onnxruntime: [Build] Unable to suppress unused variable

Describe the issue

Hi OnnxRuntime team, I’m trying to build OnnxRuntime on Jetson platform.

But keep failing due to unused variable error:

./build.sh --config Release --skip_submodule_sync --parallel --build_shared_lib --build_dir /home/nvidia/triton/onnxruntime/build --update --build --use_cuda --cuda_home /usr/local/cuda --cudnn_home /usr/lib/aarch64-linux-gnu --use_tensorrt --tensorrt_home /usr/src/tensorrt --cmake_extra_defines CMAKE_CXX_FLAGS=-Wunused-variable --cmake_extra_defines 'CMAKE_CUDA_ARCHITECTURES=53;62;72;87'
...
...
...
-- Found PythonInterp: /usr/bin/python3 (found version "3.8.10") 
Generated: /raid/home/nvidia/triton/onnxruntime/build/Release/_deps/onnx-build/onnx/onnx-ml.proto
Generated: /raid/home/nvidia/triton/onnxruntime/build/Release/_deps/onnx-build/onnx/onnx-operators-ml.proto
Generated: /raid/home/nvidia/triton/onnxruntime/build/Release/_deps/onnx-build/onnx/onnx-data.proto
-- 
-- ******** Summary ********
--   CMake version             : 3.26.3
--   CMake command             : /usr/bin/cmake
--   System                    : Linux
--   C++ compiler              : /usr/bin/c++
--   C++ compiler version      : 9.4.0
--   CXX flags                 : -Wunused-variable -ffunction-sections -fdata-sections -DCPUINFO_SUPPORTED -Wnon-virtual-dtor
--   Build type                : Release
--   Compile definitions       : ORT_ENABLE_STREAM;EIGEN_MPL2_ONLY;_GNU_SOURCE;__STDC_FORMAT_MACROS
--   CMAKE_PREFIX_PATH         : /home/nvidia/triton/onnxruntime/build/Release/installed
--   CMAKE_INSTALL_PREFIX      : /usr/local
--   CMAKE_MODULE_PATH         : /raid/home/nvidia/triton/onnxruntime/cmake/external
-- 
--   ONNX version              : 1.14.0rc1
--   ONNX NAMESPACE            : onnx
--   ONNX_USE_LITE_PROTO       : OFF
--   USE_PROTOBUF_SHARED_LIBS  : OFF
--   Protobuf_USE_STATIC_LIBS  : ON
--   ONNX_DISABLE_EXCEPTIONS   : OFF
--   ONNX_WERROR               : OFF
--   ONNX_BUILD_TESTS          : OFF
--   ONNX_BUILD_BENCHMARKS     : OFF
-- 
--   Protobuf compiler         : 
--   Protobuf includes         : 
--   Protobuf libraries        : 
--   BUILD_ONNX_PYTHON         : OFF
Finished fetching external dependencies
-- Performing Test HAS_UNUSED_BUT_SET_PARAMETER
-- Performing Test HAS_UNUSED_BUT_SET_PARAMETER - Success
-- Performing Test HAS_UNUSED_BUT_SET_VARIABLE
-- Performing Test HAS_UNUSED_BUT_SET_VARIABLE - Success
-- Performing Test HAS_UNUSED_VARIABLE
-- Performing Test HAS_UNUSED_VARIABLE - Success
...
...
...
[ 56%] Building CXX object CMakeFiles/onnxruntime_providers.dir/raid/home/nvidia/triton/onnxruntime/onnxruntime/core/providers/cpu/math/gemm.cc.o
[ 56%] Building CXX object CMakeFiles/onnxruntime_providers_cuda.dir/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc.o
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc: In instantiation of ‘onnxruntime::common::Status onnxruntime::contrib::cuda::Attention<T>::ComputeInternal(onnxruntime::OpKernelContext*) const [with T = onnxruntime::MLFloat16]’:
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.h:21:10:   required from here
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc:105:8: error: unused variable ‘is_mask_1d_key_seq_len_start’ [-Werror=unused-variable]
  105 |   bool is_mask_1d_key_seq_len_start = parameters.mask_type == AttentionMaskType::MASK_1D_KEY_SEQ_LEN_START;
      |        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc: In instantiation of ‘onnxruntime::common::Status onnxruntime::contrib::cuda::Attention<T>::ComputeInternal(onnxruntime::OpKernelContext*) const [with T = float]’:
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.h:21:10:   required from here
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc:105:8: error: unused variable ‘is_mask_1d_key_seq_len_start’ [-Werror=unused-variable]
cc1plus: all warnings being treated as errors
make[2]: *** [CMakeFiles/onnxruntime_providers_cuda.dir/build.make:2084: CMakeFiles/onnxruntime_providers_cuda.dir/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
...
...
...
[ 66%] Linking CXX static library libonnxruntime_providers.a
[ 66%] Built target onnxruntime_providers
make: *** [Makefile:166: all] Error 2
Traceback (most recent call last):
  File "/home/nvidia/triton/onnxruntime/tools/ci_build/build.py", line 2597, in <module>
    sys.exit(main())
  File "/home/nvidia/triton/onnxruntime/tools/ci_build/build.py", line 2493, in main
    build_targets(args, cmake_path, build_dir, configs, num_parallel_jobs, args.target)
  File "/home/nvidia/triton/onnxruntime/tools/ci_build/build.py", line 1432, in build_targets
    run_subprocess(cmd_args, env=env)
  File "/home/nvidia/triton/onnxruntime/tools/ci_build/build.py", line 779, in run_subprocess
    return run(*args, cwd=cwd, capture_stdout=capture_stdout, shell=shell, env=my_env)
  File "/raid/home/nvidia/triton/onnxruntime/tools/python/util/run.py", line 49, in run
    completed_process = subprocess.run(
  File "/usr/lib/python3.8/subprocess.py", line 516, in run
    raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['/usr/bin/cmake', '--build', '/home/nvidia/triton/onnxruntime/build/Release', '--config', 'Release', '--', '-j4']' returned non-zero exit status 2.

