You have to give data in $PARM and $HPARM below.
 InitRN =   17161577   121
 $PARAM
 DestEventNo = 500000, 1000,
 InitRN =   867592527   906414608
 IntModel = '"dpmjet3"',
 SeedFile = '%_Seed'
 Trace = 0,
 TraceDir = ' ',
 $END

 $HPARAM
 UserHookc = 'f 6  -1 1  10000  12  6  f f', 
           'Acrylic'
             ' '
 UserHooki = 8*1,
 TraceDev = 31 
 $END

===================================================
The first line in $HPARM is
(roots/lab) code subcode charge  E(GeV) (A Z) cms  xbyp

---------------------------------------------------

DestEventNo: If two fields are given for DestEventNo,  2nd one is used.
        This is the number of events to be generated.
InitRN  :   Two initial random seeds.

IntModel:  Specifies an interaction mdole. If '"dpmjet3"' is specified
        you have to be cautios about the 2,3rd lines in $HPARM.

       model name      K.E (GeV) region
      ----------------------------
       '"dpmjet3"'       >0.5 
       '"fritiof7.02"'   >  9 
      '"fritiof1.6"'   4 ~ 1000
        '"nucrin"'       < 4.5
       '"gheisha"'       < 1000 (always questionable)
       '"ad-hoc"'        > 4
       ('"nexus2"')      > 10 to be implemented.

Trace: If 0, numerical values of x, y, eta, pt,... etc are
       written on stdout.  If non zero, output is made on 'trace1' for
       event 1, 'trace2' for event2,... 
       The 'trace1'  etc can be an input to gnuplot by which
       you can see the track of incident and  generated particles
       as if it were observed in nuclear emulsion. 
       (However, this is only to visualize the interaction feature;
       even pi0 is not decayed here).  The basic gnuplot command
       to see the tracks is
	         set para
	         splot 'trace1' w l
      You can use Cosmos/Util/slide command to see the data
      consecutively.  The format of the trace data is the same
      as Cosmos Trace.

  **** If you make Trace != 0, make the DestEventNo small *****


TraceDri: Directory where files trace1, trace2,.. are generated.
	If ' ', /tmp/$USER is assumed.  Another typical example is
	a current directory, './'. See Trace above.

The first line of the UserHookc:
   roots:  If f, energy you give is assumed to be laboratry total   
           energy of the projectile. If t, it is root s of the
	   colliding particle (if projectile/target is nucleus,
           cms is formed by the p-p system.)
   code:   Particle code for the projectile. 
   subcode: Its subcode.
   charge:  Its charge.

	     code     subcode    charge  
	p      6        -1         1
	pbar   6         1        -1      
	n      6        -1         0 
	pi+    4        -1         1
	K+     5        -1         1
	K0     5        -1/1       0
	He etc 9         A         Z

   energy:  The energy of the projectile or the root s of the 
            colliding system.  
       A:   target mass number(integer)
       Z:   target charge    
    cms:    If f,  output quantities are in the laboratory system
            If t,  output quantities are in the CMS system
    xbyp;   If f,  x is defined as the ratio of kinetic energy of
            the produced particle to the kinetic energy of the 
            projectile. If t, it is defined  as the z-component
	    momentum ratio.

The second line of the UserHookc;
	This is used when IntModel ='"dpmjet3"'.  In this case, you can
	test a case for which a preinitialized Glanuber data is available.

	If ' ', the default Cosmos condiiton is applied. 
	The target can be only one of O, N, Ar. 
	The projectile max A is 56. You don't need worry about
        other input data.  The maximum energy /(nucleon) is 10^12 GeV.

	If 'xxx' is given, you have to prepare 'xxx.inp' data in this
	directory.  This entry is primarily intended to test preinitialized
        Glauber data in Epics/Data/Media/.  The prototype of the xxx.inp
	is test.inp.  You may copy it to xxx.inp (or any other name you want).
	Let's suppose to test Acrylic.GLB in Epics/Data/Media/.

	1) You may find Acrylic.inp in Epics/Data/Media directory and get
		EMULSION lines.  Replace EMUSLION  lines in your xxx.inp by
		these lines.
 	2) Find 'matter' in GLAUB-PAR line, and replace it by Acrylic.
	(i.e., Acrylic.GLB is assumed in Epics/Data/Media/ ).

	No other part should be changed. Then, you can start the test.

	If the preinitialized Glauber data (Acrylic.GLB in this case)
	is not in Epics/Data/Media, you can specify the directory 
	by the 3rd line (see next).

	The maximum energy (/n) is 4x10^6 GeV for Epcis/Data/Media/*.GLB.

The 3rd line of the UserHookc.
	If this is given, it is assumed to be a directory where you
	put the preinitialized Galuber data. 

	**** You need the last "/" for the directory ***

