| | 22C:122/55:132, Lecture 13, Spring 2003 |
 | | The advantage of horizontal microprogramming is that the microprogram can exploit a maximal amount of parallelism in the data part of the system, but the price is a requirement for a very large ROM. |
 | | Looking at the logic of most microprograms, it is common to find that few microprogram instructions actually have that many distinct successors because, at most points in the microprogram, many, and perhaps most of the condition tests are don't care conditions. |
 | | This approach sends nonsense constants to the data part when there is a microprogram branch or when the W register is being controlled, so it is important to make sure that none of the multiplexors are set to values that cause the constant to be used. |
| www.cs.uiowa.edu /~jones/arch/spring03/notes/13micro.html (1324 words) |