Urgency

We are trying to build it against rel-1.15.0 branch and I would like to confirm it before release date.

Target platform

aarch64

Build script

./build.sh --config Release --skip_submodule_sync --parallel --build_shared_lib --build_dir /home/nvidia/triton/onnxruntime/build --update --build --use_cuda --cuda_home /usr/local/cuda --cudnn_home /usr/lib/aarch64-linux-gnu --use_tensorrt --tensorrt_home /usr/src/tensorrt --cmake_extra_defines CMAKE_CXX_FLAGS=-Wunused-variable --cmake_extra_defines 'CMAKE_CUDA_ARCHITECTURES=53;62;72;87'

Error / output

[ 56%] Building CXX object CMakeFiles/onnxruntime_providers_cuda.dir/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc.o
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc: In instantiation of ‘onnxruntime::common::Status onnxruntime::contrib::cuda::Attention<T>::ComputeInternal(onnxruntime::OpKernelContext*) const [with T = onnxruntime::MLFloat16]’:
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.h:21:10:   required from here
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc:105:8: error: unused variable ‘is_mask_1d_key_seq_len_start’ [-Werror=unused-variable]
  105 |   bool is_mask_1d_key_seq_len_start = parameters.mask_type == AttentionMaskType::MASK_1D_KEY_SEQ_LEN_START;
      |        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc: In instantiation of ‘onnxruntime::common::Status onnxruntime::contrib::cuda::Attention<T>::ComputeInternal(onnxruntime::OpKernelContext*) const [with T = float]’:
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.h:21:10:   required from here
/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc:105:8: error: unused variable ‘is_mask_1d_key_seq_len_start’ [-Werror=unused-variable]
cc1plus: all warnings being treated as errors
make[2]: *** [CMakeFiles/onnxruntime_providers_cuda.dir/build.make:2084: CMakeFiles/onnxruntime_providers_cuda.dir/raid/home/nvidia/triton/onnxruntime/onnxruntime/contrib_ops/cuda/bert/attention.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....

Visual Studio Version

No response

GCC / Compiler Version

No response