g77 problem with -lstdc++
Paul Smith
phhs80 at gmail.com
Sun Apr 27 22:22:33 UTC 2008
On Sun, Apr 27, 2008 at 10:50 PM, Daniel B. Thurman <dant at cdkkt.com> wrote:
> Paul Smith wrote:
> |> Actually, I am trying to compile Algacan from
> |>
> |> http://www.ime.usp.br/~egbirgin/tango/codes.php
> |>
> |> and I am not being successful.
> |
> |In addition, see
> |
> |http://www.ime.usp.br/~egbirgin/tango/c.php
> |
>
> I have gome to the ime.usp,br site and downloaded
> the Tango tar-bundle, and near the bottom of the page,
> downloaded the individual files for the "toy" examples.
>
> As for building the Tango tar-bundle, I noticed straight
> away that there is no -lstdc++ for any of the make
> files. So I have to assume that you are using it for
> some other application? In any case, I was not able
> to build the tar-bundled application since the Makefiles
> do not appear to be compatable to Linux(Fedora). Doing a
> 'make clean' resulted in an infinite loop, for example.
>
> As for the "toy" applications, given the compile instructions,
> on the download page there was no problem in doing so:
>
> g77 -O4 -xf77-cpp-input algencanma.f toyprob.f algencan.f -o algencanma
> (No Errors reported.)
>
> Running ./algencanma resulted:
> ========================================================
>
> The optional specification file algencan.dat was not found in the current
> directory (this is not a problem nor an error). The default values for the
> ALGENCAN parameters will be used.
>
> ALGENCAN PARAMETERS:
> rhoauto = T
> rhotype = 1
> rhomult = 1.0000E+01
> rhofrac = 5.0000E-01
> gtype = 0
> hptype = 6
> intype = 0
> precond = QNCGNA
> checkder = F
> epsfeas = 1.0000E-04
> epsopt = 1.0000E-04
> maxoutit = 50
> maxtotit = 1000000
> maxtotfc = 5000000
> iprint = 1
> ncomp = 5
>
> Entry to ALGENCAN.
>
> Number of variables: 2 Number of constraints: 2
>
> Lower bounds (first 2 components):
> -1.0000E+01 -1.0000E+20
>
> Upper bounds (first 2 components):
> 1.0000E+01 1.0000E+20
>
> ALGENCAN outer iteration: 0
>
> Current point (first 2 components):
> 0.0000E+00 0.0000E+00
>
> Updated Lagrange multipliers (first 2 components):
> 0.0000E+00 0.0000E+00
>
> Updated penalty parameters (first 2 components):
> 4.0000E+00 4.0000E+00
>
> Objective function value = 0.0000E+00
> Sup-norm of the constraints = 2.0000E+00
> Sup-norm of the projected gradient of the Lagrangian = 1.0000E+00
>
> Up-to-now number of inner iterations = 0
> Up-to-now number of Augmented Lagrangian function evaluations = 1
> Up-to-now number of Augmented Lagrangian gradient evaluations = 1
>
> The penalty parameters are not modified
> after the resolution of the first subproblem.
>
> ALGENCAN outer iteration: 1
>
> Current point (first 2 components):
> 6.0716E-01 1.2557E+00
>
> Updated Lagrange multipliers (first 2 components):
> 4.5159E-01 5.4838E-01
>
> Updated penalty parameters (first 2 components):
> 4.0000E+00 4.0000E+00
>
> Objective function value = 1.2557E+00
> Sup-norm of the constraints = 1.3709E-01
> Sup-norm of the projected gradient of the Lagrangian = 3.5620E-05
>
> Up-to-now number of inner iterations = 6
> Up-to-now number of Augmented Lagrangian function evaluations = 9
> Up-to-now number of Augmented Lagrangian gradient evaluations = 9
>
> Desired feasibility improvement was achieved,
> so, penalty parameters will not be modified.
>
> ALGENCAN outer iteration: 2
>
> Current point (first 2 components):
> 6.1713E-01 1.3819E+00
>
> Updated Lagrange multipliers (first 2 components):
> 4.4757E-01 5.5242E-01
>
> Updated penalty parameters (first 2 components):
> 4.0000E+00 4.0000E+00
>
> Objective function value = 1.3819E+00
> Sup-norm of the constraints = 1.0109E-03
> Sup-norm of the projected gradient of the Lagrangian = 1.0746E-05
>
> Up-to-now number of inner iterations = 9
> Up-to-now number of Augmented Lagrangian function evaluations = 14
> Up-to-now number of Augmented Lagrangian gradient evaluations = 14
>
> Desired feasibility improvement was achieved,
> so, penalty parameters will not be modified.
>
> ALGENCAN outer iteration: 3
>
> Current point (first 2 components):
> 6.1796E-01 1.3819E+00
>
> Updated Lagrange multipliers (first 2 components):
> 4.4728E-01 5.5280E-01
>
> Updated penalty parameters (first 2 components):
> 4.0000E+00 4.0000E+00
>
> Objective function value = 1.3819E+00
> Sup-norm of the constraints = 9.4348E-05
> Sup-norm of the projected gradient of the Lagrangian = 8.1567E-05
>
> Up-to-now number of inner iterations = 10
> Up-to-now number of Augmented Lagrangian function evaluations = 17
> Up-to-now number of Augmented Lagrangian gradient evaluations = 17
>
> Flag of ALGENCAN = 0 (Convergence with Sup-norm of the Lagrangian projected
> gradient smaller than 1.0000E-04, constraints Sup-norm smaller than
> 1.0000E-04, and largest complementarity violation smaller than 1.0000E-04)
> ========================================================
>
>
> So, IMO, the f/g77 compiler works fine. You just need to
> work on porting Tango Makefiles to Linux(Fedora).
Thanks, Daniel. Have you tried toyprob.c? The C interface is the
problematic one as well as the R one.
Paul
More information about the users
mailing list