I think the problem is that you use icpc as the compiler and Theano don't know how to find the arch parameter automatically for it. Do you have your home directory on the network or put our local to the computer? If it is local, add -match=native to the cxxflags Theano flag.
This is probably not your problem of slowness. Fred Le lun. 20 févr. 2017 12:07, nacho arroyo <[email protected]> a écrit : > As Theano says we can sent this warning here, I do it. > > It is very important to me solving this, because I think this is the > reason I have very slow training. > > WARNING (theano.gof.cmodule): OPTIMIZATION WARNING: Theano was not able > to find the g++ parameters that tune the compilation to your specific CPU > . This can slow down the execution of Theano functions. Please submit the > following lines to Theano's mailing list so that we can fix this problem: > ['# 1 "-"\n', 'icpc version 17.0.1 (gcc version 4.8.0 compatibility)\n', > '/opt/intel/compilers_and_libraries_2017.1.132/linux/bin/intel64/mcpcom > --target_efi2 -_g -mP3OPT_inline_alloca -D__ICC=1700 > -D__INTEL_COMPILER=1700 -D__INTEL_COMPILER_UPDATE=1 -D__PTRDIFF_TYPE__=long > "-D__SIZE_TYPE__=unsigned long" -D__WCHAR_TYPE__=int > "-D__WINT_TYPE__=unsigned int" "-D__INTMAX_TYPE__=long int" > "-D__UINTMAX_TYPE__=long unsigned int" -D__LONG_MAX__=9223372036854775807L > -D__QMSPP_ -D__OPTIMIZE__ -D__NO_MATH_INLINES -D__NO_STRING_INLINES > -D__GNUC_GNU_INLINE__ -D__GNUC__=4 -D__GNUC_MINOR__=8 > -D__GNUC_PATCHLEVEL__=0 -D__LP64__ -D_LP64 -D__GXX_ABI_VERSION=1002 > "-D__USER_LABEL_PREFIX__= " -D__REGISTER_PREFIX__= -D__INTEL_RTTI__ > -D__unix__ -D__unix -D__linux__ -D__linux -D__gnu_linux__ -B -Dunix -Dlinux > "-_Asystem(unix)" -D__ELF__ -D__x86_64 -D__x86_64__ -D__amd64 -D__amd64__ > "-_Acpu(x86_64)" "-_Amachine(x86_64)" > -D__INTEL_COMPILER_BUILD_DATE=20161005 -D__INTEL_OFFLOAD -D__pentium4 > -D__pentium4__ -D__tune_pentium4__ -D__SSE2__ -D__SSE2_MATH__ -D__SSE3__ > -D__SSSE3__ -D__SSE4_1__ -D__SSE4_2__ -D__SSE__ -D__SSE_MATH__ -D__MMX__ > -D__AVX__ -_k -_8 -_l --has_new_stdarg_support -_a -_b -E > --gnu_version=40800 -_W5 --gcc-extern-inline --multibyte_chars > -mGLOB_diag_suppress_sys --array_section --simd --simd_func > --offload_mode=1 --offload_target_names=gfx,GFX,mic,MIC > --offload_unique_string=icpc1145998360jqCGIp -mP1OPT_print_version=FALSE > -mGLOB_em64t=TRUE -mP1OPT_version=17.0-intel64 -mGLOB_long_size_64 > -mGLOB_routine_pointer_size_64 -\n', '#include "..." search starts > here:\n', '#include <...> search starts here:\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/ipp/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/mkl/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/tbb/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/daal/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/ipp/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/mkl/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/tbb/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/daal/include\n', ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/compiler/include/intel64\n', > ' > /opt/intel/compilers_and_libraries_2017.1.132/linux/compiler/include/icc\n', > ' /opt/intel/compilers_and_libraries_2017.1.132/linux/compiler/include\n', > ' /usr/local/include\n', ' /usr/lib/gcc/x86_64-linux-gnu/4.8/include\n', ' > /usr/lib/gcc/x86_64-linux-gnu/4.8/include-fixed\n', ' /usr/include/\n', ' > /usr/include\n', ' /usr/include/x86_64-linux-gnu\n', 'End of search > list.\n'] > > My g++ --version output command: > > ~$ g++ --version > g++ (Ubuntu 4.8.4-2ubuntu1~14.04.3) 4.8.4 > Copyright (C) 2013 Free Software Foundation, Inc. > This is free software; see the source for copying conditions. There is NO > warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE > . > > Please help, given I have not GPUs I installed > https://github.com/intel/Theano and Intel® *Parallel* Studio XE *2017 *and > my ~/.theanorc file is as follows: > > [global] > device = cpu > floatX = float32 > cxx = icpc > mode = FAST_RUN > openmp = True > openmp_elemwise_minsize = 10 > allow_gc = False > [dnn] > enabled = mkl > [gcc] > #cxxflags = -qopenmp -march=native -O3 -qopt-report=3 -fno-alias > -qopt-prefetch > =2 -fp-trap=none > cxxflags = -qopenmp -O3 -qopt-report=3 -fno-alias -qopt-prefetch=2 -fp- > trap=non > e > [blas] > ldflags = -lmkl_rt > > I did this as one of the suggestions from someone here: > http://stackoverflow.com/questions/41583628/efficient-kernel-implementation-in-theano > In the case you suggest something else to do, it would be very appreciated > if you can tell me. > > Thanks > > > > > -- > > --- > You received this message because you are subscribed to the Google Groups > "theano-users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
