This is quite odd; due to changes we made in the 7.1 command line
parsing stuff, we've seen it grab additional flags that it wasn't
supposed to, but I haven't seen it grab *half* a command line argument
before.
Can you send your config.log file? (please compress) There should be
some details in there that could help shed some light on this problem.
Also, can you give the latest 7.1.2 beta a whirl
(http://www.lam-mpi.org/beta/)? We've made a few changes in the
command line parsing code in it.
Thanks!
On Nov 24, 2004, at 11:32 AM, Pierre Valiron wrote:
> Hi all,
>
> I have just bought a v20z bi-opteron from Sun and I am testing the
> first pre-release of Solaris 10 with 64 bit support on the Opteron.
>
> I installed the lam-7.1.1 using the native Sun compilers and 64 bit
> flags.
>
> setenv CC cc
> setenv CXX CC
> setenv FC f95
>
> setenv CFLAGS "-O -xtarget=opteron -xarch=amd64"
> setenv CXXFLAGS "-O -xtarget=opteron -xarch=amd64"
> setenv FFLAGS "-O -xtarget=opteron -xarch=amd64"
>
> setenv LDFLAGS "-xtarget=opteron -xarch=amd64"
> setenv CXXLDFLAGS "-xtarget=opteron -xarch=amd64"
>
> Configure and make run like a charm.
> Then mpicc or mpif77 invoke the compiler just fine:
>
> valiron_at_h5 ~/mul > mpicc -xtarget=opteron -xarch=amd64 -c timing.c
> -showme
> cc -I/usr/local/include -mt -xtarget=opteron -xarch=amd64 -c timing.c
>
> valiron_at_h5 ~/mul > mpif77 -xtarget=opteron -xarch=amd64 -c xerbla.f
> -showme
> f95 -I/usr/local/include -xtarget=opteron -xarch=amd64 -c xerbla.f
>
>
>
> However the compilation step add an erroneous '-xtarget' option which
> is rejected as a fatal error by the cc and f95 compilers:
>
> valiron_at_h5 ~/mul > mpicc -xtarget=opteron -xarch=amd64 timing.c -showme
> cc -I/usr/local/include -mt -xtarget=opteron -xarch=amd64 timing.c
> -xtarget -lthread -L/usr/local/lib -llammpio -llamf77mpi -lmpi -llam
> -laio -laio -lsocket -lnsl -lthread -lthread -ldl
>
> valiron_at_h5 ~/mul > mpif77 xerbla.f -showme
> f95 -I/usr/local/include -xtarget=opteron -xarch=amd64 xerbla.f
> -xtarget -lthread -L/usr/local/lib -llammpio -llamf77mpi -lmpi -llam
> -laio -laio -lsocket -lnsl -lthread -lthread -ldl
>
> Did I made any mistake ? Is there any simple patch to fix this
> problem ?
>
>
>
> Fortunately I can still link by pasting "by hand" the correct link
> options and exercise the library.
> The resulting LAM/MPI seems quite good on the biprocessor.
>
> - MPI_Alltoall latency is about 7 microseconds
>
> - Typical all2all throughput is over 500 MB/s, and peaks to about 800
> MB/s for 64k buffers.
>
> Pierre.
>
> --
> Soutenez le mouvement SAUVONS LA RECHERCHE :
> http://recherche-en-danger.apinc.org/
>
> _/_/_/_/ _/ _/ Dr. Pierre VALIRON
> _/ _/ _/ _/ Laboratoire d'Astrophysique
> _/ _/ _/ _/ Observatoire de Grenoble / UJF
> _/_/_/_/ _/ _/ BP 53 F-38041 Grenoble Cedex 9 (France)
> _/ _/ _/ http://www-laog.obs.ujf-grenoble.fr
> _/ _/ _/ mail: Pierre.Valiron_at_[hidden]
> _/ _/ _/ Phone: +33 4 7651 4787 Fax: +33 4 7644 8821
> _/ _/_/
>
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
