Nightly clang/7+cuda/9.2 build failure with -Werror: /gmres.hpp:273:21: error: using integer absolute value function 'abs' when argument is of floating point type

[ndellingwood]

The nightly clang/7+cuda/9.2 build began failing with error message:

In file included from /ascldap/users/ndellin/kokkos-kernels/example/gmres/test_real_A.cpp:54:
/ascldap/users/ndellin/kokkos-kernels/example/gmres/gmres.hpp:273:21: error: using integer absolute value function 'abs' when argument is of floating point type [-Werror,-Wabsolute-value]
      shortRelRes = abs(GVec_h(j+1))/nrmB; // this abs is in libstdc++
                    ^
/ascldap/users/ndellin/kokkos-kernels/example/gmres/test_real_A.cpp:102:27: note: in instantiation of function template specialization 'gmres<double, Kokkos::LayoutLeft, Kokkos::Cuda, int>' requested here
  GmresStats solveStats = gmres<ST, Kokkos::LayoutLeft, EXSP>(A, B, X, solverOpts);
                          ^
/ascldap/users/ndellin/kokkos-kernels/example/gmres/gmres.hpp:273:21: note: use function 'std::abs' instead
      shortRelRes = abs(GVec_h(j+1))/nrmB; // this abs is in libstdc++
                    ^~~
                    std::abs

The only change in kokkos-kernels was merge of #1171 , but it's not clear to me how those changes would impact this build (abs should be called from blas not batched blas, right?)

@e10harvey the compiler suggests explicit namespacing of abs with std, but I saw there are additional abs routines in this code to handle half precision types and wasn't certain of the best way to handle this. Can you offer a suggestion or look into the failure as well?

Reproducer (kokkos-dev-2):

module load sems-env sems-cmake/3.17.1 sems-clang/7.0.1 sems-cuda/9.2

$KOKKOSKERNELS_PATH/cm_generate_makefile.bash --with-devices=Cuda,OpenMP --arch=Volta70 --compiler=clang++ --cxxflags="-O3 -Wall -Wunused-parameter -Wshadow -pedantic -Werror -Wsign-compare -Wtype-limits -Wuninitialized " --cxxstandard="14"  --kokkos-path=$KOKKOS_PATH --kokkoskernels-path=$KOKKOSKERNELS_PATH --with-scalars='double,complex_double' --with-ordinals=int --with-offsets=int,size_t --with-layouts=LayoutLeft  --with-options=disable_deprecated_code --with-cuda-options=enable_lambda   --no-examples