FIx error when building with oneAPI release compiler

[muhammad-tanvir-1211]

This PR fixes an error seen from googlebench and gtest (warning in this case) that explicit comparison with NaN is not allowed in fast precision mode.

[pgorlani]

If the purpose of the patch is removing the warning during the unit test/benchmark compilation, it would be better to set -Wno-tautological-constant-compare in test/CMakeLists.txt and benchmark/portblas/CMakeLists.txt.

-ffp-model=precise flag is an umbrella option that could have side consequences on many aspects since it "disables optimizations that are not value-safe on floating-point data".