[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
Saturday morning update: SUCCESS! Installing the 64-bit Ubuntu did the trick. Now if only I can get my scanner to work ... Sage is a truly great community. Thanks. -Bruce On Apr 16, 10:16 am, Bruce Cohen math.co...@gmail.com wrote: A status report: Last yesterday's build died with ATLAS -- this time on a truly quiet machine. I tried again with a make clean export SAGE_FAT_BINARY=yes make but that also died with ATLAS. -Bruce On Apr 15, 10:40 am, Robert Bradshaw rober...@math.washington.edu wrote: On Apr 15, 2010, at 9:28 AM, Bruce Cohen wrote: I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. That would probably do it. It's probably not a question of being overloaded so much as being inconsistently loaded, making the ATLAS timing/tuning parameter graphs to erratic to optimize against. The suggestions on this thread are good if you're not able to get it going tonight. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group athttp://groups.google.com/group/sage-devel URL:http://www.sagemath.org -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
A status report: Last yesterday's build died with ATLAS -- this time on a truly quiet machine. I tried again with a make clean export SAGE_FAT_BINARY=yes make but that also died with ATLAS. -Bruce On Apr 15, 10:40 am, Robert Bradshaw rober...@math.washington.edu wrote: On Apr 15, 2010, at 9:28 AM, Bruce Cohen wrote: I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. That would probably do it. It's probably not a question of being overloaded so much as being inconsistently loaded, making the ATLAS timing/tuning parameter graphs to erratic to optimize against. The suggestions on this thread are good if you're not able to get it going tonight. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On 04/16/2010 12:16 PM, Bruce Cohen wrote: A status report: Last yesterday's build died with ATLAS -- this time on a truly quiet machine. I tried again with a make clean export SAGE_FAT_BINARY=yes make but that also died with ATLAS. -Bruce Here are two other things to try: 1. Download the latest ATLAS (http://sourceforge.net/projects/math-atlas/) and try building it. Does it succeed? Note to everyone: the newest version of atlas is 3.9.23, while our version is 3.8.3 (p12). On the surface, it looks like our ATLAS is way out of date. Maybe we should upgrade anyway? Does anyone know of problems with upgrading? 2. Try just installing the ATLAS spkg, using sage -i -s ATLAS SPKG FILE. This will leave your build directory there. You'll notice in the build messages you posted that there is an error file created: tar cf error_UNKNOWNx8632SSE3.tar Make.inc bin/INSTALL_LOG/* gzip --best error_UNKNOWNx8632SSE3.tar mv error_UNKNOWNx8632SSE3.tar.gz error_UNKNOWNx8632SSE3.tgz make[6]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' make[5]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' make[4]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build/bin' Error report error_ARCH.tgz has been created in your top-level ATLAS directory. Be sure to include this file in any help request. Doing sage -i -s should leave this error file around, and it might illuminate the problem better. What kind of i7 do you have? I have a quad core i7 920 that I compile Sage on just fine. I'm running 64-bit Ubuntu 9.10, though. Thanks, Jason -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On 04/16/2010 12:50 PM, Jason Grout wrote: Note to everyone: the newest version of atlas is 3.9.23, while our version is 3.8.3 (p12). On the surface, it looks like our ATLAS is way out of date. Maybe we should upgrade anyway? Does anyone know of problems with upgrading? Never mind. I see now that the latest *stable* release is 3.8.3, so apparently we are up to date. The unstable release is 3.9.23. Thanks, Jason -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On Apr 16, 10:50 am, Jason Grout jason-s...@creativetrax.com wrote: What kind of i7 do you have? I have a quad core i7 920 that I compile Sage on just fine. I'm running 64-bit Ubuntu 9.10, though. I have a quad core i7 860. I had been hesitant to try 64-bit because of problems with Java when I last tried 64 bit in 8.04, but I will try this over the weekend. Thanks -Bruce -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On 04/16/2010 01:14 PM, Bruce Cohen wrote: On Apr 16, 10:50 am, Jason Groutjason-s...@creativetrax.com wrote: What kind of i7 do you have? I have a quad core i7 920 that I compile Sage on just fine. I'm running 64-bit Ubuntu 9.10, though. I have a quad core i7 860. I had been hesitant to try 64-bit because of problems with Java when I last tried 64 bit in 8.04, but I will try this over the weekend. FYI, jmol and 3d plots seem to work just fine for me, if that's what you're worried about. I still have problems with flash videos, though. One of these days, I'll figure that out. Thanks, Jason -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
The job ran overnight and nothing else was going on. Perhaps I should try again tonight. -Bruce On Apr 14, 11:18 pm, Robert Bradshaw rober...@math.washington.edu wrote: On Apr 14, 2010, at 11:12 PM, Bruce Cohen wrote: I have a new machine with Ubuntu 9.10 (32 bit). My first build of Sage (4.3.5) did not work. Here is a 100 line (maybe too many lines, maybe too little) of install.log. Thanks for your help. -Bruce How loaded is your machine? make[3]: *** [build] Error 255 make[3]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' make[2]: *** [build] Error 2 make[2]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' Failed to build ATLAS. ATLAS failed to build for the 1-th time, possibly because of a loaded system, so we will automatically try again up to 4 more times. Waiting 9 minutes... ATLAS failed to build for the 2-th time, possibly because of a loaded system, so we will automatically try again up to 3 more times. Waiting 10 minutes... ATLAS failed to build for the 3-th time, possibly because of a loaded system, so we will automatically try again up to 2 more times. Waiting 10 minutes... ATLAS failed to build for the 4-th time, possibly because of a loaded system, so we will automatically try again up to 1 more times. Waiting 7 minutes... Too many failures to build ATLAS. Giving up! ATLAS seems to be having issues getting reliable timings for its tuning. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org To unsubscribe, reply using remove me as the subject.
[sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. -Bruce On Apr 15, 7:51 am, Bruce Cohen math.co...@gmail.com wrote: The job ran overnight and nothing else was going on. Perhaps I should try again tonight. -Bruce On Apr 14, 11:18 pm, Robert Bradshaw rober...@math.washington.edu wrote: On Apr 14, 2010, at 11:12 PM, Bruce Cohen wrote: I have a new machine with Ubuntu 9.10 (32 bit). My first build of Sage (4.3.5) did not work. Here is a 100 line (maybe too many lines, maybe too little) of install.log. Thanks for your help. -Bruce How loaded is your machine? make[3]: *** [build] Error 255 make[3]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' make[2]: *** [build] Error 2 make[2]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' Failed to build ATLAS. ATLAS failed to build for the 1-th time, possibly because of a loaded system, so we will automatically try again up to 4 more times. Waiting 9 minutes... ATLAS failed to build for the 2-th time, possibly because of a loaded system, so we will automatically try again up to 3 more times. Waiting 10 minutes... ATLAS failed to build for the 3-th time, possibly because of a loaded system, so we will automatically try again up to 2 more times. Waiting 10 minutes... ATLAS failed to build for the 4-th time, possibly because of a loaded system, so we will automatically try again up to 1 more times. Waiting 7 minutes... Too many failures to build ATLAS. Giving up! ATLAS seems to be having issues getting reliable timings for its tuning. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org To unsubscribe, reply using remove me as the subject.
Re: [sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
Atlas is supposed to find out what sort of processor your machine has automatically and then use tuning data for that. But it is not perfect at detecting the processor type, so then it goess of to do this tuning stuff which takes hours. The same happened to me when I got a new laptop in 2008. Until Sage upgraded Atlas the only was I found to stop this (very) time-consuming step happening for every Sage build (often unsuccessfully) was to keep the previous successful build. Once you have a good build of Atlas, take a copy of SAGE_ROOT/local; and set SAGE_ATLAS to point to that copied directory. then when you build Sage it will use that copy of Atlas instead of rebuilding it. John On 15 April 2010 17:28, Bruce Cohen math.co...@gmail.com wrote: I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. -Bruce On Apr 15, 7:51 am, Bruce Cohen math.co...@gmail.com wrote: The job ran overnight and nothing else was going on. Perhaps I should try again tonight. -Bruce On Apr 14, 11:18 pm, Robert Bradshaw rober...@math.washington.edu wrote: On Apr 14, 2010, at 11:12 PM, Bruce Cohen wrote: I have a new machine with Ubuntu 9.10 (32 bit). My first build of Sage (4.3.5) did not work. Here is a 100 line (maybe too many lines, maybe too little) of install.log. Thanks for your help. -Bruce How loaded is your machine? make[3]: *** [build] Error 255 make[3]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' make[2]: *** [build] Error 2 make[2]: Leaving directory `/home/bic/src/sage-4.3.5/spkg/build/ atlas-3.8.3.p12/ATLAS-build' Failed to build ATLAS. ATLAS failed to build for the 1-th time, possibly because of a loaded system, so we will automatically try again up to 4 more times. Waiting 9 minutes... ATLAS failed to build for the 2-th time, possibly because of a loaded system, so we will automatically try again up to 3 more times. Waiting 10 minutes... ATLAS failed to build for the 3-th time, possibly because of a loaded system, so we will automatically try again up to 2 more times. Waiting 10 minutes... ATLAS failed to build for the 4-th time, possibly because of a loaded system, so we will automatically try again up to 1 more times. Waiting 7 minutes... Too many failures to build ATLAS. Giving up! ATLAS seems to be having issues getting reliable timings for its tuning. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org To unsubscribe, reply using remove me as the subject. -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
Re: [sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On Thu, Apr 15, 2010 at 9:37 AM, John Cremona john.crem...@gmail.com wrote: Atlas is supposed to find out what sort of processor your machine has automatically and then use tuning data for that. But it is not perfect at detecting the processor type, so then it goess of to do this tuning stuff which takes hours. The same happened to me when I got a new laptop in 2008. Until Sage upgraded Atlas the only was I found to stop this (very) time-consuming step happening for every Sage build (often unsuccessfully) was to keep the previous successful build. Once you have a good build of Atlas, take a copy of SAGE_ROOT/local; and set SAGE_ATLAS to point to that copied directory. then when you build Sage it will use that copy of Atlas instead of rebuilding it. John On 15 April 2010 17:28, Bruce Cohen math.co...@gmail.com wrote: I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. If that doesn't work, you might try building after first typing export SAGE_FAT_BINARY=yes This will result in a Sage that is non-optimal at numerical linear algebra computations. But ATLAS will I think be much more likely to build, since it just uses some minimal default configuration. William -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org To unsubscribe, reply using remove me as the subject.
Re: [sage-devel] Re: Build did not work 4.3.5 Ubuntu 9.10 (atlas-3.8.3.p12)
On Apr 15, 2010, at 9:28 AM, Bruce Cohen wrote: I now remember that I did have a 5 hour scp file transfer running that night. It is a brand new machine, so I was bringing over a tar file from an old machine on my network. I will try again tonight. It is disappointing that that the new quad core i7 (purchased to do Sage work) could not handle two tasks. We'll see. That would probably do it. It's probably not a question of being overloaded so much as being inconsistently loaded, making the ATLAS timing/tuning parameter graphs to erratic to optimize against. The suggestions on this thread are good if you're not able to get it going tonight. - Robert -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org To unsubscribe, reply using remove me as the subject.