From jeff.sullivan at intel.com Wed Feb 25 20:19:34 2004 From: jeff.sullivan at intel.com (Sullivan, Jeff) Date: Wed, 25 Feb 2004 15:19:34 -0500 Subject: Problem with configure on Linux/ia32: Red Hat 8.0 Message-ID: <88E10B707C51824DA13AA135A402DF470109C33A@hdsmsx403.hd.intel.com> Hi, I hope that I am not asking a FAQ, but the FAQ doesn't appear to be there yet... I'm trying to build POOMA 2.4.0 on Linux/ia32, but am running into some problems with the configure and make. Here is my set-up % perl --version This is perl, v5.8.0 built for i386-linux-thread-multi % make --version GNU Make version 3.78.1, by Richard Stallman and Roland McGrath. % gmake --version GNU Make version 3.79.1, by Richard Stallman and Roland McGrath. Here's what I tried: % configure --arch LINUXgcc --opt --suite KO % setenv POOMASUITE KO // make causes these errors... % make { Lots of warnings about mt1, mt2, mt3, mt4, mt5... } make: /tmp/src/pooma-2.4.0/config/Shared/tail.mk:110: pipe: Too many open files /tmp/src/pooma-2.4.0/config/Shared/tail.mk:110: /config/Shared/rules.mk: No such file or directory make: /tmp/src/pooma-2.4.0/config/Shared/tail.mk:113: pipe: Too many open files make: *** No rule to make target `/config/Shared/rules.mk'. Stop. // gmake causes these errors... % gmake { Similar warnings about mt1, mt2, mt3, mt4, mt5... } gmake: stat:/usr7/jpsulliv/src/pooma-2.4.0/Utilities/Utilities/Utilities/Utilit ies/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilitie s/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/ Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Ut ilities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Util ities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilities/Utilit ies/Utilities/Utilities/Utilities/Utilities/.../subdir.mk: File name too long gmake: *** No rule to make target `/config/Shared/rules.mk'. Stop. The gmake errors look more "interesting" and might provide more clues as to what is wrong. Let me know if you have any suggestions. Any help would be appreciated. Thanks, -Jeff From rguenth at tat.physik.uni-tuebingen.de Thu Feb 26 09:08:46 2004 From: rguenth at tat.physik.uni-tuebingen.de (Richard Guenther) Date: Thu, 26 Feb 2004 10:08:46 +0100 Subject: [pooma-dev] Problem with configure on Linux/ia32: Red Hat 8.0 In-Reply-To: <88E10B707C51824DA13AA135A402DF470109C33A@hdsmsx403.hd.intel.com> References: <88E10B707C51824DA13AA135A402DF470109C33A@hdsmsx403.hd.intel.com> Message-ID: <403DB79E.2000101@tat.physik.uni-tuebingen.de> Sullivan, Jeff wrote: > Hi, > > I hope that I am not asking a FAQ, but the FAQ doesn't appear to be > there yet... > > I'm trying to build POOMA 2.4.0 on Linux/ia32, but am running into some > problems > with the configure and make. Here is my set-up > > % perl --version > This is perl, v5.8.0 built for i386-linux-thread-multi > > % make --version > GNU Make version 3.78.1, by Richard Stallman and Roland McGrath. > > % gmake --version > GNU Make version 3.79.1, by Richard Stallman and Roland McGrath. > > Here's what I tried: > > % configure --arch LINUXgcc --opt --suite KO > > % setenv POOMASUITE KO > > // make causes these errors... > % make > > { Lots of warnings about mt1, mt2, mt3, mt4, mt5... } Looks familiar, try removing lib/KO/depend.mk - maybe even configure spit out some messages about errors during dependency generation? Also rather than the 2.4.0 release you probably want to check out current CVS as it has many bugs fixed (notably compilation issues with the intel compiler and upcoming gcc 3.4 release). Hope this helps, Richard. From jeff.sullivan at intel.com Thu Feb 26 20:48:05 2004 From: jeff.sullivan at intel.com (Sullivan, Jeff) Date: Thu, 26 Feb 2004 15:48:05 -0500 Subject: [pooma-dev] Problem with configure on Linux/ia32: Red Hat 8.0 Message-ID: <88E10B707C51824DA13AA135A402DF470109C34C@hdsmsx403.hd.intel.com> -----Original Message----- From: Richard Guenther > Looks familiar, try removing lib/KO/depend.mk - maybe even