I notice that your bit-twiddling version avoids the conditional branch for choosing whether to round up or down. Nice, but not easier to explain. (0x00800000 >

I notice that your bit-twiddling version avoids the conditional branch for choosing whether to round up or down. Nice, but not easier to explain. (0x00800000 >> exponent) is easier to explain that what I did (1 << (23 - e)). > C tends to make this sort of thing relatively verbose. If this were assembly, there would be a chance that the test for whether to return the argument without change can be factored with the computation of the quantity (23 - e) (which would be computed directly from the unbiased exponent), so I would really have to try both before deciding on a variant. > although I write this sort of stuff professionally,… Since you are reading this (assuming you are reading this), may I ask you what tools you use? I have been making do with CRlibm and a good C compiler, occasionally relying on long double when extra precision was needed, but it has been close to inadequate a few times. Is there a standard multi-precision floating-point calculator that all floating-point experts use for quick computations, with hexadecimal input and output, or does one have to build one's own? I was thinking of forcing myself to learn either bc or Gappa, but they both seem to have horrible syntax (that is, not a syntax similar to something I'm used to).