Re: [Help-glpk] Error message with glpsol on PowerPC

[Andrew Makhorin]

>   Thanx for the assistance.  Unfortunately, the results from the source
> and target platforms differed significantly.  Here is the result from my
> linux PowerPC 440EPx target:

> lpx_read_mps: reading problem data from `plan.mps'...
> lpx_read_mps: problem PLAN
> lpx_read_mps: 8 rows, 7 columns, 48 non-zeros
> lpx_read_mps: 54 cards were read
> glp_simplex: original LP has 8 rows, 7 columns, 48 non-zeros
> PROBLEM HAS NO PRIMAL FEASIBLE SOLUTION
> Time used:   0.0 secs
> Memory used: 0.0 Mb ((null) bytes)

There is something totally wrong. The fragment that displays two
last messages is the following:

      /* display statistics */
      print("Time used:   %.1f secs", xdifftime(xtime(), start));
      {  glp_ulong tpeak;
         char buf[50];
         lib_mem_usage(NULL, NULL, NULL, &tpeak);
         print("Memory used: %.1f Mb (%s bytes)",
            (4294967296.0 * tpeak.hi + tpeak.lo) / 1048576.0,
            ultoa(tpeak, buf, 10));
      }

Ultoa always returns a string pointer; it cannot return NULL.

Why not to try testing cross-compilation on a simple C program?

>   I repeated the same on the Windows/Cygwin source PC:

> lpx_read_mps: reading problem data from `plan.mps'...
> lpx_read_mps: problem PLAN
> lpx_read_mps: 8 rows, 7 columns, 48 non-zeros
> lpx_read_mps: 54 cards were read
> glp_simplex: original LP has 8 rows, 7 columns, 48 non-zeros
> glp_simplex: presolved LP has 7 rows, 7 columns, 41 non-zeros
> lpx_adv_basis: size of triangular part = 7
>       0:   objval =   6.500000000e+02   infeas =   1.000000000e+00 (0)
>       2:   objval =   4.376770833e+02   infeas =   0.000000000e+00 (0)
> *     2:   objval =   4.376770833e+02   infeas =   0.000000000e+00 (0)
> *     8:   objval =   2.962166065e+02   infeas =   0.000000000e+00 (0)
> OPTIMAL SOLUTION FOUND
> Time used:   0.0 secs
> Memory used: 0.1 Mb (54416 bytes)

>   The point of departure between the two appears to be during/immediately
> after the presolve step.  Hopefully one of them is correct, presumably the
> PC!  Not sure if this provides enough information on what to try next?