Mixed Precision Linear Algebra Algorithms
- Date:
- Time: 14:00 - 15:30
- Address: Sokolovská 83, Praha
- Room: K1
- Speaker: Yuxin Ma
Mixed precision algorithms have been developed to utilize several floating-point arithmetic formats. Some of these algorithms aim to use a higher precision to improve the accuracy of an algorithm; others provide results of the same quality as algorithms running in a fixed precision but at a much lower cost. In recent years, since lower precision arithmetic is in general faster than higher precision arithmetic especially in GPU, it has attracted a lot of attention to use lower precision to improve the performance.