You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2020/03/19 20:18:40 UTC
[GitHub] [incubator-mxnet] leezu opened a new issue #17875: USE_TVM_OP=1
build broken with DMLC_LOG_FATAL_THROW=0
leezu opened a new issue #17875: USE_TVM_OP=1 build broken with DMLC_LOG_FATAL_THROW=0
URL: https://github.com/apache/incubator-mxnet/issues/17875
## Description
Setting `-DDMLC_LOG_FATAL_THROW=0` is useful for debugging with `gdb` or `lldb`. Unfortunately we can't compile MXNet with `USE_TVM_OP=1` in that case.
### Error Message
```
[658/658] Linking CXX shared library libmxnet.so
FAILED: libmxnet.so
: && /usr/bin/c++ -fPIC -Wall -Wno-sign-compare -O3 -std=c++11 -fopenmp -shared -Wl,-soname,libmxnet.so -o libmxnet.so CMakeFiles/mxnet.dir/dummy.c.o -L/home/ubuntu/src/mxnet-master/build/3rdparty/tvm -Wl,-rpath,/usr/local/lib:/home/ubuntu/src/mxnet-master/build/3rdparty/tvm: -Wl,--whole-archive libmxnet.a -Wl,--no-whole-archive libmxnet.a 3rdparty/dmlc-core/libdmlc.a -ldl 3rdparty/mkldnn/src/libdnnl.a -lpthread -Wl,--start-group /opt/intel/mkl/lib/intel64/libmkl_intel_lp64.a /opt/intel/mkl/lib/intel64/libmkl_intel_thread.a /opt/intel/mkl/lib/intel64/libmkl_core.a /usr/local/lib/libiomp5.so -Wl,--end-group /usr/lib/x86_64-linux-gnu/librt.so /usr/lib/x86_64-linux-gnu/libopencv_highgui.so.3.2.0 /usr/lib/x86_64-linux-gnu/libopencv_imgcodecs.so.3.2.0 /usr/lib/x86_64-linux-gnu/libopencv_imgproc.so.3.2.0 /usr/lib/x86_64-linux-gnu/libopencv_core.so.3.2.0 -llapack -ltvm_runtime /usr/lib/gcc/x86_64-linux-gnu/7/libgomp.so -lpthread -lrt && cd /home/ubuntu/src/mxnet-master/build && /home/ubuntu/.pyenv/versions/3.8.2-debug/lib/python3.8/site-packages/cmake/data/bin/cmake -E copy /home/ubuntu/src/mxnet-master/build/3rdparty/mkldnn/include/dnnl_config.h /home/ubuntu/src/mxnet-master/include/mkldnn/ && /home/ubuntu/.pyenv/versions/3.8.2-debug/lib/python3.8/site-packages/cmake/data/bin/cmake -E copy /home/ubuntu/src/mxnet-master/build/3rdparty/mkldnn/include/dnnl_version.h /home/ubuntu/src/mxnet-master/include/mkldnn/ && cd /home/ubuntu/src/mxnet-master/build && /home/ubuntu/.pyenv/versions/3.8.2-debug/lib/python3.8/site-packages/cmake/data/bin/cmake -E env PYTHONPATH="/home/ubuntu/src/mxnet-master/3rdparty/tvm/python:/home/ubuntu/src/mxnet-master/3rdparty/tvm/topi/python:/home/ubuntu/src/mxnet-master/contrib" LD_LIBRARY_PATH=/home/ubuntu/src/mxnet-master/build:/home/ubuntu/src/mxnet-master/build/3rdparty/tvm:/usr/local/cuda/lib64:/usr/local/lib:/usr/lib:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/mpi/lib:/home/ubuntu/local/lib:/home/ubuntu/local/lib:/usr/local/cuda/lib64:/usr/local/lib:/usr/lib:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/mpi/lib:/home/ubuntu/local/lib /home/ubuntu/.pyenv/versions/3.8.2-debug/bin/python3.8 /home/ubuntu/src/mxnet-master/contrib/tvmop/compile.py -o/home/ubuntu/src/mxnet-master/build --config /home/ubuntu/src/mxnet-master/build/tvmop.conf -L /home/ubuntu/src/mxnet-master/build/3rdparty/tvm
[19:56:21] ../3rdparty/tvm/src/node/reflection.cc:113: AttributeError: Variable object has no attributed shape
Child aborted
ninja: build stopped: subcommand failed.
```
## To Reproduce
### Steps to reproduce
1. Apply patch below
2. Compile with `cmake -GNinja -DUSE_CUDA=0 -DUSE_TVM_OP=ON ..; ninja`
``` diff
diff --git a/CMakeLists.txt b/CMakeLists.txt
index 365da2d90..b4c9d267b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -151,7 +151,7 @@ endif()
add_definitions(-DDMLC_MODERN_THREAD_LOCAL=0)
# disable stack trace in exception by default.
add_definitions(-DDMLC_LOG_STACK_TRACE_SIZE=0)
-add_definitions(-DDMLC_LOG_FATAL_THROW=1)
+add_definitions(-DDMLC_LOG_FATAL_THROW=0)
if(MSVC)
add_definitions(-DWIN32_LEAN_AND_MEAN)
```
----------------------------------------------------------------
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
[GitHub] [incubator-mxnet] leezu commented on issue #17875: USE_TVM_OP=1
build broken with DMLC_LOG_FATAL_THROW=0
Posted by GitBox <gi...@apache.org>.
leezu commented on issue #17875: USE_TVM_OP=1 build broken with DMLC_LOG_FATAL_THROW=0
URL: https://github.com/apache/incubator-mxnet/issues/17875#issuecomment-601477244
This failure is from `tvmop/compile.py`. Backtrace:
```
frame #2: 0x00007fff4cc3a496 libtvm.so`dmlc::LogMessageFatal::~LogMessageFatal() + 166 frame #3: 0x00007fff4cdbf0cd libtvm.so`tvm::ReflectionVTable::GetAttr(tvm::runtime::Object*, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&
) const + 605 frame #4: 0x00007fff4cdbf79e libtvm.so`tvm::NodeGetAttr(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) + 766
frame #5: 0x00007fff4cdc0d84 libtvm.so`std::_Function_handler<void (tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*), void (*)(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)>::_M_invoke(std::_Any_data const&, tvm::runtime::TVMArgs&&, tvm::runtime::TVMRetValue*&&) + 20
frame #6: 0x00007fff4d49c75e libtvm.so`TVMFuncCall + 110
frame #7: 0x00007ffff5965dae libffi.so.6`ffi_call_unix64 + 76
frame #8: 0x00007ffff596571f libffi.so.6`ffi_call + 559
```
----------------------------------------------------------------
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
[GitHub] [incubator-mxnet] leezu closed issue #17875: USE_TVM_OP=1 build
broken with DMLC_LOG_FATAL_THROW=0
Posted by GitBox <gi...@apache.org>.
leezu closed issue #17875: USE_TVM_OP=1 build broken with DMLC_LOG_FATAL_THROW=0
URL: https://github.com/apache/incubator-mxnet/issues/17875
----------------------------------------------------------------
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