[sage-devel] Re: Track performance regressions
Thank you for working towards giving us extra tools to help us prevent regressions. It is definitely something we can put to good use. However, the total time it takes to test a file is somewhat of a bad metric as adding lots of (sometimes relatively computationally intensive) tests to a file will make it look like it had a regression, which is the case you mentioned for Lie algebras (e.g., see https://trac.sagemath.org/ticket/22883). Perhaps it would be better to normalize the differences based on the change in the number of doctests, so when doctests are added, it is less likely to show a large increase. For me, the x-axis marks on the graphs are obscured by the top bar, so it makes it harder to understand what the commits are referring to. Best, Travis On Saturday, August 18, 2018 at 12:47:29 AM UTC+10, Julian Rüth wrote: > > Hello. > > At #25262 [1] I am trying to use airspeed velocity (asv) [2] to track > performance regressions in Sage. In a first metaclass hack, I had > asv benchmark all "long time" doctests in Sage for all the tags between > 8.2 and 8.3 (i.e., all the beta and rc versions of 8.3.) > > I ran this on my laptop computer over the past 48 or so hours. Since I > used it for quite some other stuff there is a lot of noise in the data > but you might still like to have a look at the result [3]. Strangely, > the 8.2 tag did not run so you can't really see 8.2 as a baseline > unfortunately. > > Btw., the tool detected serious regressions in > finite_dimensional_lie_algebras_with_basis and classical_lie_algebra. If > care about these you might want to double check what happened there. > > If you think it's a good idea to do this kind of regression tracking or > would like to see this being done differently, please join the > discussion at #25262 :) > > > julian > > [1] https://trac.sagemath.org/ticket/25262 > [2] https://github.com/airspeed-velocity/asv > [3] https://saraedum.github.io/sage/ > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] NTL 11.3.0
I just uploaded NTL 11.3.0 to https://www.shoup.net/ntl/ Brief summary of what's new: - An AVX-based small-prime FFT - For a variety of reasons, this is an experimental feature that is not enabled by default. - Works with both AVX2 and AVX512. - It can give a 2-3x speedup for the small-prime FFT itself, but it can actually slow other things down, like the CRT code. It's a work in progress. To really get an overall speedup, I have to come up with a way to make the CRT code also exploit AVX instructions. - AVX512 integration in the Mat algorithms (which already use AVX). - Fast GCD and XGCD algorithms for GF2EX, zz_pEX, and ZZ_pEX. Thanks to Luis Felipe Tabera Alonso for porting and testing. So now all of NTL's classes for polynomials over finite fields come equipped with a fast GCD. See https://www.shoup.net/ntl/doc/tour-changes.html for more details. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Arch: error building Sage, rpy2 fails
El viernes, 17 de agosto de 2018, 17:31:13 (UTC+2), Nicolás Sirolli escribió: > > Hello, > > I'm trying to build Sage 8.4.beta1 (following the develop branch from git) > on Arch Linux, and it fails when trying to build the rpy2-2.8.2.p0 package. > I'm attaching the corresponding log file. > > Please, let me know if any further information can be helpful to solve the > issue. > > You need to rebuild R against the latest ICU version. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Arch: error building Sage, rpy2 fails
Hello, I'm trying to build Sage 8.4.beta1 (following the develop branch from git) on Arch Linux, and it fails when trying to build the rpy2-2.8.2.p0 package. I'm attaching the corresponding log file. Please, let me know if any further information can be helpful to solve the issue. Thanks, Nico. