<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Hi Andreas and Dominique.<br>
<br>
First of all, I should thank you both very very much for your kind help!<br>
<br>
Here, I can find "lib" subdirectory under $IPOPT. "libipopt.a" is
inside.<br>
However, in my installation, I can't find "libiptron.a". The following
is<br>
all my *.a:<br>
[lihong@frosty]Ipopt> find . -name "*.a" -ls<br>
123226370 4 -rw-r--r-- 1 lihong postdoc_f 2824 Nov 8 23:02
./IPOPT/AMPL_interface/libipampl.a<br>
139592829 980 -rw-r--r-- 1 lihong postdoc_f 998556 Nov 8 23:02
./IPOPT/ipopt/libipcore.a<br>
145015332 160 -rw-r--r-- 1 lihong postdoc_f 156452 Nov 8 22:56
./OTHERS/blas/libipblas.a<br>
128582542 732 -rw-r--r-- 1 lihong postdoc_f 742080 Nov 8 23:01
./OTHERS/ASL/amplsolver.a<br>
115196859 192 -rw-r--r-- 1 lihong postdoc_f 191048 Nov 8 22:57
./OTHERS/lapack/libiplapack.a<br>
118669402 80 -rw-r--r-- 1 lihong postdoc_f 76920 Nov 8 22:57
./OTHERS/HSL/libiphsl.a<br>
153551147 1396 -rw-r--r-- 1 lihong postdoc_f 1422756 Nov 8 23:03
./lib/libipopt.a<br>
<br>
I checked the $IPOPT/doc/INSTALL and got to know "libiptron.a" may <br>
or may not be used in the new release of IPOPT. Should I include it<br>
in my installation?<br>
<br>
Now let me show the result obtained here. Something is still wrong<br>
(maybe my compilation problem).<br>
I followed the instruction from Dominique to revise the
$MYCUTER/bin/ipopt<br>
(ie, Line168 and Line181). <br>
1) If I define ipopt.liblist as follows:<br>
-L/homes/lihong/cvs/COIN/Ipopt/lib -lipopt<br>
<br>
I got the following after issueing "sdipopt HS11.SIF":<br>
****************************************************<br>
[lihong@si]tmp> sdipopt HS11.SIF<br>
<br>
Problem name: HS11 <br>
<br>
Double precision version will be formed. <br>
<br>
The objective function uses 1 nonlinear group <br>
<br>
There is 1 nonlinear inequality constraint <br>
<br>
There are 2 free variables<br>
<br>
<br>
ld: warning: symbol `evals_' has differing sizes:<br>
(file
/homes/lihong/cvs/cuter/CUTEr.medium.sun.sol.g77/double/bin/ipoptma.o
value=0x8; file <br>
/homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ipopt.o) value=0xc);<br>
/homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ipopt.o) definition
taken<br>
ld: fatal: symbol `idamax_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(idamax.o)
type=FUNC; file /homes/lihong/cv<br>
s/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dswap_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dswap.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dcopy_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dcopy.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dscal_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dscal.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `daxpy_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(daxpy.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dasum_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dasum.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dnrm2_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dnrm2.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `ddot_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ddot.o)
type=FUNC; file /homes/lihong/cvs/<br>
cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: File processing errors. No output written to ipoptmin<br>
collect2: ld returned 1 exit status<br>
/atm/vlsifiles/d0/mscad/home/lihong/cvs/cuter/common/sif/tmp/ipoptmin:
Command not found<br>
****************************************************<br>
<br>
2) If I define ipopt.liblist as follows:<br>
-L/homes/lihong/cvs/COIN/Ipopt/lib -lipopt -llapack -lblas<br>
<br>
I got the following after issueing "sdipopt HS11.SIF":<br>
****************************************************<br>
[lihong@si]tmp> sdipopt HS11.SIF<br>
<br>
Problem name: HS11 <br>
<br>
Double precision version will be formed. <br>
<br>
The objective function uses 1 nonlinear group <br>
<br>
There is 1 nonlinear inequality constraint <br>
<br>
There are 2 free variables<br>
<br>
<br>
ld: warning: symbol `evals_' has differing sizes:<br>
(file
/homes/lihong/cvs/cuter/CUTEr.medium.sun.sol.g77/double/bin/ipoptma.o
value=0x8; file <br>
/homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ipopt.o) value=0xc);<br>
/homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ipopt.o) definition
taken<br>
ld: fatal: library -llapack: not found<br>
ld: fatal: library -lblas: not found<br>
ld: fatal: symbol `idamax_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(idamax.o)
type=FUNC; file /homes/lihong/cv<br>
s/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dswap_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dswap.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dcopy_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dcopy.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dscal_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dscal.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `daxpy_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(daxpy.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dasum_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dasum.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `dnrm2_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(dnrm2.o)
type=FUNC; file /homes/lihong/cvs<br>
/cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: symbol `ddot_' is multiply-defined:<br>
(file /homes/lihong/cvs/COIN/Ipopt/lib/libipopt.a(ddot.o)
type=FUNC; file /homes/lihong/cvs/<br>
cuter/CUTEr.medium.sun.sol.g77/double/bin/linpac.o type=FUNC);<br>
ld: fatal: File processing errors. No output written to ipoptmin<br>
collect2: ld returned 1 exit status<br>
/atm/vlsifiles/d0/mscad/home/lihong/cvs/cuter/common/sif/tmp/ipoptmin:
Command not found<br>
*******************************************************<br>
<br>
3) If I define ipopt.liblist as follows:<br>
-L/homes/lihong/cvs/COIN/Ipopt/IPOPT/ipopt -lipcore
-L/homes/lihong/cvs/COIN/Ipopt/OTHERS/TRON -liptron
-L/homes/lihong/cvs/COIN/Ipopt/IPOPT/ipopt -lipcore
-L/homes/lihong/cvs/COIN/Ipopt/OTHERS/HSL -liphsl
-L/homes/lihong/cvs/COIN/Ipopt/OTHERS/lapack -liplapack
-L/homes/lihong/cvs/COIN/Ipopt/OTHERS/blas -lipblas
-L/homes/lihong/cvs/COIN/Ipopt/lib -lipopt<br>
<br>
I got the following after issueing "sdipopt HS11.SIF":<br>
****************************************************<br>
[lihong@si]tmp> sdipopt HS11.SIF<br>
<br>
Problem name: HS11 <br>
<br>
Double precision version will be formed. <br>
<br>
The objective function uses 1 nonlinear group <br>
<br>
There is 1 nonlinear inequality constraint <br>
<br>
There are 2 free variables<br>
<br>
<br>
ld: warning: symbol `evals_' has differing sizes:<br>
(file
/homes/lihong/cvs/cuter/CUTEr.medium.sun.sol.g77/double/bin/ipoptma.o
value=0x8; file
/homes/lihong/cvs/COIN/Ipopt/IPOPT/ipopt/libipcore.a(ipopt.o)
value=0xc);<br>
/homes/lihong/cvs/COIN/Ipopt/IPOPT/ipopt/libipcore.a(ipopt.o)
definition taken<br>
ld: fatal: library -liptron: not found<br>
ld: fatal: File processing errors. No output written to ipoptmin<br>
collect2: ld returned 1 exit status<br>
/atm/vlsifiles/d0/mscad/home/lihong/cvs/cuter/common/sif/tmp/ipoptmin:
Command not found<br>
*****************************************************<br>
<br>
Above, my IPOPT installation directory is /homes/lihong/cvs/COIN/Ipopt
(ie, $IPOPT).<br>
By the way, my computer is Sun workstation using Solaris.<br>
<br>
Could you please help me to figure out the reason? If you need<br>
any further information about my installation environment, please<br>
let me know. Thank you very much for your help.<br>
<br>
Best regards,<br>
<br>
Lihong<br>
<br>
<br>
Andreas Waechter wrote:
<blockquote
cite="midPine.A41.4.44.0511091442560.43442-100000@bigopt.watson.ibm.com"
type="cite">
<pre wrap="">Hi Dominique,
</pre>
<blockquote type="cite">
<pre wrap="">What you say makes total sense and also corresponds to my memories, but
something seems wrong as when I unpack IPOPT, there is no lib
subdirectory! It is not created either when I say 'make'. So there's no
libipopt.a to be found.
</pre>
</blockquote>
<pre wrap=""><!---->
'make' doesn't install the code, you need to type 'make install'. You are
right, there is no lib subdir when you unpack the code, but it will be
created (and filled) with 'make install'
Could you please try it again?
Thanks! :)
Andreas
</pre>
</blockquote>
<br>
<br>
<pre class="moz-signature" cols="72">
</pre>
</body>
</html>