Chapter06Rev07 - Chapter 6 (Revision number 7) Rudiments of...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Chapter 6 (Revision number 7) Rudiments of Pipelined Processor Design 6.1 Introduction Welcome to Bill’s Sandwich shop! He has a huge selection of breads, condiments, cheese, meats, and veggies to choose from in his shop, all neatly organized into individual stations. When Bill was starting out in business, he was a one-man team! He took the orders, went through the various stations, and made the sandwich to the specs. Now his business has grown. He has 5 employees one for each of the five stations involved with the sandwich assembly: taking the order, bread and condiments selection, cheese selection, meat selection, and veggies selection. The following figure shows the sandwich assembly process. Station 1 station II station III station IV station V (place order) (select bread) (cheese) (meat) (veggies) New (5 th order) 4 th order 3 rd order 2 nd order 1 st order Each station is working on a different order; while the last station (station V) is working on the very first order, the first station is taking a new order for a sandwich. Each station after doing “its thing” for the sandwich passes the partially assembled sandwich to the next station along with the order. Bill is a clever manager! He carefully chose the work to be done in each station to be roughly the same, so that no employee is going to be twiddling his thumbs. Of course, if a particular sandwich order does not need a specific ingredient (say cheese) then the corresponding station simply passes on the partially assembled sandwich on to the next station. Nevertheless, most of the time (especially during peak time) all his employees are kept busy rolling out sandwiches in rapid succession. Bill has quintupled the rate at which he can serve his customers! 6.2 Towards an instruction processing assembly line You can see where we are going with this…in the simple implementation of LC-2200, the FSM executes one instruction at a time taking it all the way through the FETCH, DECODE, EXECUTE macro states before turning its attention to the next instruction. What is wrong with that? Well, if we look at the datapath for LC-2200, we can see that not all the hardware resources are used all the time in each of the states that the FSM takes an instruction through. Figure 6.1 shows the parts of the datapath that are in use for each of the macro states for a couple of instructions:
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Macro State Functional Units in Use _______________ FETCH IR ALU PC MEM DECODE IR EXECUTE (ADD) IR ALU Reg-file EXECUTE (LW) IR ALU Reg-file MEM Sign extender Figure 6.1: Datapaths in use for different macro states We can immediately make the following two observations: 1. The IR is in use in every macro state. This is not surprising since IR contains the instruction, and parts of the IR are used in different macro states of its execution. IR is equivalent to the “order” being passed from station to station in the
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 11/25/2010 for the course CENG 100 taught by Professor Ceng during the Spring '10 term at Universidad Europea de Madrid.

Page1 / 32

Chapter06Rev07 - Chapter 6 (Revision number 7) Rudiments of...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online