Optimized pow() approximation for Java, C / C++, and C# – Martin Ankerl
I don’t know if this is helpful to you 7 years later, but maybe it will be helpful to someone else. I derived this pow method for 32-bit floats from Schraudolph’s 1999 paper. The new value to subtract and add the the int-interpreted value is 1065307417 (r...