Custom Query (122 matches)


Show under each result:

Results (91 - 93 of 122)

Ticket Resolution Summary Owner Reporter
#64 fixed improve capability to use precompiled MUMPS library andreasw stefan


the current way to provide a user-compiled MUMPS library is to specify a directory via --with-mumps-dir. Then the build system assumes to find the MUMPS header files there and the libraries $mumps_dir/lib/libdmumps.$libe, $mumps_dir/lib/libpord.$libe, $mumps_dir/libseq/libmpiseq.$libe. This probably works fine if MUMPS had been compiled with the MUMPS Makefiles.

However, in case that one has a mumps library (maybe as one instead of three files as produced by ThirdParty/Mumps) and header files in other locations, it seem to be extremly difficult to make Ipopt use this library. For other 3rd party codes, one can specify the link line and include flags via --with-PROJ-lib and --with-PROJ-incdir. This seem to offer more flexibility. Further, a --with-mumps-lib would allow to specify mumps-library-specific linking flags, e.g., in case that fortran runtime libraries from a different compiler are needed.

Finally, having also a --enable-mumps-libcheck option would be great, since it allows to build an Ipopt library that includes a Mumps interface without having the MUMPS library actually present or linking.

Thank you, Stefan

#66 fixed Lapack hangs in dlamc1 if in optimized build andreasw stefan


in some cases, CBC (2.1 stable) and SCIP (1.0; using Clp as LP solver) hang in the Lapack routine dlamc1 if I have compiled in optimized mode (i.e., default).

The compiler is a GCC 4.2.1, the machine a Intel Core2 Duo.

John pointed me to this ticket in the Octave mailing list. I followed the suggestions there and the problem went away when I added -ffloat-store to the FFLAGS.

Should this flag be added per default (when gcc is used), at least for dlamch.f ?

Best, Stefan

PS: This is from the gcc manual on what -ffloat-store is doing:

Do not store floating point variables in registers, and inhibit other options that might change whether a floating point value is taken from a register or memory.

This option prevents undesirable excess precision on machines such as the 68000 where the floating registers (of the 68881) keep more precision than a "double" is supposed to have. Similarly for the x86 architecture. For most programs, the excess precision does only good, but a few programs rely on the precise definition of IEEE floating point. Use -ffloat-store for such programs, after modifying them to store all pertinent intermediate computations into variables.

#67 fixed Running configure without g++ => cryptic error andreasw plb

While setting up bcp (with CoinUtils?? 2.2.5), I ran into the following problem :

configure: error: Cannot find integer type with 64 bits configure: error: /bin/bash '../../CoinUtils?/configure' failed for CoinUtils?

I'm enclosing config.log.

I tried several times. I had gcc installed, but not g++. I set up g++ and ran configure, and it went just fine. I'm enclosing the config.log of the successful run. If the failure was indeed caused by the absence of g++, it might be useful to make the error message more specific.

Note: See TracQuery for help on using queries.