mxnet-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Anton Chernov <mecher...@gmail.com>
Subject Re: [Discussion] Remove bundled llvm OpenMP
Date Thu, 22 Nov 2018 15:12:11 GMT
Hi Chris,

Thank you for your answer. If you have noticed the initial email comes from
me, Anton Chernov (@lebeg on Github) and thus the proposal is not from any
'Ci' team that you've mentioned, but from me personally.

You are writing:

> someone is doing something unhealthy when they fork ...

I'm missing any context to understand what you mean.

> we get a lot of performance gain from OMP ...

There is no data that would prove this statement and therefore it is a
random guess.

> in many months, no investigation has occurred as to WHY the assertion is
failing.

The investigation has concluded that this is happening due to undefined
behaviour which is, in my opinion, a suffient answer that does not require
to go any deeper.

> the pr is vetoed until such a time that the actual root cause of the
problem is known.

And considering the statements above there is no valid reason to veto the
PR.


Best
Anton

чт, 22 нояб. 2018 г. в 15:38, Chris Olivier <cjolivier01@gmail.com>:

> 3x less overhead*
>
> On Thu, Nov 22, 2018 at 6:25 AM Chris Olivier <cjolivier01@gmail.com>
> wrote:
>
> > someone is doing something unhealthy when they fork, which is causing an
> > assertion in the openmp library. the same assertion that would fire in
> mkl,
> > which is linked to libiomp5 (exact same omp library). this is new
> behavior
> > and most likely due to an error or suboptimal approach in the forking
> logic
> > in mxnet.
> >
> > in order to circumvent the assert, the Ci team is proposing to remove the
> > library completely which is equivalent to cutting off your leg to make
> the
> > pain from stubbing your toe go away.
> >
> > we get a lot of performance gain from OMP. is has about a 1/3 less
> > overhead for entering omp regions and also supports omp regions after a
> > fork, which libgomp does not.
> >
> > in many months, no investigation has occurred as to WHY the assertion is
> > failing.
> >
> > the pr is vetoed until such a time that the actual root cause of the
> > problem is known.
> >
> >
> > thanks,
> >
> > -Chris.
> >
> >
> >
> >
> > On Thu, Nov 22, 2018 at 4:36 AM Anton Chernov <mechernov@gmail.com>
> wrote:
> >
> >> Dear MXNet community,
> >>
> >> I would like to drive attention to an important issue that is present in
> >> the MXNet CMake build: usage of bundled llvm OpenMP library.
> >>
> >> I have opened a PR to remove it:
> >> https://github.com/apache/incubator-mxnet/pull/12160
> >>
> >> The issue was closed, but I am strong in my oppinion that it's the right
> >> thing to do.
> >>
> >> *Background*
> >> If you want to use OpenMP pragmas in your code for parallelization you
> >> would supply a special flag to the compiler:
> >>
> >> - Clang / -fopenmp
> >> https://openmp.llvm.org/
> >>
> >> - GCC / -fopenmp
> >> https://gcc.gnu.org/onlinedocs/libgomp/Enabling-OpenMP.html
> >>
> >> - Intel / [Q]openmp
> >>
> >>
> https://software.intel.com/en-us/node/522689#6E24682E-F411-4AE3-A04D-ECD81C7008D1
> >>
> >> - Visual Studio: /openmp (Enable OpenMP 2.0 Support)
> >> https://msdn.microsoft.com/en-us/library/tt15eb9t.aspx
> >>
> >> Each of the compilers would enable the '#pragma omp' directive during
> >> C/C++
> >> compilation and arrange for automatic linking of the OpenMP runtime
> >> library
> >> supplied by each complier separately.
> >>
> >> Thus, to use the advantages of an OpenMP implementation one has to