configure > spit out some messages about errors during dependency generation? OK, I think that did help (removing lib/KO/depend.mk) Thanks! > Also rather than the 2.4.0 release you probably want to check out > current CVS as it has many bugs fixed (notably compilation issues with > the intel compiler and upcoming gcc 3.4 release). I am interested in building with the Intel compiler (on Linux), so I probably would be interested in the latest sources. Is there a config file for this already (LINUXicc)? I was planning to just modify LINUXgcc. I wasn't able to check out the sources, though: % cvs -d :pserver:anoncvs at pooma.codesourcery.com:/home/pooma/Repository login Logging in to :pserver:anoncvs at pooma.codesourcery.com:2401/home/pooma/Repository CVS password: cvs [login aborted]: connect to pooma.codesourcery.com(65.74.133.5):2401 failed: Connection timed out If you have other ideas that I should try, please let me know. Thanks again, -Jeff From oldham at codesourcery.com Fri Feb 27 15:45:06 2004 From: oldham at codesourcery.com (Jeffrey D. Oldham) Date: Fri, 27 Feb 2004 07:45:06 -0800 Subject: [pooma-dev] Problem with configure on Linux/ia32: Red Hat 8.0 In-Reply-To: <88E10B707C51824DA13AA135A402DF470109C34C@hdsmsx403.hd.intel.com> References: <88E10B707C51824DA13AA135A402DF470109C34C@hdsmsx403.hd.intel.com> Message-ID: <403F6602.30203@codesourcery.com> Sullivan, Jeff wrote: > -----Original Message----- > From: Richard Guenther > > >>Looks familiar, try removing lib/KO/depend.mk - maybe even configure >>spit out some messages about errors during dependency generation? > > > OK, I think that did help (removing lib/KO/depend.mk) Thanks! > > >>Also rather than the 2.4.0 release you probably want to check out >>current CVS as it has many bugs fixed (notably compilation issues with > > >>the intel compiler and upcoming gcc 3.4 release). > > > I am interested in building with the Intel compiler (on Linux), so I > probably would be interested in the latest sources. Is there a config > file for this already (LINUXicc)? I was planning to just modify > LINUXgcc. > > I wasn't able to check out the sources, though: > > % cvs -d :pserver:anoncvs at pooma.codesourcery.com:/home/pooma/Repository > login > Logging in to > :pserver:anoncvs at pooma.codesourcery.com:2401/home/pooma/Repository > CVS password: > cvs [login aborted]: connect to pooma.codesourcery.com(65.74.133.5):2401 > failed: Connection timed out > > If you have other ideas that I should try, please let me know. This morning, I did not have any difficulty using the directions at the bottom of http://www.codesourcery.com/pooma/pooma_download. The password is 'anoncvs'. If you still have problems, please send as much information as possible. I believe cvs's '-v' option produces more information, which can be useful for debugging. -- Jeffrey D. Oldham oldham at codesourcery.com From rguenth at tat.physik.uni-tuebingen.de Fri Feb 27 14:58:49 2004 From: rguenth at tat.physik.uni-tuebingen.de (Richard Guenther) Date: Fri, 27 Feb 2004 15:58:49 +0100 Subject: AltView* classes Message-ID: <403F5B29.6040005@tat.physik.uni-tuebingen.de> Hi! Does anyone remember the exact reason for the existance of the AltView* classes in Array.h and Field.h? Removing them and fixing their usage to use View* seems to work (full test with gcc 3.3 and 3.4 in progress). Thanks, Richard. From jeff.sullivan at intel.com Fri Feb 27 19:29:50 2004 From: jeff.sullivan at intel.com (Sullivan, Jeff) Date: Fri, 27 Feb 2004 14:29:50 -0500 Subject: [pooma-dev] Problem with configure on Linux/ia32: Red Hat 8.0 Message-ID: <88E10B707C51824DA13AA135A402DF470109C356@hdsmsx403.hd.intel.com> Thanks again for your help, but I am still having trouble checking out the cvs sources. I also was unable to ping pooma.codesourcery.com (from Linux). % ping pooma.codesourcery.com PING pooma.codesourcery.com (65.74.133.5) from .... : 56(84) bytes of data. --- pooma.codesourcery.com ping statistics --- 4 packets