Hi Sakthi Vinayaga Moorthy V,
thanks for your detailed error description. I think we already found the reason for this issue - which is more or less a compiler bug.
We solved the issue in V14.01.00.00 of the Build Environments for Linux based platforms. See the Release Notes:
Bug-fix: Incorrect floating point calculation on targets with NEON-VFPV4 architecture in combination with GCC compiler. This behavior can be avoided by using #pragma GCC optimize ("-ffp-contract=off").
Are you using the latest version? Otherwise try the mentioned pragma.
Does this help?
Best regards,
Manfred.