 platforms. It is now possible to
pass different optimization options to gcj by adding these options to ammc.


As of v. 1.10:

MMC code was restructured and placed into 4 packages. This should not affect
the behavior of the static (gcj) executables on this page in any way. However,
several new features (cross sections and effects) were added to this version,
which required removing some old flags (-allm, -phnu) and adding some new ones
(-ph, -bb, -sh). To use exactly the same cross sections as in the previous
version in mass/simu packages, replace "-allm" with "-ph=3 -bb=2 -sh=1". This
selects ALLM parametrization (-ph=3), year 97 version (-bb=2) with the same
shadowing function as before (-sh=1). Parametrization table names have changed
too, so the new ones will be generated when you run MMC, and the old ones can
be removed.

Alpha executables could not be produced with -O2 or -O3 flags. Gcj aborts the
compilation with the "virtual memory exhausted: Not enough space" error. The
ones produced with -Os appear to be about 7% faster than those produced with
-O1 and about 55% faster than those compiled with -O0. To change this option
search ammc for "-O3" and replace it with your option, then recompile with
"ammc -gcj".



As of v. 1.09:
Why two different executables for Linux?

It was found that executables compiled on RedHat 7.2 fail on RedHat 9
computers. Executables mmc_v1.09_gcc3xx_rh9.Linux and have been compiled on a
RedHat 9 laptop and appear to run on the RedHat 7.2 machines, too. Executable
mmc_v1.09_gcc332_rh9.Linux is compiled using the newer version of gcc and is
~20-30% faster, but for some reason continuously grows during runtime, so use
it if you are processing relatively small files and do not experience problems
with available memory. Executable mmc_v1.09_gcc322_rh9.Linux is compiled using
the older gcc which comes with RedHat 9 and is not as fast, but doesn't have
the expanding used memory problem (should stop growing at ~25 Mb for standard
AMANDA processing).