[Build] onnxruntime build fails on ordino xu4

[DefinitiveIch]

Describe the issue

When I try to build onnxruntime on the ordino xu4 I get an error running the following command:

./build.sh --config MinSizeRel --update --build

Urgency

No response

Target platform

Ubuntu 22.04.3 LTS

Build script

pip3 install --upgrade pip
pip3 install --upgrade setuptools
pip3 install --upgrade wheel
pip3 install numpy

Build the latest cmake

mkdir /code
cd /code
wget https://github.com/Kitware/CMake/releases/download/v3.26.4/cmake-3.26.4.tar.gz
tar zxf cmake-3.26.4.tar.gz

cd /code/cmake-3.26.4
./configure --system-curl
make
sudo make install

Prepare onnxruntime Repo

cd /code
git clone --recursive https://github.com/Microsoft/onnxruntime.git

Start the basic build

cd /code/onnxruntime
./build.sh --config MinSizeRel --update --build

Build Shared Library

./build.sh --config MinSizeRel --build_shared_lib

Build Python Bindings and Wheel

./build.sh --config MinSizeRel --enable_pybind --build_wheel

Build Output

ls -l /code/onnxruntime/build/Linux/MinSizeRel/.so
ls -l /code/onnxruntime/build/Linux/MinSizeRel/dist/.whl

Error / output

/mnt/onnxruntime/build/Linux/MinSizeRel/_deps/eigen-src/Eigen/src/SparseCore/CompressedStorage.h:93:80: error: overflow in conversion from ‘long long int’ to ‘int’ changes value from ‘9223372036854775807’ to ‘-1’ [-Werror=overflow]
   93 |         Index realloc_size = (std::min<Index>)(NumTraits<StorageIndex>::highest(),  size + Index(reserveSizeFactor*double(size)));
      |                                                ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~
In file included from /mnt/onnxruntime/include/onnxruntime/core/framework/sparse_tensor.h:9,
                 from /mnt/onnxruntime/onnxruntime/contrib_ops/cpu/math/sparse_dense_matmul.cc:6:
/mnt/onnxruntime/include/onnxruntime/core/framework/tensor_shape.h: In member function ‘virtual onnxruntime::common::Status onnxruntime::contrib::SparseToDenseMatMul::Compute(onnxruntime::OpKernelContext*) const’:
/mnt/onnxruntime/include/onnxruntime/core/framework/tensor_shape.h:86:106: note: parameter passing for argument of type ‘gsl::span<const long long int>’ changed in GCC 7.1
   86 |   TensorShape(std::initializer_list<int64_t> dims) : TensorShape(gsl::make_span(dims.begin(), dims.end())) {}
      |                 

Visual Studio Version

No response

GCC / Compiler Version

11.4.0

[RyanUnderhill]

This might explain our issue: https://gitlab.com/libeigen/eigen/-/issues/2517
"Setting Eigen::Index to 32-bit does not appear to work with using a StorageIndex type that's 64-bit."

Can you give more context on the build error? What file is including this, etc?

[DefinitiveIch]

Sorry the close was a miss click.
I updated the error message.

[github-actions]

This issue has been automatically marked as stale due to inactivity and will be closed in 7 days if no further activity occurs. If further support is needed, please provide an update and/or more details.

[github-actions]

This issue has been automatically closed due to inactivity. Please reactivate if further support is needed.