10.1 The Importance of the Design of the Instruction Set

The design of the CPU's instruction set is one of the most interesting and important aspects of CPU design. The instruction set architecture (ISA) is something that a designer must get correct from the start of the design cycle. Features like caches, pipelining, superscalar implementation, and so on, can all be grafted on o a CPU long after the original design is obsolete. However, it is very difficult to change the instruction set once a CPU is in production and people are writing software using those instructions. Therefore, instruction set design requires very careful consideration.

You might be tempted to think that the kitchen sink approach to instruction set design (as in 'everything, including the kitchen sink'), in which you include every instruction you can dream up, is best. This approach fails for several reasons, as you'll soon see. Instruction set design is the epitome of compromise management. Good instruction set design involves selecting what to throw out rather than what to keep. It's easy enough to say, 'Let's include everything.' The hard part is deciding what to leave out once it becomes clear that you can't have every instruction on the chip. Why can't we have it all? Well, in the real world some nasty realities prevent this:

Nasty reality 1: Silicon real estate The first problem with 'putting it all on the chip' is that each feature requires some number of transistors on the CPU's silicon die (chip). CPU designers work with a 'silicon budget' and are given a finite number of transistors to work with. This means that there aren't enough transistors to support ' putting every possible feature' on a CPU. The original 8086 processor, for example, had a tran-sist or budget of fewer than 30 thousand transistors. The Pentium III processor had a budget of over 8 million transistors . These two budgets reflect the differences in semiconductor technology in 1978 versus 1998.

Nasty reality 2: Cost Although it is possible to use millions of transistors on a CPU today, the more transistors that are used the more expensive the CPU becomes. For example, at the beginning of 2004, Pentium IV processors using millions of transistors cost hundreds of dollars. Con-temporary CPUs with 30,000 transistors cost only a few dollars.

Nasty reality 3: Expandability One problem with the kitchen sink approach is that it's very difficult to anticipate all the features people will want. For example, Intel's MMX and SIMD instruction enhancements were added to make multimedia programming more practical on the Pentium processor. Back in 1978, when Intel created the first 8086 processor, very few people could have possibly anticipated the need for these instructions. Therefore, a CPU designer needs to allow for making extensions to the instruction set in future members of the CPU family to handle unanticipated needs.

Nasty reality 4: Legacy support This is almost the opposite of expand-ability. Often, an instruction that the CPU designer feels is important turns out to be less useful than anticipated. For example, the loop instruction on the 80x86 CPU sees very little use in modern high-perfor-mance programs. The 80x86 enter instruction is another good example. When designing a CPU using the kitchen sink approach, it is common to discover that most programs never use some of the instructions. Unfortunately, removing instructions from later versions of a processor will break existing programs that use those instructions. Generally, once an instruction is added to the instruction set, it will have to be supported in all future members of the processor. Unless very few programs use the instruction, and CPU designers are willing to let them break, removing instructions is a difficult thing to do.

Nasty reality 5: Complexity The popularity of a new processor is easily measured by how much software people write for that processor. Most CPU designs die a quick death because no one writes software specific to that CPU. Therefore, a CPU designer must consider the assembly programmers and compiler writers who will be using the chip upon introduction. While a kitchen sink approach might seem to appeal to such programmers, the truth is that no one wants to learn an overly complex system. A CPU that does everything under the sun appeals mostly to someone who is already familiar with the CPU. However, pity the poor soul who doesn't know the chip and has to learn it all at once.

These problems with the kitchen sink approach all have a common solution: design a simple instruction set for the first version of the CPU, and leave room for later expansion. This is one of the main reasons the 80x86 has proven to be so popular and long-lived. Intel started with a relatively simple CPU and figured out how to extend the instruction set over the years to accommodate new features.