19/11/2015

Intervenant(s) : Olga Kupriianova (AriC, ENS de Lyon)

This work investigates the two ways of enlarging the current floating-point environment: the first is to support several implementation versions for each mathematical function (elementary as exp, log and special as erf, Γ), the second one is to provide IEEE754 operations that mix the inputs and the output of different radices. As the number of various implementations for each mathematical function is large, this work is focused on code generation. Our code generator is addressed to the huge variety of functions: it generates parametrized implementations for the user-specified functions. So it may be considered as a black-box function generator. Since the revision of the IEEE754 Standard in 2008 it is possible to mix numbers of different precisions in one operation. However, there is no mechanism that allows users to mix numbers of different radices in one operation.

marc (at) nullmezzarobba.net