[build] Fix -Wconversion double-to-float truncation warnings

[oznogon]

A build with the -Wconversion compiler flag generates 97 float-conversion double-to-float truncation warnings in GCC, such as:

EmptyEpsilon/src/beamTemplate.cpp:16:26:
warning:
conversion from ‘double’ to ‘float’ changes value from ‘2.0e-2’ to ‘1.99999996e-2f’
[-Wfloat-conversion]

This PR f-suffixes decimal numbers generating these warnings to make them explicitly floats, and others that aren't generating warnings but are inconsistently presented near changed values.

Values aren't changed, only suffixed.