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout. Found local metadata for rpy2-2.8.2.p0 Attempting to download package rpy2-2.8.2.tar.gz from mirrors https://mirror.koddos.net/sagemath/spkg/upstream/rpy2/rpy2-2.8.2.tar.gz [..] rpy2-2.8.2.p0 Setting up build directory for rpy2-2.8.2.p0 Finished extraction Applying patches from ../patches... Applying ../patches/cygwin.patch patching file rpy/rinterface/na_values.c patching file rpy/rinterface/_rinterface.c Hunk #1 succeeded at 1156 (offset -6 lines). Hunk #2 succeeded at 3750 (offset 1 line). Hunk #3 succeeded at 3760 (offset 1 line). Hunk #4 succeeded at 3770 (offset 1 line). Hunk #5 succeeded at 3780 (offset 1 line). patching file setup.py Hunk #1 succeeded at 200 (offset 40 lines). Host system: Linux pc-sirolli 4.17.2-1-ARCH #1 SMP PREEMPT Sat Jun 16 11:08:59 UTC 2018 x86_64 GNU/Linux C compiler: gcc C compiler version: Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-pc-linux-gnu/8.1.1/lto-wrapper Target: x86_64-pc-linux-gnu Configured with: /build/gcc/src/gcc/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/ --enable-languages=c,c++,ada,fortran,go,lto,objc,obj-c++ --enable-shared --enable-threads=posix --enable-libmpx --with-system-zlib --with-isl --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --disable-libssp --enable-gnu-unique-object --enable-linker-build-id --enable-lto --enable-plugin --enable-install-libiberty --with-linker-hash-style=gnu --enable-gnu-indirect-function --enable-multilib --disable-werror --enable-checking=release --enable-default-pie --enable-default-ssp Thread model: posix gcc version 8.1.1 20180531 (GCC) Installing rpy2-2.8.2.p0 Warning: This package has a badly-behaved setup.py which outputs more than the package name for 'setup.py --name'; using the last line as the package name: rpy2 Installing package rpy2 using pip2 Ignoring indexes: https://pypi.python.org/simple Processing /home/nico/appz/sage/local/var/tmp/sage/build/rpy2-2.8.2.p0/src Running setup.py (path:/tmp/pip-mXHN6p-build/setup.py) egg_info for package from file:///home/nico/appz/sage/local/var/tmp/sage/build/rpy2-2.8.2.p0/src Running command python setup.py egg_info R version 3.4.4 (2018-03-15) -- "Someone to Lean On" /home/nico/appz/sage/local/lib/R/bin/R CMD config --ldflags /home/nico/appz/sage/local/lib/R/bin/R CMD config --cppflags Compilation parameters for rpy2's C components: include_dirs= ['/home/nico/appz/sage/local/lib/R//include'] library_dirs= ['/home/nico/appz/sage/local/lib', '/home/nico/appz/sage/local/lib/R//lib'] libraries = ['readline', 'R', 'pcre', 'lzma', 'bz2', 'z', 'tirpc', 'rt', 'dl', 'm', 'icuuc', 'icui18n'] extra_link_args = ['-Wl,--export-dynamic', '-fopenmp', '-Wl,-rpath,/home/nico/appz/sage/local/lib'] running egg_info creating pip-egg-info/rpy2.egg-info writing requirements to pip-egg-info/rpy2.egg-info/requires.txt writing pip-egg-info/rpy2.egg-info/PKG-INFO writing top-level names to pip-egg-info/rpy2.egg-info/top_level.txt writing dependency_links to pip-egg-info/rpy2.egg-info/dependency_links.txt writing manifest file 'pip-egg-info/rpy2.egg-info/SOURCES.txt' R version 3.4.4 (2018-03-15) -- "Someone to Lean On" reading manifest file 'pip-egg-info/rpy2.egg-info/SOURCES.txt' reading manifest template 'MANIFEST.in' warning: no previously-included files matching '*patch*' found anywhere in distribution warning: no previously-included files matching '*diff*' found anywhere in distribution warning: no previously-included files matching '.hg' found anywhere in distribution warning: no files found matching 'MANIFEST' warning: no files found matching 'README' warning: no files found matching 'MPL_LICENSE' warning: no files found matching 'GPL_LICENSE'
Re: [sage-devel] Parallelization issue on macOS
Le vendredi 17 août 2018 15:26:37 UTC+2, Erik Bray a écrit : > > On Fri, Aug 17, 2018 at 2:02 PM Eric Gourgoulhon > wrote: > > > > Btw, I would be curious to know whether the notebook > > > http://nbviewer.jupyter.org/github/sagemanifolds/SageManifolds/blob/master/Worksheets/v1.3/SM_Kerr.ipynb > > > runs successfully with nproc>1 on Cygwin (I don't have any Windows > computer at hand either...) > > It works fine for me with nproc=8. Good to hear! Many thanks for having tried. > A few of the lines are > frustratingly slow though, such as calls like `nabla(g)` (I'm > surprised this isn't cached when called repeatedly, especially when > it's the connection on g itself).Also computing the Ricci tensor > felt very slow. Is that normal? > Yes that's normal: most of CPU time is actually spent, not in the computation itself, but in simplifications of symbolic expressions (via Maxima). On my Linux laptop (Intel Core i7-6700HQ), with nproc=8, nabla(g) takes 5 s and the computation of the Ricci tensor takes 27 s. Regarding caching, nabla(T), where T is any tensor field, is not cached, even if T=g. The cached operations are g.connection() g.ricci() g.riemann() Best regards, Eric. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Track performance regressions
