From tradke at aei.mpg.de Thu Dec 4 06:59:36 2008 From: tradke at aei.mpg.de (Thomas Radke) Date: Thu, 04 Dec 2008 13:59:36 +0100 Subject: [Developers] [Users] Cross-compiling trigger. In-Reply-To: <49D70175-B5FD-4527-8A9D-E2D38FE07A14@cct.lsu.edu> References: <4935E5EE.40108@cct.lsu.edu> <493658BE.1050400@gravity.psu.edu> <49D70175-B5FD-4527-8A9D-E2D38FE07A14@cct.lsu.edu> Message-ID: <4937D438.3070204@aei.mpg.de> Erik Schnetter wrote: > Yes, I think that a new option CROSS_COMPILE would be the best way to > go, defaulting to "no" (obviously). > > As far as I can see, this requires changes to configure.in. Unless I'm > mistaken, the only change there would be to abort if cross-compiling is > detected unless CROSS_COMPILE=yes. That sounds rather simple -- am I > missing something? I can't see what else would need to be changed. So I propose the attached patch which implements the new CROSS_COMPILE option. -- Cheers, Thomas. --- StripMime Report -- processed MIME parts --- multipart/mixed text/plain (text body -- kept) text/x-patch --- From bentivegna at cct.lsu.edu Thu Dec 4 16:18:15 2008 From: bentivegna at cct.lsu.edu (Eloisa Bentivegna) Date: Thu, 04 Dec 2008 16:18:15 -0600 Subject: [Developers] [Users] Cross-compiling trigger. In-Reply-To: <4937D438.3070204@aei.mpg.de> References: <4935E5EE.40108@cct.lsu.edu> <493658BE.1050400@gravity.psu.edu> <49D70175-B5FD-4527-8A9D-E2D38FE07A14@cct.lsu.edu> <4937D438.3070204@aei.mpg.de> Message-ID: <49385727.60408@cct.lsu.edu> Thomas Radke ha scritto: > Erik Schnetter wrote: >> Yes, I think that a new option CROSS_COMPILE would be the best way to >> go, defaulting to "no" (obviously). >> >> As far as I can see, this requires changes to configure.in. Unless I'm >> mistaken, the only change there would be to abort if cross-compiling is >> detected unless CROSS_COMPILE=yes. That sounds rather simple -- am I >> missing something? > > I can't see what else would need to be changed. So I propose the > attached patch which implements the new CROSS_COMPILE option. Thanks! Eloisa From tradke at aei.mpg.de Fri Dec 5 05:12:50 2008 From: tradke at aei.mpg.de (Thomas Radke) Date: Fri, 05 Dec 2008 12:12:50 +0100 Subject: [Developers] [Users] Cross-compiling trigger. In-Reply-To: <49385727.60408@cct.lsu.edu> References: <4935E5EE.40108@cct.lsu.edu> <493658BE.1050400@gravity.psu.edu> <49D70175-B5FD-4527-8A9D-E2D38FE07A14@cct.lsu.edu> <4937D438.3070204@aei.mpg.de> <49385727.60408@cct.lsu.edu> Message-ID: <49390CB2.3040407@aei.mpg.de> Eloisa Bentivegna wrote: > Thomas Radke ha scritto: >> Erik Schnetter wrote: >>> Yes, I think that a new option CROSS_COMPILE would be the best way to >>> go, defaulting to "no" (obviously). >>> >>> As far as I can see, this requires changes to configure.in. Unless I'm >>> mistaken, the only change there would be to abort if cross-compiling is >>> detected unless CROSS_COMPILE=yes. That sounds rather simple -- am I >>> missing something? >> I can't see what else would need to be changed. So I propose the >> attached patch which implements the new CROSS_COMPILE option. Hi Eloisa, the patch I sent around yesterday contained an error which is fixed in the version I just committed to the cactusdevcvs repository. -- Cheers, Thomas. From tradke at aei.mpg.de Fri Dec 5 05:13:27 2008 From: tradke at aei.mpg.de (Thomas Radke) Date: Fri, 05 Dec 2008 12:13:27 +0100 Subject: [Developers] new configure option to enable/disable OpenMP support In-Reply-To: <49146B51.8070106@aei.mpg.de> References: <49146B51.8070106@aei.mpg.de> Message-ID: <49390CD7.3060705@aei.mpg.de> Thomas Radke wrote: > Hi, > > I propose the attached patch which introduces a new configure option > OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus > configuration. So far Intel, PathScale, and GNU compilers under Linux > and IBM compilers under AIX are supported. Ok to apply ? -- Cheers, Thomas. From jtao at cct.lsu.edu Fri Dec 5 11:08:48 2008 From: jtao at cct.lsu.edu (Jian Tao) Date: Fri, 05 Dec 2008 11:08:48 -0600 Subject: [Developers] new configure option to enable/disable OpenMP support In-Reply-To: <49390CD7.3060705@aei.mpg.de> References: <49146B51.8070106@aei.mpg.de> <49390CD7.3060705@aei.mpg.de> Message-ID: <49396020.90508@cct.lsu.edu> Hi Thomas, On my laptop, I used the following: MPI=OpenMPI OPENMPI_LIB_DIR=/usr/lib64/openmpi/1.2.4-gcc OPENMPI_INC_DIR=/usr/include/openmpi/1.2.4-gcc As one of MPI implementations, OpenMPI might be better configured with the MPI option. Are there any reasons why we want an options for OpenMPI itself ? Regards, Jian Thomas Radke wrote: > Thomas Radke wrote: >> Hi, >> >> I propose the attached patch which introduces a new configure option >> OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus >> configuration. So far Intel, PathScale, and GNU compilers under Linux >> and IBM compilers under AIX are supported. > > Ok to apply ? > From bentivegna at cct.lsu.edu Fri Dec 5 11:45:39 2008 From: bentivegna at cct.lsu.edu (Eloisa Bentivegna) Date: Fri, 05 Dec 2008 11:45:39 -0600 Subject: [Developers] [Users] Cross-compiling trigger. In-Reply-To: <49390CB2.3040407@aei.mpg.de> References: <4935E5EE.40108@cct.lsu.edu> <493658BE.1050400@gravity.psu.edu> <49D70175-B5FD-4527-8A9D-E2D38FE07A14@cct.lsu.edu> <4937D438.3070204@aei.mpg.de> <49385727.60408@cct.lsu.edu> <49390CB2.3040407@aei.mpg.de> Message-ID: <493968C3.4040402@cct.lsu.edu> Thomas Radke ha scritto: > Hi Eloisa, > > the patch I sent around yesterday contained an error which is fixed in > the version I just committed to the cactusdevcvs repository. OK. I see you've also cut the error redirection to /dev/null; thanks! Eloisa From schnetter at cct.lsu.edu Fri Dec 5 12:54:21 2008 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Fri, 5 Dec 2008 10:54:21 -0800 Subject: [Developers] [Patches] new configure option to enable/disable OpenMP support In-Reply-To: <49390CD7.3060705@aei.mpg.de> References: <49146B51.8070106@aei.mpg.de> <49390CD7.3060705@aei.mpg.de> Message-ID: On Dec 5, 2008, at 03:13:27, Thomas Radke wrote: > Thomas Radke wrote: >> Hi, >> >> I propose the attached patch which introduces a new configure option >> OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus >> configuration. So far Intel, PathScale, and GNU compilers under Linux >> and IBM compilers under AIX are supported. > > Ok to apply ? Thomas I must have missed your original email -- I can't find it on my system, but I do see it in the list archives. I also didn't see the email about your configuration changes. Maybe my subscription went wrong; I'll check. The patch looks fine. A very minor issue: In the documentation, I would change "[...] to use no OpenMP support" to "[...] to disable OpenMP", as "use no" sounds confusing. -erik PS: I usually exclude the file "configure" when I send patches, since this file is large and automatically generated. -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from www.keyserver.net. -------------- next part -------------- A non-text attachment was scrubbed... Name: PGP.sig Type: application/pgp-signature Size: 194 bytes Desc: This is a digitally signed message part Url : http://www.cactuscode.org/pipermail/developers/attachments/20081205/99558709/attachment.bin From schnetter at cct.lsu.edu Fri Dec 5 12:55:15 2008 From: schnetter at cct.lsu.edu (Erik Schnetter) Date: Fri, 5 Dec 2008 10:55:15 -0800 Subject: [Developers] new configure option to enable/disable OpenMP support In-Reply-To: <49396020.90508@cct.lsu.edu> References: <49146B51.8070106@aei.mpg.de> <49390CD7.3060705@aei.mpg.de> <49396020.90508@cct.lsu.edu> Message-ID: Jian Unfortunately there are both OpenMPI and OpenMP, and both are very different. The first (with an "I" at the end) is an MPI implementation, available as a library. The latter (without the "I") is a standard for multithreading that needs to be implemented by the compiler. -erik On Dec 5, 2008, at 09:08:48, Jian Tao wrote: > Hi Thomas, > > On my laptop, I used the following: > > MPI=OpenMPI > OPENMPI_LIB_DIR=/usr/lib64/openmpi/1.2.4-gcc > OPENMPI_INC_DIR=/usr/include/openmpi/1.2.4-gcc > > As one of MPI implementations, OpenMPI might be better configured > with the MPI option. Are there any reasons why we want an