[sage-devel] Re: compiling on AIX/PowerPPC
Oops, I am in the middle of giving the gcc-4.2.1 package a try. I hope I am using it correctly - I installed it, and then removed everything from spkg/installed, and now I am recompiliing by using make. Will that rebuild with gcc-4.2.1? It looks like it from what I can see during compilation. If this doesn't help at all, I will try your suggestion of a local binutils. Thanks for the help, Marshall On Sep 10, 11:10 pm, mabshoff [EMAIL PROTECTED] dortmund.de wrote: On Sep 10, 8:36 pm, Hamptonio [EMAIL PROTECTED] wrote: cat /etc/issue gives: Hello Marshall, Welcome to SUSE LINUX Enterprise Server 9 (x86_64) - Kernel \r (\l). The tail end of the polymake build, when it crashes, has the following: make[2]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/modules/graph' g++ -I/home/bc1/hamptonm/sage-2.8.3.6/local/lib -I/home/bc1/ hamptonm/sage-2.8.3.6/local/lib -o cayley_embedding cayley_embedding.o libpolytope.a ../../lib/libpoly.a -lgmp collect2: ld terminated with signal 11 [Segmentation fault] Okay, this is not a gcc issue (at least the output makes it very unlikely), but that the linker shoots itself. Check with the sysadmin if there is a more current ld somewhere in the systems, otherwise compile your own binutils locally and make sure that they are in $PATH before the systems ld. Try again with that ld and let us know if the problem goes away. You should also make sure the polymake doesn't hard code the path to ld somewhere in the makefile in case the error doesn't go away with the new binutils. make[2]: *** [cayley_embedding] Error 1 make[1]: *** [do_all] Error 2 make[1]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/apps/polytope' make: *** [all] Error 2 cp: cannot create regular file `/home/bc1/hamptonm/sage-2.8.3.6/local/ polymake/bin/polymake': No such file or directory Do not worry if there is an error message above, as long as the build says it worked below. Error building and installing polymake real6m1.033s user5m19.608s sys 0m31.232s Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On Sep 11, 2:07 pm, Hamptonio [EMAIL PROTECTED] wrote: Oops, I am in the middle of giving the gcc-4.2.1 package a try. I hope I am using it correctly - I installed it, and then removed everything from spkg/installed, and now I am recompiliing by using make. Will that rebuild with gcc-4.2.1? It looks like it from what I can see during compilation. Yep. Assuming you have source sage-env 'which g++' should point to the one in local/bin. But it might be needed to modify local/bin/sage-env slightly and include lib64 in LD_LIBRARY_PATH like the following LD_LIBRARY_PATH=$SAGE_ROOT/local/lib/64:$SAGE_ROOT/local/lib/: $LD_LIBRARY_PATH export LD_LIBRARY_PATH to pick up the new libstdc++.so before the system's one. The problem is due to Linux on Opterons having a 32 and 64 bit userspace and gcc installing the 64 bit libraries into lib64. If this doesn't help at all, I will try your suggestion of a local binutils. Well, that certainly is somewhat of a last resort, but if the box has gcc 3.3 the binutils ought to be also fairly ancient. Thanks for the help, Marshall SNIP Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On Sep 11, 3:09 pm, mabshoff [EMAIL PROTECTED] dortmund.de wrote: On Sep 11, 2:07 pm, Hamptonio [EMAIL PROTECTED] wrote: Sorry, but this is slightly wrong but causes bad consequences: LD_LIBRARY_PATH=$SAGE_ROOT/local/lib/64:$SAGE_ROOT/local/lib/: $LD_LIBRARY_PATH export LD_LIBRARY_PATH It needs to be: LD_LIBRARY_PATH=$SAGE_ROOT/local/lib64:$SAGE_ROOT/local/lib/: $LD_LIBRARY_PATH export LD_LIBRARY_PATH to pick up the new libstdc++.so before the system's one. The problem is due to Linux on Opterons having a 32 and 64 bit userspace and gcc installing the 64 bit libraries into lib64. Sorry for the noise. Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
