17.05.2015 Views

16-Bit Microprocessor Handbook

16-Bit Microprocessor Handbook

16-Bit Microprocessor Handbook

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Thus a low-to-high transition of one signal triggering changes in two other signal levels would be illustrated as<br />

follows:<br />

10) All signal level changes are shown as square waves. Thus rise and fall times are ignored. These times are given in<br />

the data sheets which appear at the end of every chapter.<br />

INSTRUCTION SET CONVENTIONS<br />

Every microcomputer instruction set is described with two tables. One table identifies the operations which<br />

occur when the instruction set is executed, while the second table defines object codes and instruction times.<br />

Because of the wide differences that exist between one instruction set and another, we have elected not to<br />

use a single set of codes and symbols to describe the operations for all instructions in all instruction sets. We<br />

believe any type of universal convention is like to confuse rather than clarify; therefore each instruction set<br />

table is preceded by a list of symbols as used within the table alone.<br />

A short benchmark program is given to illustrate each instruction set. Some comments regarding benchmark<br />

programs in general are, however, in order. We are not attempting to highlight strengths or weaknesses of<br />

different devices, nor does this book make any attempt to comparative analyses, since the criteria which make<br />

one microcomputer better than another are simply too dependent on the application.<br />

Consider an application which requires relatively high speed processing. The only important criterion<br />

will be program execution speed, which may limit the choice to just one of the microcomputers<br />

we are describing.<br />

COMPARATIVE<br />

ANALYSIS<br />

Execution speeds of all of the microcomputers may, on the other hand, be quite adequate for a second application; in<br />

this case, price may be the only overriding factor. In a third application, a manufacturer may have already invested in a<br />

great deal of engineering development expense, using one particular microcomputer that was available in quantity earlier<br />

than any others; the advantages or disadvantages of using a different microcomputer, based on minor cost of performance<br />

advantages, will likely be overwhelmed by the extra expense and time delays involved with switching in<br />

midstream.<br />

And what about benchmark programs 7<br />

There have been a number of benchmark programs in the literature, purporting to show the<br />

strengths or weaknesses of one microcomputer versus another; individual manufacturers<br />

have added to the confusion by putting out their own competing benchmarks, aimed at showing their product to<br />

be superior to an immediate rival.<br />

Benchmark programs are misleading, irrelevant and worthless for these reasons:<br />

1) In a majority of microcomputer applications, program execution speed, and minor variations in program<br />

length, are simply overwhelmed by pricing considerations.<br />

2) Even assuming that for some specific application, program length and execution speed are important, trivial<br />

changes in the benchmark program definition can profoundly alter the results that are obtained. This is one<br />

point we will demonstrate in this book, while describing individual instruction sets.<br />

3) Benchmark programs are invariable written by the smartest programmers in an organization, and they take<br />

an enormous amount of time to ensure programming accuracy and excellence. This is not the level at which<br />

any user should anticipate "run of the mill" programmers working; indeed, a far more realistic evaluation of<br />

a microcomputer's instruction set could be generated by giving an average programmer too little time in<br />

which to implement an incompletely defined benchmark. This will more closely approximate the working<br />

conditions under which real products are developed. Of course, defining the "average programmer," "too<br />

little time" and an "incomplete specification" are all sufficiently subjective that they defy resolution.<br />

xiv

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!