This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Testing g77 using LAPACK -
- To: "Billinghurst, David (RTD)" <David dot Billinghurst at riotinto dot com dot au>
- Subject: Re: Testing g77 using LAPACK -
- From: Toon Moene <toon at moene dot indiv dot nluug dot nl>
- Date: Sun, 16 Aug 98 23:17:19 +0200
- Cc: "'egcs at cygnus dot com'" <egcs at cygnus dot com>
- Organization: Moene Computational Physics, Maartensdijk, The Netherlands
- References: <BE186FF1FA63D111B6510000C07BDF0D4D04B5@CRCMAIL.crts.techaust.riotinto.com.au>
> Just built and tested LAPACK using g77 from egcs-19980803
> on mips-sgi-irix6.2. This is a fair test as it involves
> building the BLAS, LINPACK, EISPACK and LAPACK libraries
> I have modified the makefiles to enable the package to
> be built outside the source directory, so it should be
> simple to run it regularly and check for regressions.
That's interesting to have - could you send me the modified
makefiles; I think I might be able to put them to some good use ...
> There is one failure - a core dump in directory
> LAPACK/TIMING for "xeigtimz < zseptim.in" Before I follow
> it up, is it a known problem?
Well, it's not one that I know of. Could you make a standalone
program of it that fails ? As far as I can see from the Makefile in
LAPACK/TIMING/EIG, this cannot be to hard. If it's not too large -
note the egcs list's limit of 40K - you might send it to the list,
*including* the compile flags that were used *and* the input file.
Please, make sure before you do so that the standalone source indeed
fails.
Unfortunately, the one (Fortran) person I know who has access to
IRIX 6.2 systems is Dave Love, and he just went away for a week.
Thanks,
Toon.