cat /etc/issue gives: Welcome to SUSE LINUX Enterprise Server 9 (x86_64) - Kernel \r (\l). The tail end of the polymake build, when it crashes, has the following: make[2]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/modules/graph' g++ -I/home/bc1/hamptonm/sage-2.8.3.6/local/lib -I/home/bc1/ hamptonm/sage-2.8.3.6/local/lib -o cayley_embedding cayley_embedding.o libpolytope.a ../../lib/libpoly.a -lgmp collect2: ld terminated with signal 11 [Segmentation fault] make[2]: *** [cayley_embedding] Error 1 make[1]: *** [do_all] Error 2 make[1]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/apps/polytope' make: *** [all] Error 2 cp: cannot create regular file `/home/bc1/hamptonm/sage-2.8.3.6/local/ polymake/bin/polymake': No such file or directory Do not worry if there is an error message above, as long as the build says it worked below. Error building and installing polymake real6m1.033s user5m19.608s sys 0m31.232s On Sep 9, 10:35 pm, William Stein [EMAIL PROTECTED] wrote: On 9/9/07, Hamptonio [EMAIL PROTECTED] wrote: From the MSI webpage, it says the bladecenter is: ...a Linux Cluster from IBM. It is a IBM BladeCenter H with 307 LS 21 nodes. Each node has two dual-core 2.6 GHz AMD Opteron processors sharing 8 GB of memory. The interactive node that I can read /proc/cpuinfo on says there are 4 cores, each with: vendor_id : AuthenticAMD cpu family : 15 model : 65 model name : Dual-Core AMD Opteron(tm) Processor 2218 stepping: 2 cpu MHz : 2600.209 cache size : 1024 KB uname -a: Linux blade288 2.6.5-7.244-smp #1 SMP Mon Dec 12 18:32:25 UTC 2005 x86_64 x86_64 x86_64 GNU/Linux Which linux distribution is it? Try cat /etc/issue It's probably SUSE, since that is listed in the gcc output below. Also, regarding polymake, maybe you should report what the error is when it builds? William Jobs are scheduled with something called PBS. Polymake fails to compile, which is bad news for me - although I haven't tried to install polymake on a fresh 2.8.4 install so it might be a more pervasive problem. One problem is that it looks like an older gcc, is this a problem(?): GCC Version gcc -v Reading specs from /usr/lib64/gcc-lib/x86_64-suse-linux/3.3.3/specs Configured with: ../configure --enable-threads=posix --prefix=/usr -- with-local-prefix=/usr/local -- infodir=/usr/share/info --mandir=/usr/share/man --enable-languages=c,c+ +,f77,objc,java,ada --disable -checking --libdir=/usr/lib64 --enable-libgcj --with-gxx-include-dir=/ usr/include/g++ --with-slibdir =/lib64 --with-system-zlib --enable-shared --enable-__cxa_atexit x86_64-suse-linux Thread model: posix gcc version 3.3.3 (SuSE Linux) It would be very cool if I could use this machine for polytope/ algebraic geometry/groebner basis stuff. At the moment I am using it for pretty routine bioinformatics stuff that I can do without sage if necessary. -Marshall -- William Stein Associate Professor of Mathematics University of Washingtonhttp://www.williamstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On Sep 10, 8:36 pm, Hamptonio [EMAIL PROTECTED] wrote: cat /etc/issue gives: Hello Marshall, Welcome to SUSE LINUX Enterprise Server 9 (x86_64) - Kernel \r (\l). The tail end of the polymake build, when it crashes, has the following: make[2]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/modules/graph' g++ -I/home/bc1/hamptonm/sage-2.8.3.6/local/lib -I/home/bc1/ hamptonm/sage-2.8.3.6/local/lib -o cayley_embedding cayley_embedding.o libpolytope.a ../../lib/libpoly.a -lgmp collect2: ld terminated with signal 11 [Segmentation fault] Okay, this is not a gcc issue (at least the output makes it very unlikely), but that the linker shoots itself. Check with the sysadmin if there is a more current ld somewhere in the systems, otherwise compile your own binutils locally and make sure that they are in $PATH