options > for OpenMPI itself ? > > Regards, > Jian > > Thomas Radke wrote: >> Thomas Radke wrote: >>> Hi, >>> >>> I propose the attached patch which introduces a new configure option >>> OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus >>> configuration. So far Intel, PathScale, and GNU compilers under >>> Linux >>> and IBM compilers under AIX are supported. >> >> Ok to apply ? >> > > _______________________________________________ > Developers mailing list > Developers at cactuscode.org > http://www.cactuscode.org/mailman/listinfo/developers -- Erik Schnetter http://www.cct.lsu.edu/~eschnett/ My email is as private as my paper mail. I therefore support encrypting and signing email messages. Get my PGP key from www.keyserver.net. -------------- next part -------------- A non-text attachment was scrubbed... Name: PGP.sig Type: application/pgp-signature Size: 194 bytes Desc: This is a digitally signed message part Url : http://www.cactuscode.org/pipermail/developers/attachments/20081205/7d4aa772/attachment.bin From jtao at cct.lsu.edu Fri Dec 5 13:16:21 2008 From: jtao at cct.lsu.edu (Jian Tao) Date: Fri, 05 Dec 2008 13:16:21 -0600 Subject: [Developers] new configure option to enable/disable OpenMP support In-Reply-To: References: <49146B51.8070106@aei.mpg.de> <49390CD7.3060705@aei.mpg.de> <49396020.90508@cct.lsu.edu> Message-ID: <49397E05.20809@cct.lsu.edu> Sorry, I misread the post. My mind was trying to do automatic completion. Regards, Jian Erik Schnetter wrote: > Jian > > Unfortunately there are both OpenMPI and OpenMP, and both are very > different. The first (with an "I" at the end) is an MPI implementation, > available as a library. The latter (without the "I") is a standard for > multithreading that needs to be implemented by the compiler. > > -erik > > On Dec 5, 2008, at 09:08:48, Jian Tao wrote: > >> Hi Thomas, >> >> On my laptop, I used the following: >> >> MPI=OpenMPI >> OPENMPI_LIB_DIR=/usr/lib64/openmpi/1.2.4-gcc >> OPENMPI_INC_DIR=/usr/include/openmpi/1.2.4-gcc >> >> As one of MPI implementations, OpenMPI might be better configured >> with the MPI option. Are there any reasons why we want an options >> for OpenMPI itself ? >> >> Regards, >> Jian >> >> Thomas Radke wrote: >>> Thomas Radke wrote: >>>> Hi, >>>> >>>> I propose the attached patch which introduces a new configure option >>>> OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus >>>> configuration. So far Intel, PathScale, and GNU compilers under Linux >>>> and IBM compilers under AIX are supported. >>> >>> Ok to apply ? >>> >> >> _______________________________________________ >> Developers mailing list >> Developers at cactuscode.org >> http://www.cactuscode.org/mailman/listinfo/developers > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Developers mailing list > Developers at cactuscode.org > http://www.cactuscode.org/mailman/listinfo/developers From tradke at aei.mpg.de Mon Dec 8 06:27:57 2008 From: tradke at aei.mpg.de (Thomas Radke) Date: Mon, 08 Dec 2008 13:27:57 +0100 Subject: [Developers] [Patches] new configure option to enable/disable OpenMP support In-Reply-To: References: <49146B51.8070106@aei.mpg.de> <49390CD7.3060705@aei.mpg.de> Message-ID: <493D12CD.1030100@aei.mpg.de> Erik Schnetter wrote: > On Dec 5, 2008, at 03:13:27, Thomas Radke wrote: > >> Thomas Radke wrote: >>> Hi, >>> >>> I propose the attached patch which introduces a new configure option >>> OpenMP=yes|no to enable/disable OpenMP compilation of a Cactus >>> configuration. So far Intel, PathScale, and GNU compilers under Linux >>> and IBM compilers under AIX are supported. >> >> Ok to apply ? > > > Thomas > > I must have missed your original email -- I can't find it on my system, > but I do see it in the list archives. I also didn't see the email about > your configuration changes. Maybe my subscription went wrong; I'll check. > > The patch looks fine. > > A very minor issue: In the documentation, I would change "[...] to use > no OpenMP support" to "[...] to disable OpenMP", as "use no" sounds > confusing. Thanks, Erik ! I committed the patch. -- Cheers, Thomas.