I am attempting to create an PDL interface to a Fortran function that
solves quadratic programming problems. I've made a fair bit of
progress getting up the PDL::PP, Fortran and XS learning curve, but I
am currently stumped by follow error message:
Error in qpgen2:Wrong dims
at .../PDL-QP/blib/lib/PDL/Opt/QP.pm line 148.
PDL::Opt::QP::qp('PDL=SCALAR(0x10082e280)',
'PDL=SCALAR(0x1008362e0)', 'PDL=SCALAR(0x100891188)',
'PDL=SCALAR(0x1008360e8)', 'PDL=SCALAR(0x10082e3b8)')
I was able to trace it to a line in the .xs file, but I'm still at a
loss. I assume it is telling me that I'm calling the pp_def'd function
qpgen2 with the wrong arguments. I think the relevant code is:
pp_def("qpgen2",
Pars => 'dmat(m,m); dvec(m); int fddmat(); int n();
[o]sol(m); [o]lagr(q); [o]crval();
amat(m,q); bvec(q); int fdamat(); int q(); int meq();
int [o]iact(q); int [o]nact();
int [o]iter(s=2); [t]work(z); int [io]ierr();
', ....
and the call:
my $res = qpgen2(
$Dmat, # isPDL: Double D [3,3]
$dvec, # isPDL: Double D [3]
$n, # isPDL: Double D []
$n, # isPDL: Double D []
$sol, # isPDL: Double D [3]
$lagr, # isPDL: Double D [2]
$crval, # isPDL: Double D []
$Amat, # isPDL: Double D [2,3]
$bvec, # isPDL: Double D [1,2]
$n, # isPDL: Double D []
$q, # isPDL: Double D []
$meq, # isPDL: Double D []
$iact, # isPDL: Double D []
$nact, # isPDL: Double D []
$iter, # isPDL: Double D [2]
$work, # isPDL: Double D [18]
$ierr, # isPDL: Double D []
);
I've uploaded the project to github, if anyone would like dig deeper:
https://github.com/mvgrimes/pdl-opt-qp
Thanks,
Mark
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
