> > > To be clear, I was just mentioning that I use single precision floats in > our main application that is all written in C++, which is where the > performance is critical/required. I'm trying to use Racket as a tool for > understanding how we can maximize accuracy of our numerical algorithms > under the restriction of single precision floats and the Racket code can be > significantly slower since it is all intended to be run offline for the > purposes of understanding how the floating point numbers behave against a > reference implementation. >
Have you looked at Neil Toronto's paper Practically Accurate Floating-Point Math ( https://www.cs.umd.edu/~ntoronto/papers/toronto-2014cise-floating-point.pdf). It is an excellent paper that discusses the use of Racket to analyze the accuracy of floating-point computations. Doug -- > You received this message because you are subscribed to the Google Groups > "Racket Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Racket Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
