[BUG] Warning Message for Host call in device function for Einsum

[tylera-nvidia]

Describe the Bug When using the einsum API, a compile warning is made warning that a host function is called from a host/device function, which is not allowed.

I have verified this is not an issue, so I'm not sure if we are only calling this from the host, or if the warning about a host call being used on the device is incorrect.

below is the warning:

[ 50%] Building CUDA object CMakeFiles/einsumTest.dir/einsumTest.cu.o
/scratch/playground/matxExample/MatX/include/matx/core/tensor_impl.h(657): warning #20011-D: calling a __host__ function("std::__cxx11::basic_string<char,     ::std::char_traits<char> , ::std::allocator<char> > ::basic_string(const ::std::__cxx11::basic_string<char,     ::std::char_traits<char> , ::std::allocator<char> > &)") from a __host__ __device__ function("matx::detail::EinsumOp< ::matx::tensor_t<float, (int)3,  ::matx::basic_storage< ::matx::raw_pointer_buffer<float,  ::matx::matx_allocator<float> > > ,  ::matx::tensor_desc_t< ::cuda::std::__4::array<long long, (unsigned long)3ul> ,  ::cuda::std::__4::array<long long, (unsigned long)3ul> , (int)3> > ,  ::matx::tensor_t<float, (int)3,  ::matx::basic_storage< ::matx::raw_pointer_buffer<float,  ::matx::matx_allocator<float> > > ,  ::matx::tensor_desc_t< ::cuda::std::__4::array<long long, (unsigned long)3ul> ,  ::cuda::std::__4::array<long long, (unsigned long)3ul> , (int)3> >  > ::EinsumOp") is not allowed

Remark: The warnings can be suppressed with "-diag-suppress <warning-number>"

[100%] Linking CUDA executable einsumTest

To Reproduce

make any einsum call

Expected Behavior

No warning is presented

Code Snippets

  auto a = matx::make_tensor<float>({3,1,1});
  auto b = matx::make_tensor<float>({1,3,2});
  auto c = matx::make_tensor<float>({1,2});
  
  a.SetVals({{{1,2,3}}});
  b.SetVals({{{1,2},{3,4},{5,6}}});
  (c = matx::cutensor::einsum("ijk,jil->kl", a, b)).run();
  matx::print(c);

System Details (please complete the following information):

• OS: Ubuntu 22.04
• CUDA version: 12.5
• g++ version: 11.4