before the systems ld. Try again with that ld and let us know if the problem goes away. You should also make sure the polymake doesn't hard code the path to ld somewhere in the makefile in case the error doesn't go away with the new binutils. make[2]: *** [cayley_embedding] Error 1 make[1]: *** [do_all] Error 2 make[1]: Leaving directory `/home/bc1/hamptonm/sage-2.8.3.6/spkg/build/ polymake-2.2.p2/build/apps/polytope' make: *** [all] Error 2 cp: cannot create regular file `/home/bc1/hamptonm/sage-2.8.3.6/local/ polymake/bin/polymake': No such file or directory Do not worry if there is an error message above, as long as the build says it worked below. Error building and installing polymake real6m1.033s user5m19.608s sys 0m31.232s Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
I didn't really try any further with the AIX compilation after starting on the linux bladecenter. I could try a bit if it is of interest. On the BladeCenter, only one test failed after upgrading to 2.8.4: sage -t devel/sage-main/sage/rings/real_rqdf.pyx ** File real_rqdf.pyx, line 24: sage: RQDF( 123.2) * RR (.543) Expected: 66.897600624851281827432114309792736749325567465385058291 Got: NaN ** 1 items had failures: 1 of 9 in __main__.example_0 ***Test Failed*** 1 failures. For whitespace errors, see the file .doctest_real_rqdf.pyx On Sep 8, 10:24 am, mabshoff [EMAIL PROTECTED] dortmund.de wrote: On Sep 8, 4:33 pm, Hamptonio [EMAIL PROTECTED] wrote: It turns out that I also have access to a linux-based supercomputer, (an IBM BladeCenter with about 300 quad-processor nodes), which should be faster than the Power4 system anyway. So I will give up, for the moment, trying to install sage on AIX. Ok, how far did you get? I have installed sage-2.8.3.6 on the BladeCenter successfully, with the following make test failures: sage -t devel/sage-main/sage/modular/modform/ambient_g1.py sage -t devel/sage-main/sage/modular/modform/ eisenstein_submodule.py sage -t devel/sage-main/sage/modular/modform/submodule.py sage -t devel/sage-main/sage/rings/real_rqdf.pyx Total time for all tests: 2137.4 seconds Probably some of those have been fixed in 2.8.4. Please do an ./sage -upgrade and then rerun the testsuite. If any of those persist we should fix them. There was a last minute 32 bit only regression at the very end of 2.8.4. It was so late that William respun the tarball ;) Now I'll have to learn more about DSage; I am not sure how to exploit all those processors the way I usually code. Cheers, Marshall Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On 9/9/07, Hamptonio [EMAIL PROTECTED] wrote: On the BladeCenter, only one test failed after upgrading to 2.8.4: sage -t devel/sage-main/sage/rings/real_rqdf.pyx ** File real_rqdf.pyx, line 24: sage: RQDF( 123.2) * RR (.543) Expected: 66.897600624851281827432114309792736749325567465385058291 Got: NaN ** 1 items had failures: 1 of 9 in __main__.example_0 ***Test Failed*** 1 failures. For whitespace errors, see the file .doctest_real_rqdf.pyx What is the BladeCenter hardware-wise exactly? Failure of the above test is very weird/curious. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
From the MSI webpage, it says the bladecenter is: ...a Linux Cluster from IBM. It is a IBM BladeCenter H with 307 LS 21 nodes. Each node has two dual-core 2.6 GHz AMD Opteron processors sharing 8 GB of memory. The interactive node that I can read /proc/cpuinfo on says there are 4 cores, each with: vendor_id : AuthenticAMD cpu family : 15 model : 65 model name : Dual-Core AMD Opteron(tm) Processor 2218 stepping: 2 cpu MHz : 2600.209 cache size : 1024 KB uname -a: Linux blade288 2.6.5-7.244-smp #1 SMP Mon Dec 12 18:32:25 UTC 2005 x86_64 x86_64 x86_64 GNU/Linux Jobs are scheduled with something called PBS. Polymake fails to compile, which is bad news for me - although I haven't tried