Hello. At #25262 [1] I am trying to use airspeed velocity (asv) [2] to track performance regressions in Sage. In a first metaclass hack, I had asv benchmark all "long time" doctests in Sage for all the tags between 8.2 and 8.3 (i.e., all the beta and rc versions of 8.3.) I ran this on my laptop computer over the past 48 or so hours. Since I used it for quite some other stuff there is a lot of noise in the data but you might still like to have a look at the result [3]. Strangely, the 8.2 tag did not run so you can't really see 8.2 as a baseline unfortunately. Btw., the tool detected serious regressions in finite_dimensional_lie_algebras_with_basis and classical_lie_algebra. If care about these you might want to double check what happened there. If you think it's a good idea to do this kind of regression tracking or would like to see this being done differently, please join the discussion at #25262 :) julian [1] https://trac.sagemath.org/ticket/25262 [2] https://github.com/airspeed-velocity/asv [3] https://saraedum.github.io/sage/ -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout. signature.asc Description: PGP signature
Re: [sage-devel] Parallelization issue on macOS
On Fri, Aug 17, 2018 at 2:02 PM Eric Gourgoulhon wrote: > > Hi Erik, > > Le jeudi 16 août 2018 13:53:41 UTC+2, Erik Bray a écrit : >> >> >> Love how fork() is even more broken on OSX now than it is on Cygwin =_= > > > Btw, I would be curious to know whether the notebook > http://nbviewer.jupyter.org/github/sagemanifolds/SageManifolds/blob/master/Worksheets/v1.3/SM_Kerr.ipynb > runs successfully with nproc>1 on Cygwin (I don't have any Windows computer > at hand either...) It works fine for me with nproc=8. A few of the lines are frustratingly slow though, such as calls like `nabla(g)` (I'm surprised this isn't cached when called repeatedly, especially when it's the connection on g itself).Also computing the Ricci tensor felt very slow. Is that normal? -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Parallelization issue on macOS
Hi Erik, Le jeudi 16 août 2018 13:53:41 UTC+2, Erik Bray a écrit : > > Love how fork() is even more broken on OSX now than it is on Cygwin =_= > Btw, I would be curious to know whether the notebook http://nbviewer.jupyter.org/github/sagemanifolds/SageManifolds/blob/master/Worksheets/v1.3/SM_Kerr.ipynb runs successfully with nproc>1 on Cygwin (I don't have any Windows computer at hand either...) Best regards, Eric. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Parallelization issue on macOS
Hi Jeroen, Le jeudi 16 août 2018 10:27:37 UTC+2, Jeroen Demeyer a écrit : > > On 2018-08-15 17:46, Eric Gourgoulhon wrote: > > Various Sage users have reported on that parallelization of tensor field > > computations does not work on macOS. > > This might be https://trac.sagemath.org/ticket/25921 > > Does it help to set the environment variable > OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES > Thanks for this suggestion. I've just got the report from a macOS user: he sets this environment variable in his .bashrc file, but unfortunately, this does not help. Best regards, Eric. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] GNU/Linux debian sid/testing: "Error building Sage" 8.3 from git
Hi, Le 17/08/2018 à 11:17, Jörg-Volker a écrit : Another question I have after looking at the binary package of sage for debian 9 is: how was that built? Since in this package all packages using blas and lapack are linked against the system libraries libblas.so, liblapack.so, and libopenblas.so. Building sage isn't as simple as we'd like yet: https://wiki.debian.org/DebianScience/Sage jpuydt on irc.debian.org -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] GNU/Linux debian sid/testing: "Error building Sage" 8.3 from git
