| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
This is somewhat reminiscent of the staggered arithmetic techniques used, for example, to achieve rapid, "almost accurate" dot productsin software with 754 numbers. Is that the underlying sense (that is, roughly, to usetwo floating point datums, or possibly two datums with different precisions) to represent the radius?Yes, this is essentially staggered correction, in that staggered correction interval techniques use an interval to represent the error. However, staggered correction simulates multiprecision by a sum of floats, which is somewhat inefficient compared to true multiprecision arithmetic.
True for higher precision (6- to 8-fold and beyond), for a little higher precision (2- to 4-fold) this technique is faster than multiprecision. -- ===================================================== Prof. Dr. Siegfried M. Rump Institute for Reliable Computing Hamburg University of Technology Schwarzenbergstr. 95 21071 Hamburg Germany phone +49 40 42878 3027 fax +49 40 42878 2489 http://www.ti3.tu-harburg.de and Visiting Professor at Waseda University Faculty of Science and Engineering Shinjuku Lambdax Bldg. 902 2-4-12 Okubo, Shinjuku-ku Tokyo 169-0072 Japan phone/fax in Japan +81 3 5286 3414