to install polymake on a fresh 2.8.4 install so it might be a more pervasive problem. One problem is that it looks like an older gcc, is this a problem(?): GCC Version gcc -v Reading specs from /usr/lib64/gcc-lib/x86_64-suse-linux/3.3.3/specs Configured with: ../configure --enable-threads=posix --prefix=/usr -- with-local-prefix=/usr/local -- infodir=/usr/share/info --mandir=/usr/share/man --enable-languages=c,c+ +,f77,objc,java,ada --disable -checking --libdir=/usr/lib64 --enable-libgcj --with-gxx-include-dir=/ usr/include/g++ --with-slibdir =/lib64 --with-system-zlib --enable-shared --enable-__cxa_atexit x86_64-suse-linux Thread model: posix gcc version 3.3.3 (SuSE Linux) It would be very cool if I could use this machine for polytope/ algebraic geometry/groebner basis stuff. At the moment I am using it for pretty routine bioinformatics stuff that I can do without sage if necessary. -Marshall On Sep 9, 10:19 am, William Stein [EMAIL PROTECTED] wrote: On 9/9/07, Hamptonio [EMAIL PROTECTED] wrote: On the BladeCenter, only one test failed after upgrading to 2.8.4: sage -t devel/sage-main/sage/rings/real_rqdf.pyx ** File real_rqdf.pyx, line 24: sage: RQDF( 123.2) * RR (.543) Expected: 66.897600624851281827432114309792736749325567465385058291 Got: NaN ** 1 items had failures: 1 of 9 in __main__.example_0 ***Test Failed*** 1 failures. For whitespace errors, see the file .doctest_real_rqdf.pyx What is the BladeCenter hardware-wise exactly? Failure of the above test is very weird/curious. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On 9/9/07, Hamptonio [EMAIL PROTECTED] wrote: From the MSI webpage, it says the bladecenter is: ...a Linux Cluster from IBM. It is a IBM BladeCenter H with 307 LS 21 nodes. Each node has two dual-core 2.6 GHz AMD Opteron processors sharing 8 GB of memory. The interactive node that I can read /proc/cpuinfo on says there are 4 cores, each with: vendor_id : AuthenticAMD cpu family : 15 model : 65 model name : Dual-Core AMD Opteron(tm) Processor 2218 stepping: 2 cpu MHz : 2600.209 cache size : 1024 KB uname -a: Linux blade288 2.6.5-7.244-smp #1 SMP Mon Dec 12 18:32:25 UTC 2005 x86_64 x86_64 x86_64 GNU/Linux Which linux distribution is it? Try cat /etc/issue It's probably SUSE, since that is listed in the gcc output below. Also, regarding polymake, maybe you should report what the error is when it builds? William Jobs are scheduled with something called PBS. Polymake fails to compile, which is bad news for me - although I haven't tried to install polymake on a fresh 2.8.4 install so it might be a more pervasive problem. One problem is that it looks like an older gcc, is this a problem(?): GCC Version gcc -v Reading specs from /usr/lib64/gcc-lib/x86_64-suse-linux/3.3.3/specs Configured with: ../configure --enable-threads=posix --prefix=/usr -- with-local-prefix=/usr/local -- infodir=/usr/share/info --mandir=/usr/share/man --enable-languages=c,c+ +,f77,objc,java,ada --disable -checking --libdir=/usr/lib64 --enable-libgcj --with-gxx-include-dir=/ usr/include/g++ --with-slibdir =/lib64 --with-system-zlib --enable-shared --enable-__cxa_atexit x86_64-suse-linux Thread model: posix gcc version 3.3.3 (SuSE Linux) It would be very cool if I could use this machine for polytope/ algebraic geometry/groebner basis stuff. At the moment I am using it for pretty routine bioinformatics stuff that I can do without sage if necessary. -Marshall -- William Stein Associate Professor of Mathematics University of Washington http://www.williamstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
