The iAPX 86/20 (8086 with the 8087 numeric coprocessor) is considered for digital filtering. The advantage in using the iAPX 86/20 lies in the 80-bit width of the 8087 floating-point arithmetic-registers. With such large arithmetic registers, the effects of coefficient roundoff and arithmetic roundoff errors on the filter output are reduced. The price paid for the improved numerical performance is the increased time spent by the system moving data to and from memory. The method of Knowles and Olcayto for measuring the effect of coefficient roundoff is studies in detail. This method is applied to an example filter in order to demonstrate that the iAPX 86/20 can meet filter specifications that the 8086 without the numeric coprocessor (iAPX 86/10) cannot meet.
| Identifer | oai:union.ndltd.org:ucf.edu/oai:stars.library.ucf.edu:rtd-1672 |
| Date | 01 October 1983 |
| Creators | Canright, Robert E. |
| Publisher | STARS |
| Source Sets | University of Central Florida |
| Language | English |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Retrospective Theses and Dissertations |
| Rights | Public Domain |
Page generated in 0.002 seconds