[GitHub] [incubator-mxnet] leezu edited a comment on issue #14979: [BUG] Using a package with MKL and GPU versions, using python to open a new process will cause an error

[GitBox <gi...@apache.org>]

leezu edited a comment on issue #14979: [BUG] Using a package with MKL and GPU versions, using python to open a new process will cause an error
URL: https://github.com/apache/incubator-mxnet/issues/14979#issuecomment-562910168
 
 
   Thanks @pengzhao-intel. I observe the issue still happens when compiling MXNet master with the llvm-openmp and MKLDNN. It is my understanding, that llvm-openmp and intel openmp share the same codebase, so I suppose this is expected ([reference](https://software.intel.com/en-us/forums/intel-c-compiler/topic/793552))? What do you think? 
   
   For now I'll reopen this issue to facilitate further investigation.
   
   You can build mxnet master version with `cmake -DUSE_CUDA=1 -DUSE_MKLDNN=1 -DCMAKE_EXPORT_COMPILE_COMMANDS=1 -DBUILD_CYTHON_MODULES=1 ..` and you will observe the issue still happens.
   
   This only affects the CMake build, because https://github.com/apache/incubator-mxnet/pull/8730/ only enabled the llvm openmp for the cmake build.
   
   Thus I suggest we revert https://github.com/apache/incubator-mxnet/pull/8730 to fix this issue.
   @pengzhao-intel do you think that is sensible? 
   
   ```
   Segmentation fault: 11
   
   Stack trace:
     [bt] (0) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(+0x14ab799) [0x7f4e85a66799]
     [bt] (1) /lib/x86_64-linux-gnu/libc.so.6(+0x3ef20) [0x7f4f1324bf20]
     [bt] (2) /home/ubuntu/src/mxnet-dc/build/3rdparty/openmp/runtime/src/libomp.so(+0x34250) [0x7f4e84327250]
     [bt] (3) /home/ubuntu/src/mxnet-dc/build/3rdparty/openmp/runtime/src/libomp.so(+0x34d3e) [0x7f4e84327d3e]
     [bt] (4) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(mxnet::engine::OpenMP::set_reserve_cores(int)+0x8ca) [0x7f4e8598e2fa]
     [bt] (5) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(mxnet::engine::ThreadedEnginePerDevice::PushToExecute(mxnet::engine::OprBlock*, bool)::{lambda()#2}::operator()() const+0x4f) [0x7f4e859a05bf]
     [bt] (6) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(std::shared_ptr<mxnet::engine::ThreadedEnginePerDevice::ThreadWorkerBlock<(dmlc::ConcurrentQueueType)1> > mxnet::common::LazyAllocArray<mxnet::engine::ThreadedEnginePerDevice::ThreadWorkerBlock<(dmlc::ConcurrentQueueType)1> >::Get<mxnet::engine::ThreadedEnginePerDevice::PushToExecute(mxnet::engine::OprBlock*, bool)::{lambda()#2}>(int, mxnet::engine::ThreadedEnginePerDevice::PushToExecute(mxnet::engine::OprBlock*, bool)::{lambda()#2})+0x414) [0x7f4e859a1b14]
     [bt] (7) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(mxnet::engine::ThreadedEnginePerDevice::PushToExecute(mxnet::engine::OprBlock*, bool)+0x481) [0x7f4e859a3291]
     [bt] (8) /home/ubuntu/src/mxnet-dc/python/mxnet/../../build/libmxnet.so(mxnet::engine::ThreadedEngine::Push(mxnet::engine::Opr*, mxnet::Context, int, bool)+0x1a8) [0x7f4e859939b8]
   ```

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services