Hi, On Tuesday, August 14, 2018 at 5:09:05 PM UTC+2, Erik Bray wrote: > > On Tue, Aug 14, 2018 at 5:02 PM Jörg-Volker > wrote: > > > > > > > > On Tuesday, August 14, 2018 at 4:44:58 PM UTC+2, Snark wrote: > >> > >> Hi, > >> > >> Le 14/08/2018 à 11:20, Jörg-Volker a écrit : > >> > > >> > Hi, > >> > trying to build Sage 8.3 on my GNU/Linux debian testing/sid system > with > >> > local gcc suite version 8.2.0 fails while compiling scipy. This is on > a > >> > desktop computer with 4 cores (8 threads) and 32 GB RAM. > >> > >> I have no clue what "testing/sid" might be... since as far as I know > >> "sid==unstable" > >> > > It's a mixture of sid and testing. Mainly sid packages but also a few > from testing. > > If you could explain exactly how to reproduce your environment (e.g. > in a Docker container) then I could have a look at it. Otherwise the > best I think anyone can do is make random guesses. > > I took a second look at the package versions on my systems. The development tools, openblas, and gcc are all versions from debian testing. So a testing installation should do it ( Live-CD's are available, for example, here https://cdimage.debian.org/cdimage/weekly-live-builds/amd64/iso-hybrid/ ) Another question I have after looking at the binary package of sage for debian 9 is: how was that built? Since in this package all packages using blas and lapack are linked against the system libraries libblas.so, liblapack.so, and libopenblas.so. > > >> > The build command was > >> > > >> > MAKE='make -j8' SAGE_INSTALL_GCC=no SAGE_BUILD_DIR=/tmp/sage-b > make > >> > build > >> > > >> > > >> > The relevant log file shows > >> > > >> > Traceback (most recent call last): > >> >File "", line 1, in > >> >File "/tmp/.cache-walt/pip-9Uwgks-build/setup.py", line > 416, in > >> > setup_package() > >> >File "/tmp/.cache-walt/pip-9Uwgks-build/setup.py", line > 396, in setup_package > >> > from numpy.distutils.core import setup > >> >File > >> > > "/tmp/sage/sage.git-8.3/local/lib/python2.7/site-packages/numpy/__init__.py", > line 142, in > >> > from . import add_newdocs > >> >File > >> > > "/tmp/sage/sage.git-8.3/local/lib/python2.7/site-packages/numpy/add_newdocs.py", > > line 13, in > >> > from numpy.lib import add_newdoc > >> >File > >> > > "/tmp/sage/sage.git-8.3/local/lib/python2.7/site-packages/numpy/lib/__init__.py", > > line 8, in > >> > from .type_check import * > >> >File > >> > > "/tmp/sage/sage.git-8.3/local/lib/python2.7/site-packages/numpy/lib/type_check.py", > > line 11, in > >> > import numpy.core.numeric as _nx > >> >File > >> > > "/tmp/sage/sage.git-8.3/local/lib/python2.7/site-packages/numpy/core/__init__.py", > > line 26, in > >> > raise ImportError(msg) > >> > ImportError: > >> > Importing the multiarray numpy extension module failed. > Most > >> > likely you are trying to import a failed build of numpy. > >> > If you're working with a numpy git repo, try `git clean > -xdf` (removes all > >> > files not under version control). Otherwise reinstall > numpy. > >> > > >> > Original error was: /usr/lib/x86_64-linux-gnu/libblas.so.3: > undefined symbol: sgemv_thread_n > >> > > >> > Running setup.py install for scipy: finished with status > 'error' > >> > > >> > The system has OpenBLAS installed but with a newer version 0.3.2. > >> > > >> > Any ideas? > >> > > >> > I've built previous versions of Sage on this system before (with gcc > 7.3.0). > >> > > >> > >> Notice that there is a sagemath 8.3 in preparation for Debian, see : > >> https://people.debian.org/~thansen/debian-sage-status.html > >> > >> so perhaps you could lend a hand making it to unstable? > > > > I'll take a look at it. > >> > >> > >> jpuydt on irc.debian.org > > Regards, Jörg. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] document and doctest constructing polyhedra over number fields
While I appreciate dark humour, this is my morning What the Hell moment: illustrating a feature by showing how it does not work, but not by how it works: see https://trac.sagemath.org/ticket/26077 -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.