Shadow computation with BFloat16 to estimate thenumerical accuracy of summations
Résumé
In this article, we propose to exploit the new computational capability offered by the Bfloat16 representation format to perform shadow computations and compute estimations of the relative error. We demonstrate and evaluate the assumptions under which shadow computation is valid for the summation problem.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|