From: Daniel P. <dp...@gm...> - 2014-02-06 20:23:14
|
OK, I see the problem. This is a known issue with gcc-4.8.2.1, there is a bug in the implementation of std::sort in the standard library. We were not able to work around this bug, which is a pretty severe one. The next minor revision number of gcc should fix it. I'm surprised this version even appears in a widely used repo, because this is a bad bug. Dan On Thu, Feb 6, 2014 at 2:56 PM, Simon Klüpfel <sim...@gm...>wrote: > Hi Dan, > > here is the output of: > > valgrind ./matrix-lib-test > matrix-lib-test.valgrind.log 2>&1 > > I hope it helps in some way. > > Simon > > > On 02/06/2014 07:42 PM, Daniel Povey wrote: > >> I don't think revision number is relevant here. >> Dan >> >> >> >> On Thu, Feb 6, 2014 at 2:41 PM, Simon Klüpfel <sim...@gm... >> <mailto:sim...@gm...>> wrote: >> >> Hi Dan, >> >> installing valgrind right now. Shall I also update to the latest >> revision, or had there not been any changes in the relevant sources? >> >> Simon >> >> >> On 02/06/2014 07:40 PM, Daniel Povey wrote: >> >> Please run it in valgrind, e.g. >> valgrind ./matrix-lib-test >> Dan >> >> >> On Thu, Feb 6, 2014 at 2:36 PM, Simon Klüpfel >> <sim...@gm... <mailto:sim...@gm...> >> <mailto:simon.kluepfel@gmail.__com >> >> <mailto:sim...@gm...>>> wrote: >> >> Hi, >> >> I have been today experiencing a problem with my kaldi >> installation >> (most of today, actually... unfortunately). I had an update >> on my >> system (LMDE) recently, and perhaps some update on gcc >> might have >> caused it, but perhaps there might also just be a bug in >> Kaldi. >> >> I had set up a script to install a certain revision of >> Kaldi on our >> machines, to have something solid to work with and build >> around, and >> just update to a newer revision if it might be >> necessary/beneficial. >> >> I had so far chosen revision 3460 of kaldi-trunk. I wrote a >> script >> to compile the needed dependencies (LAPACK/BLAS, portaudio, >> openfst) >> and went all the way compiling kaldi and the external >> components. >> >> Today I included the tests to the script, and the 'make >> test' fails >> in matrix-lib-test. I have tried the following setups: >> >> OpenBLAS + Kaldi shared >> ATLAS + Kaldi shared >> ATLAS + Kaldi static >> >> The error occurred in all these cases and reads as in the >> attached >> 'error.log' file (of the last setup, hence no mention of >> libatlas.so >> or alike). It was a very similar error in all cases. >> >> I did not yet try the latest revision, as I could not find >> any hint >> to changes connected to this error in the commits, but I >> could do >> this. If you need any more detailed output or me invoking a >> debugger >> or something else, please tell me. >> >> My system specs: >> >> $ uname -a >> Linux g500s 3.10-2-amd64 #1 SMP Debian 3.10.5-1 >> (2013-08-07) x86_64 >> GNU/Linux >> >> $ gcc -v >> Using built-in specs. >> COLLECT_GCC=gcc >> >> COLLECT_LTO_WRAPPER=/usr/lib/____gcc/x86_64-linux-gnu/4.8/ >> lto-____wrapper >> >> >> Target: x86_64-linux-gnu >> Configured with: ../src/configure -v >> --with-pkgversion='Debian >> 4.8.2-1' >> --with-bugurl=file:///usr/____share/doc/gcc-4.8/README.Bugs >> --enable-languages=c,c++,java,____go,d,fortran,objc,obj-c++ >> >> >> --prefix=/usr --program-suffix=-4.8 --enable-shared >> --enable-linker-build-id --libexecdir=/usr/lib >> --without-included-gettext --enable-threads=posix >> --with-gxx-include-dir=/usr/____include/c++/4.8 >> >> --libdir=/usr/lib >> >> --enable-nls --with-sysroot=/ --enable-clocale=gnu >> --enable-libstdcxx-debug --enable-libstdcxx-time=yes >> --enable-gnu-unique-object --enable-plugin --with-system-zlib >> --disable-browser-plugin --enable-java-awt=gtk >> --enable-gtk-cairo >> --with-java-home=/usr/lib/jvm/____java-1.5.0-gcj-4.8-amd64/ >> jre >> --enable-java-home >> --with-jvm-root-dir=/usr/lib/____jvm/java-1.5.0-gcj-4.8- >> amd64 >> >> --with-jvm-jar-dir=/usr/lib/____jvm-exports/java-1.5.0-gcj- >> 4.____8-amd64 >> --with-arch-directory=amd64 >> --with-ecj-jar=/usr/share/____java/eclipse-ecj.jar >> --enable-objc-gc >> --enable-multiarch --with-arch-32=i586 --with-abi=m64 >> --with-multilib-list=m32,m64,____mx32 --with-tune=generic >> >> >> --enable-checking=release --build=x86_64-linux-gnu >> --host=x86_64-linux-gnu --target=x86_64-linux-gnu >> Thread model: posix >> gcc version 4.8.2 (Debian 4.8.2-1) >> >> >> I hope there is a way to get this to work. >> >> All the best and thanks for this great code! >> >> Simon >> >> >> ------------------------------__---------------------------- >> --__------------------ >> >> Managing the Performance of Cloud-Based Applications >> Take advantage of what the Cloud has to offer - Avoid >> Common Pitfalls. >> Read the Whitepaper. >> http://pubads.g.doubleclick.__net/gampad/clk?id=121051231&__ >> iu=/4140/ostg.clktrk >> <http://pubads.g.doubleclick.net/gampad/clk?id=121051231& >> iu=/4140/ostg.clktrk> >> _________________________________________________ >> Kaldi-developers mailing list >> Kaldi-developers@lists.__sourceforge.net >> <mailto:Kal...@li...> >> <mailto:Kaldi-developers@__lists.sourceforge.net >> <mailto:Kal...@li...>> >> https://lists.sourceforge.net/__lists/listinfo/kaldi-__developers< >> https://lists.sourceforge.net/lists/listinfo/kaldi-developers> >> >> >> >> >> > |