-------- Forwarded Message -------- From: Prabhu Manyem <[email protected]> To: Andrew Makhorin <[email protected]> Cc: [email protected] Subject: Re: GLPK Exact option not working 2 different glpsol Date: Sun, 18 Sep 2022 12:44:00 +0930
> Dear Andrew, > > OK, I re-installed GLPK again (as root)... But now, after > re-installation, there are two versions of GLPK, and both have almost > the same time stamp (Sept 18 12:28 and 12:29): > > root> ls -al /opt/glpk-5.0/examples/glpsol > -rwxr-xr-x 1 root root 6149 Sep 18 12:28 /opt/glpk-5.0/examples/glpsol > > root> ls -al /usr/local/bin/glpsol > -rwxr-xr-x 1 root root 98904 Sep 18 12:29 /usr/local/bin/glpsol > > The first glpsol (in the glpk examples directory) works well with the > "--exact" option. > > The second glpsol (at /usr/local/bin/) does NOT work with the > "--exact" option. > > I don't understand why GLPK installs 2 different versions of glpsol in > 2 different directories.. Strange. > > Anyway, I will start using the first glpsol at: /opt/glpk- > 5.0/examples/glpsol > > Thank you! > > -Prabhu > > > On Sun, 18 Sept 2022 at 10:56, Andrew Makhorin <[email protected]> wrote: > > > > Hi Prabhu, > > > > Thank you for your interest in glpk. > > > > > > I used the following commands to build glpk 5.0: > > > > ./configure --with-gmp > > > > make > > > > cd examples > > > > ./glpsol --mps ./murtagh.mps --max --exact > > > > and all works correctly: > > > > GLPSOL--GLPK LP/MIP Solver 5.0 > > Parameter(s) specified in the command line: > > --mps ./murtagh.mps --max --exact > > Reading problem data from './murtagh.mps'... > > Problem: OILREFI > > Objective: PROFIT > > 74 rows, 81 columns, 504 non-zeros > > 600 records were read > > One free row was removed > > glp_exact: 73 rows, 81 columns, 474 non-zeros > > GNU MP bignum library is being used > > * 0: objval = 0 (42) > > * 103: objval = 126.057124110517 (0) > > OPTIMAL SOLUTION FOUND > > Time used: 0.1 secs > > Memory used: 0.1 Mb (114012 bytes) > > > > Please check if the GNU MP library is correctly installed. > > If so, after running the configure script check the file > > config.h created by the script and make sure that using the > > GMP library is enabled: > > > > #define HAVE_GMP 1 > > /* defined if the GNU MP bignum library is available */ > > /* requires <gmp.h> and -lgmp */ > > > > > > Best regards, > > > > Andrew Makhorin > > > > > > PS: Please report bugs to <[email protected]>. Thanks. > > > > > > > > > > On Sat, 2022-09-17 at 12:29 +0930, Prabhu Manyem wrote: > > > Dear Andrew, > > > > > > The "exact" option in my GLPK 5.0 (command line) is failing.. It > > > says > > > "GLPK bignum module is being used (Consider installing GNU MP to > > > attain a much better performance.)" > > > > > > However, I have installed GMP, and the GLPK 5.0 compilation and > > > installation have also used GMP... Strange! > > > > > > Let me attach my installation config file.. I copied the original > > > config file into a new file called config-log.txt. > > > > > > My configure command: > > > ./configure --with-gmp --enable-dl=ltdl > > > > > > Some lines about GMP in the config file: > > > configure:8316: checking gmp.h usability > > > configure:8316: gcc -c -g -O2 conftest.c >&5 > > > configure:8316: $? = 0 > > > configure:8316: result: yes > > > configure:8316: checking gmp.h presence > > > configure:8316: gcc -E conftest.c > > > configure:8316: $? = 0 > > > configure:8316: result: yes > > > configure:8316: checking for gmp.h > > > configure:8316: result: yes > > > configure:8326: checking whether to use GNU MP bignum library > > > configure:8329: result: yes > > > : > > > ac_cv_header_gmp_h=yes > > > : > > > LIBS='-lltdl -lgmp -lm ' > > > > > > I downloaded and installed gmp-6.2.1, and then, I installed GLPK > > > again. > > > > > > LTDL has also been installed... Let me also attach the results of > > > my > > > "locate gmp" command in my Linux machine. > > > > > > Any idea why the GLPK (at the command line) is unable to recognise > > > GMP? Thank you very much! > > > > > > -Prabhu > > > > > > Dr Prabhu Manyem > > > Retired Professor of Applied Mathematics > > > Nanchang Institute of Technology > >
