cosmos comilation  f90jxflat -Hefsux
real   50 min

compilation:  f90jx  -Hefs
real    26m50.526s
user    5m46.632s
sys     12m1.965s


Fe 30 GeV. dpmjet3
  dpmjet3 init takes very long time.
2events generation with trace; same IR
          maja        LHCf2      LHCf
real    52m22.720s   1m10.890s
	52m38.765s   1m6.080s
        58m37.451s   1m9.771s
user    52m14.746s   0m53.594s 
        52m35.512s   0m50.886s
        58m29.234s   0m52.781s
sys     0m5.384s     0m17.044s  
        0m5.877s     0m14.911s
        0m2.654s     0m16.680s


   no trace  
real    51m33.481s   0m25.008s
        48m2.660s    0m25.275s
        54m28.028s
	45m41.618s
user    51m28.014s   0m24.795s
        47m59.175s   0m25.064s
        54m24.157s
	45m39.705s
sys     0m3.079s     0m0.147s
        0m3.123s     0m0.144s
        0m2.596s
	0m1.304s
Trace size: 
    230843776        261738948
    256545012        275208708
                     246879132
                     260674644
stdout size
    4745733          5240955
    4826147
EPOS: epos-xan-lhc-v3700.f
	 write( ),J...
   , should be omitted.

Import/Hidden/NUCFRG/Interface/csampNucFrag.f 
   IsNaN not useable.  comment out (4 places)
	(already it should not be needed)

cdefByMagAndE.f  
	"," was missing in each line;  correct one is
         write(0,*)
     *    "and prepare cmyEfield subroutine in your apps.",
     *    "Interface is cmyEfield(aTrack,Efout); see",
     *    "manual or cEfield0 in Cosmos/Module"

cfixModel.f
	stop 123456==> stop 23456  (must be 5 digit or less)


rnd.f
	call rnd1u(u,1), rnd2u(u,1)  is complained because
	     rnd1u assumes array for u
        so 
	real(8)::u1(1)
	call rnd1u(u1,1)
	u = u1 is used


HOST or HOSTNAME is not given.  so set it to something

cosmos/Zcmain.h
   comment out
	 call  mydummy
	 call myPhitsDummy


cGetXsec.f   changed for 0 divide

csampGintL.f    
          result of  cxrayp:   xs =0  is avoided

cinitObs.f
c////////////
  kcelei( ..   height undef)

ctransVectZx.f
      dir2 = temp ==> undefined temp.  
	       so temp.sys = dir1.sys is put before it.

cinteNuc.f
	debug output bef and after
	   call chAcol
  are removed.


phojet.f
	KACC.. = 0 is moved before pho_eveini

Util/Gencol.
	Gencol.mk
	  at top $(EPICSTOP) --> $(COSMOSTOP)
	Gencol.f
	  omit #include "ZepTrack.h"
	  activate #include "Zptcl.h"

Bertin:
	bert.f    0/0 happens when out(12)=0.
	(at around line #1000) 
	if out(12)== 0, next is forced.
         ppmda = 1.0
         ppnda = 1.0
         ppnna = 1.0

Import/DPM.
	dpmjet3.f  and pythia6115.f
	flaating erro 
	  In DT_INUCAS of dpmjet3
        0 division error happens 
        In PYZDIS  aroung line 35709~ 
        avoided by compiling with -g.
aftet this, IntModel "dpmjet3" works fine.



Import/Sofia
	jetset74dp.f
	in LUKFDI  0 divide. but position is not well
	know.
  ==>	-g option compile.
		
After this, '"phits" 2 "dpmjet3"' works fine

 Import/Jam
	same as pythia in DPM.
	jampyjet.f has same problem.
        0 divided in  pjzdis.

 Improt/QGS qgsjetII-04
   A floating division exception was detected.
 error occurs at qgprox_  line 11696. but this is fake.
 could not pin-point the location.
  So the qgprox part is separated as qgprox.f
  and compiled with -g option. (-->cppFCJaxa)


After above modification, for 1000TeV Fe withmin 50TeVmin,
run,error happend.
 An invalid floating point operation exception was detected.
 error occurs at qghot_   line 12061

So aghot is moved to qgprox.



 Import/EPOS
  epos-int-lhc-v3700.f
  exp overflow for single pre. (exp(x) x > 88.77 )
 !!!    yrmaxi=yrmaxi*log(exp(yradpi/yradmx)+amctot-1.)     
--->    yrmaxi=yrmaxi*log(exp(dble(yradpi/yradmx))+amctot-1.)     

there are  other exp. and the same was applied.
 
above correction OK


jwe0013i-e A floating division exception was detected.
 error occurs at psahot_  line 824 loc 000000010089c248 offset 0000000000003d08 
 psahot_      at loc

-g option compile