> compile
> >> the code with the corresponding compiler.
> >>
> >> Currently, in MXNet CMake build scripts a bundled version of llvm OpenMP
> >> is
> >> used ([1] and [2]) to replace the OpenMP library supplied by the
> compiler.
> >>
> >> I will quote here the README from the MKL-DNN (Intel(R) Math Kernel
> >> Library
> >> for Deep Neural Networks):
> >>
> >> "Intel MKL-DNN uses OpenMP* for parallelism and requires an OpenMP
> runtime
> >> library to work. As different OpenMP runtimes may not be binary
> compatible
> >> it's important to ensure that only one OpenMP runtime is used throughout
> >> the application. Having more than one OpenMP runtime initialized may
> lead
> >> to undefined behavior resulting in incorrect results or crashes." [3]
> >>
> >> And:
> >>
> >> "Using GNU compiler with -fopenmp and -liomp5 options will link the
> >> application with both Intel and GNU OpenMP runtime libraries. This will
> >> lead to undefined behavior of the application." [4]
> >>
> >> As can be seen from ldd for MXNet:
> >>
> >> $ ldd build/tests/mxnet_unit_tests | grep omp
> >>     libomp.so => /.../mxnet/build/3rdparty/openmp/runtime/src/libomp.so
> >> (0x00007f697bc55000)
> >>     libgomp.so.1 => /usr/lib/x86_64-linux-gnu/libgomp.so.1
> >> (0x00007f69660cd000)
> >>
> >> *Performance*
> >>
> >> The only performance data related to OpenMP in MXNet I was able to find
> is
> >> here:
> >>
> >>
> https://github.com/apache/incubator-mxnet/issues/9744#issuecomment-367711172
> >>
> >> Which in my understanding is testing imact of different environment
> >> variables for the same setup (using same bundled OpenMP library).
> >>
> >> The libraries may differ in implementation and the Thread Affinity
> >> Interface [5] may have significant impact on performance.
> >>
> >> All compliers support it:
> >>
> >> - Clang / KMP_AFFINITY
> >>
> >>
> https://github.com/clang-ykt/openmp/blob/master/runtime/src/kmp_affinity.cpp
> >>
> >> - GCC / GOMP_CPU_AFFINITY
> >>
> >>
> https://gcc.gnu.org/onlinedocs/gcc-4.7.1/libgomp/GOMP_005fCPU_005fAFFINITY.html
> >>
> >> - Intel / KMP_AFFINITY
> >>
> >>
> https://software.intel.com/en-us/node/522689#6E24682E-F411-4AE3-A04D-ECD81C7008D1
> >>
> >> - Visual Studio / SetThreadAffinityMask
> >>
> >>
> https://docs.microsoft.com/en-us/windows/desktop/api/winbase/nf-winbase-setthreadaffinitymask
> >>
> >> *Issues*
> >>
> >> Failed OpenMP assertion when loading MXNet compiled with DEBUG=1
> >> https://github.com/apache/incubator-mxnet/issues/10856
> >>
> >> libomp.so dependency (need REAL fix)
> >> https://github.com/apache/incubator-mxnet/issues/11417
> >>
> >> mxnet-mkl (v0.12.0) crash when using (conda-installed) numpy with MKL
> >> https://github.com/apache/incubator-mxnet/issues/8532
> >>
> >> Performance regression when OMP_NUM_THREADS environment variable is not
> >> set
> >> https://github.com/apache/incubator-mxnet/issues/9744
> >>
> >> Poor concat CPU performance on CUDA builds
> >> https://github.com/apache/incubator-mxnet/issues/11905
> >>
> >> I would appreciate hearing your thoughts.
> >>
> >>
> >> Best
> >> Anton
> >>
> >> [1]
> >>
> >>
> https://github.com/apache/incubator-mxnet/blob/master/CMakeLists.txt#L400-L405
> >> [2] https://github.com/apache/incubator-mxnet/tree/master/3rdparty
> >> [3] https://github.com/intel/mkl-dnn/blame/master/README.md#L261-L265
> >> [4] https://github.com/intel/mkl-dnn/blame/master/README.md#L278-L280
> >> [5] https://software.intel.com/en-us/node/522691
> >>
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message