glibc floating point math functions provide slightly different results between RHEL major releases

Issue

• exp2f() provides different results in RHEL 7 and RHEL 8
• We have identified a difference in behavior in libm provided by glibc between the version on RHEL 7 (glibc 2.17) and the version on RHEL8 (glibc-2.28). It's a small difference in precision of the exp2f() function; it's so small that it's within the limit of precision of glibc, so it's not an error.
• glibc EXP function calculates a different result between RH7 and RH8 and back again in RH9