Dear Igor. I has the same problem and few days a go Prof. Hulpke write to
me the
following message:

 "You have reached the limits of the capabilities of GAP 3 as documented
in       
the manual.                                                                     
The subgroup lattice program in GAP 3 is only guaranteed to work if the         
groups solvable residuum is not larger than 5000. (The algorithm uses a
data    
base of recomputed information and this information is not available for        
larger groups). Your group <g> is perfect an thus equal to its solvable         
residuum, so there is no easy way to compute the subgroup lattice.              
The only way around this problem would be to provide the pwerfect
subgroups     
a priori from theory.                                                           
The manual secion on `Lattice' gives an example on how to do this.              
                                                                                
The subgroup lattice function in GAP 4 has been improved and will work for      
much larger groups (and -- provided memory being available -- should work       
for all subgroups of S12).                                                      
However the code is quite different and thus it is not possible to make
this    
available for GAP 3."

I hope this help you.

Mario Pineda.