transmitted, 0 received, 100% loss, time 3017ms % /usr/bin/cvs --version Concurrent Versions System (CVS) 1.11.2 (client/server) % cvs -d :pserver:anoncvs at pooma.codesourcery.com:/home/pooma/Repository login Logging in to :pserver:anoncvs at pooma.codesourcery.com:2401/home/pooma/Repository CVS password: cvs [login aborted]: connect to pooma.codesourcery.com(65.74.133.5):2401 failed: Connection timed out Could this be due to my corporate firewall or other security settings? Is that the right IP address for the cvs server? Thanks, -Jeff From oldham at codesourcery.com Fri Feb 27 21:12:00 2004 From: oldham at codesourcery.com (Jeffrey D. Oldham) Date: Fri, 27 Feb 2004 13:12:00 -0800 Subject: [pooma-dev] Problem with configure on Linux/ia32: Red Hat 8.0 In-Reply-To: <88E10B707C51824DA13AA135A402DF470109C356@hdsmsx403.hd.intel.com> References: <88E10B707C51824DA13AA135A402DF470109C356@hdsmsx403.hd.intel.com> Message-ID: <403FB2A0.1050603@codesourcery.com> Sullivan, Jeff wrote: > Thanks again for your help, but I am still having trouble checking out > the cvs sources. I also was unable to ping pooma.codesourcery.com (from > Linux). > > % ping pooma.codesourcery.com > PING pooma.codesourcery.com (65.74.133.5) from .... : 56(84) bytes of > data. > > --- pooma.codesourcery.com ping statistics --- > 4 packets transmitted, 0 received, 100% loss, time 3017ms > > % /usr/bin/cvs --version > Concurrent Versions System (CVS) 1.11.2 (client/server) > > % cvs -d :pserver:anoncvs at pooma.codesourcery.com:/home/pooma/Repository > login > Logging in to > :pserver:anoncvs at pooma.codesourcery.com:2401/home/pooma/Repository > CVS password: > cvs [login aborted]: connect to pooma.codesourcery.com(65.74.133.5):2401 > failed: Connection timed out > > Could this be due to my corporate firewall or other security settings? > Is that the right IP address for the cvs server? Pinging works for me from inside and outside codesourcery.com: (outside codesourcery.com) ping pooma.codesourcery.com PING pooma.codesourcery.com (65.74.133.5) from 171.64.66.201 : 56(84) bytes of data. 64 bytes from voldemort.codesourcery.com (65.74.133.5): icmp_seq=0 ttl=243 time=11.402 msec 64 bytes from voldemort.codesourcery.com (65.74.133.5): icmp_seq=1 ttl=243 time=8.084 msec 64 bytes from voldemort.codesourcery.com (65.74.133.5): icmp_seq=2 ttl=243 time=7.647 msec 64 bytes from voldemort.codesourcery.com (65.74.133.5): icmp_seq=3 ttl=243 time=8.738 msec --- pooma.codesourcery.com ping statistics --- 4 packets transmitted, 4 packets received, 0% packet loss round-trip min/avg/max/mdev = 7.647/8.967/11.402/1.462 ms Inside codesourcery.com yields similiar data. -- Jeffrey D. Oldham oldham at codesourcery.com From ron_hylton at hotmail.com Fri Feb 27 23:05:41 2004 From: ron_hylton at hotmail.com (ron hylton) Date: Fri, 27 Feb 2004 18:05:41 -0500 Subject: [pooma-dev] AltView* classes Message-ID: > >Does anyone remember the exact reason for the existance of the AltView* >classes in Array.h and Field.h? Removing them and fixing their usage to >use View* seems to work (full test with gcc 3.3 and 3.4 in progress). > Here'a another question for anyone familiar with the development history here. Array<3> a(5,5,5); a(LeftDomain<3>(1,2,3)) fails because LeftDomain (& correspondingly RightDomain) doesn't define NewDomain1_t. Defining NewDomain1_t and specializing TemporaryNewDomain1 for Left/RightDomain as is done for AllDomain fixes the compiler error, but I was concerned that there is some subtle reason why this wasn't done. Unfortunately I'm having a very annoying problem with the Intel Windows compiler refusing to let me debug the change, it tells me no source code is available whenever I try to step into the view operation or set a breakpoint there. Ron Hylton _________________________________________________________________ Get fast, reliable access with MSN 9 Dial-up. Click here for Special Offer! http://click.atdmt.com/AVE/go/onm00200361ave/direct/01/