Hello, On the BladeCenter, only one test failed after upgrading to 2.8.4: sage -t devel/sage-main/sage/rings/real_rqdf.pyx ** File real_rqdf.pyx, line 24: sage: RQDF( 123.2) * RR (.543) Expected: 66.897600624851281827432114309792736749325567465385058291 Got: NaN ** 1 items had failures: 1 of 9 in __main__.example_0 ***Test Failed*** 1 failures. For whitespace errors, see the file .doctest_real_rqdf.pyx In addition to what William said: - If you run make check with the mpfr spkg do all tests pass? - The compiler is definitely a little long in the tooth, The is an experimental gcc 4.2.1 spkg at sagemath.org. You could install that, wipe out spkg/installed and rebuild Sage from scratch. - What happens if you install a binary release of Sage? Does the problem go away? Chances are with that crufty an install unless some current libstdc++.so is around Sage 2.8.4[.1] won't run, especially on x86-64. You probably are better of with building the experimental gcc spkg. If you have any problems with that let me know. Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
It turns out that I also have access to a linux-based supercomputer, (an IBM BladeCenter with about 300 quad-processor nodes), which should be faster than the Power4 system anyway. So I will give up, for the moment, trying to install sage on AIX. I have installed sage-2.8.3.6 on the BladeCenter successfully, with the following make test failures: sage -t devel/sage-main/sage/modular/modform/ambient_g1.py sage -t devel/sage-main/sage/modular/modform/ eisenstein_submodule.py sage -t devel/sage-main/sage/modular/modform/submodule.py sage -t devel/sage-main/sage/rings/real_rqdf.pyx Total time for all tests: 2137.4 seconds Probably some of those have been fixed in 2.8.4. Now I'll have to learn more about DSage; I am not sure how to exploit all those processors the way I usually code. Cheers, Marshall On Sep 7, 1:50 pm, William Stein [EMAIL PROTECTED] wrote: On 9/7/07, Hamptonio [EMAIL PROTECTED] wrote: Thanks guys. I can't use the notebook anyway from that machine, I would be running pre-written scripts to do heavy calculations. It got far enough that I am hopeful I can get it to where it does what I need. Please post anything you figure out. We would be very happy to have AIX officially supported one day!! William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On 9/7/07, Hamptonio [EMAIL PROTECTED] wrote: I am hoping to use sage on some supercomputers and/or clusters. Currently I am trying to compile it on the IBM Power4 machine at the Minnesota Supercomputing Center (see http://www.msi.umn.edu/power4/index.html for some details on that). Its about 300 Power4 processors running AIX. Porting SAGE to a new architecture / operating system is a potentially difficult project, that should not be underestimated. Nobody has ever compiled SAGE (or probably many of its components) under AIX, so I predict based on experience that the number of problems that one would have will be at least enough to keep 2 or 3 very very smart people occupied for several weeks, and maybe more. In short, SAGE isn't supported on AIX or Power4, and isn't likely to be unless several SAGE developers had access to such a machine -- then it probably would get supported there. I of course really really wish SAGE were supported on a wider range of machines, especially those used in supercomputing (!) I just tried compiliing sage-2.8.3.6 on it, and it started off OK but failed while compiling libgcrypt. It seems this is often a difficult package, since it looks like a sticking point on Solaris and cygwin as well. Is there a way to build around that? Would it help to try an earlier sage version? If a package called foo.spkg won't build, if you do touch spkg/installed/foo then type make the build will continue as if the package did build successfully. One can fully build SAGE without libgcrypt or any of the gnutls stuff. It will work -- the only thing that might fail is the notebook, but then only in secure mode. