The data here may be slightly different than the ones in other similar archives (e.g. GC3). The 81920 particles are in the following order:
16384 Gal. disk
16384 And. disk
8192 Gal. bulge
8192 And. bulge
16384 Gal. halo
16384 And. halo
These initial conditions start at a point less than 1 Gyr before
the collision since during the 3-4 Gyr or so leading up to the collision
nothing much happens. So Andromeda should look bigger than it is at the
moment. The simulation starts when Andromeda is only about 120 kpc
away from the Galaxy instead of the current 700 kpc so it should look
about 5 times bigger than it looks now.
In NEMO this table can be read into a snapshot as follows:
zcat dubinski.tab.gz |\
tabtos - m31.ini block1=m,pos,vel nbody=81920 times=0
If you want to run a quick simulation, you could take a subset of the
particles, multiply the masses appropriately and integrate, a subset of
640 particles, (320 per galaxy: 128 disk, 64 bulge, 128 halo):
zcat dubinski.tab.gz |\
tabtos - - block1=m,pos,vel nbody=81920 times=0 |\
snapcopy - - select=i%128 |\
snapscale - m31_small.ini mscale=128
hackcode1 m31_small.ini m31_small.dat tstop=80 freqout=1 > m31_small.log