Thanks for pointing out the issue with the passing of the sigma
variable (which strange enough did not appear on our machines).
For future reference and completeness, could you please tell us which
compilers are you using? Thanks!
On 8/27/07, gmar...@uci.edu <gmar...@uci.edu> wrote:
>
> We had to use the -cxxlib flag on the ifort compiler in order for it
> to link the c++ libraries (added to the FFLAGS line). I suppose you
> could also use the -lstdc++ flag with a -L$(PATH_TO_CPP_LIBS) added to
> the 'LIBS = ...' line. We also had to add the path to the cxa, cprts,
> irc, unwind libraries (we just added a -L$(PATH_TO_ALL_THOSE_LIBS) to
> the 'LIBS =' line). To ensure that feynhiggs compiled in 32 bit and
> not 64 bit, we added the -m32 flag to the CFLAGS in the feynhiggs'
> makefile.
>
> We also got a run-time segmentation fault. To solve this, in
> calclike.f90, we changed line 258 to
>
> function SmearedBound(y,x0, sigmain, tau,lower)result(lnprob)
>
> and line 277 to
>
> real(8) :: lnprob, x0, sigmain, tau, tstar, exparg, Zfunc2
>
> and right after added the lines
>
> real(8) sigma
> sigma = sigmain
>
>
> >
>
We also got a run-time segmentation fault. To solve this, in
calclike.f90, in function 'GetLogLikeBsmumu' (around line 852) we
replaced
lnlike = SmearedBound(1d0*Bsmumu,1d0*xstar, 0d0, 1.0d0*tau,
lower= .false.)
with
lnlike = SmearedBound(1d0*Bsmumu,1d0*xstar, 1d0*sigma, 1.0d0*tau,
lower= .false.)
and added the line
sigma = 0.0
after the line 'xstar = 1.0'