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On Sep 7, 7:00 pm, Hamptonio [EMAIL PROTECTED] wrote: Hi, I am hoping to use sage on some supercomputers and/or clusters. Currently I am trying to compile it on the IBM Power4 machine at the Minnesota Supercomputing Center (seehttp://www.msi.umn.edu/power4/index.html for some details on that). Its about 300 Power4 processors running AIX. I just tried compiliing sage-2.8.3.6 on it, and it started off OK but failed while compiling libgcrypt. It seems this is often a difficult package, since it looks like a sticking point on Solaris and cygwin as well. Is there a way to build around that? Would it help to try an earlier sage version? Marshall Hampton Hey Marshall, I believe libgcrypt is only the openssl replacement. So if you have an openssl on that box you can probably ignore it. Do a touch in spkg/ installed/name_of_spkg to skip over it. I also believe to recall that all the SLL stuff is only needed for the notebook, so I am fairly certain that you won't run your computations that way. I am not sure if AIX these days is pure 64 bit or a hyprid 32/64 userspace. If it is pure 64 bit you shouldn't run into too much trouble. Let me know if you have any problems. I anybody has access to an AIX box for me I would be willing to get Sage to work on AIX, too - I just love UNIX :) Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
Thanks guys. I can't use the notebook anyway from that machine, I would be running pre-written scripts to do heavy calculations. It got far enough that I am hopeful I can get it to where it does what I need. Cheers, Marshall On Sep 7, 11:10 am, mabshoff [EMAIL PROTECTED] dortmund.de wrote: On Sep 7, 7:00 pm, Hamptonio [EMAIL PROTECTED] wrote: Hi, I am hoping to use sage on some supercomputers and/or clusters. Currently I am trying to compile it on the IBM Power4 machine at the Minnesota Supercomputing Center (seehttp://www.msi.umn.edu/power4/index.html for some details on that). Its about 300 Power4 processors running AIX. I just tried compiliing sage-2.8.3.6 on it, and it started off OK but failed while compiling libgcrypt. It seems this is often a difficult package, since it looks like a sticking point on Solaris and cygwin as well. Is there a way to build around that? Would it help to try an earlier sage version? Marshall Hampton Hey Marshall, I believe libgcrypt is only the openssl replacement. So if you have an openssl on that box you can probably ignore it. Do a touch in spkg/ installed/name_of_spkg to skip over it. I also believe to recall that all the SLL stuff is only needed for the notebook, so I am fairly certain that you won't run your computations that way. I am not sure if AIX these days is pure 64 bit or a hyprid 32/64 userspace. If it is pure 64 bit you shouldn't run into too much trouble. Let me know if you have any problems. I anybody has access to an AIX box for me I would be willing to get Sage to work on AIX, too - I just love UNIX :) Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On Sep 7, 8:04 pm, Hamptonio [EMAIL PROTECTED] wrote: Thanks guys. I can't use the notebook anyway from that machine, I would be running pre-written scripts to do heavy calculations. It got far enough that I am hopeful I can get it to where it does what I need. Cheers, Marshall Cool, in case you need SSL (for DSAGE) you can install the optional openssl package from http://www.sagemath.org/packages/optional/openssl-0.9.8d.p1.spkg Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: compiling on AIX/PowerPPC
On 9/7/07, Hamptonio [EMAIL PROTECTED] wrote: Thanks guys. I can't use the notebook anyway from that machine, I would be running pre-written scripts to do heavy calculations. It got far enough that I am hopeful I can get it to where it does what I need. Please post anything you figure out. We would be very happy to have AIX officially supported one day!! William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
