The Software Choice - (lesson 1.7)

The Software Choice

Until recently the only software choice open to a PC user was witch version of MS-DOS (PC-DOS)to run ! MS-DOS was defanged to work with the 8088/86 and is subject to the well know 648 Kbyte limit on memory . in addition all the programs that run under MS-DOS were written for the 8088/86 As has already been mentioned  the 286 and the 386 can pretend to be high speed version of the 8080/8086 and as such can quite happily run MS-DOS and all the  programs that run under it. if all you want to do is run MS-DOS and MS-DOS performance is processes clock speed and in this case you could do just as well with a 286 system with a fast enough clock. However there are still reasons to prefer a 386 based system and these are concerned with its use of memory.

Memory Manege meant

Even if you intend to say with MS-DOS you can still benefit form a 386 machine br adding a memory manager such as from QEMM or 386MAX or by using EMM386 which is included with MS-DOS 5.The 386 has the ability to rearrange the layout of memory in a machine memory manager can make use of this  ability to move parts of MS-DOS such as network drives etc. out of the usual 640 Kbyte area and so increase the proportion of it that is available to applications programs. This is usually referred to as making use of upper memory blocks(or UMBs) which are discussed in more detail later.

Expended Memory

Expended memory is a type of memory used by order MS-DOS programs needing to use more than 640 Kbyte In non 386 based machine expended memory has to be implemented expanded memory using nothing but software. you can by utilis that do nothing but crate expanded memory-so called Limulators but in  a full 386 memory manger of the type described in the previous situation.

DOS Extenders

Some software Louts 1-2-3 Release 3 being the best known, makes better use of the 386 boys  by incorporating DOS extenders. in this case you can only run the application on a 286 or 386 under MS-DOS but the application has access to all of the memory installed on the machine. In other Words DOS extenders are nothing more than a way around the 640KByte limit.

To get the best out the 386 the is no doubt that you have to move to something beyond MS-DOS.There are varies alternatives some  of witch don't involve a complete abandonment of MS-DOS.

DOS multi-taskers

A DOS multi tracker makes use  of the 386's ability look like more than one complete 8088/86 system run multiples copies of MS-DOS and application. The Best know does  multi tracker are windows 3 and DESK view. both will allow you to run multiple DOS application and switch between them but window 3 goes some what further it also supporters its own type of application that isn't limited by the 640kbyte  barrier.In this sense Windows 3 is hallway to binging a completely new operating system that replaces MS-DOS.

 New operating system

The most attractive solutions from the technical point of view at least is to get rid or MS-DOS and all its limitation and starts again with a more appropriate operating system.this is exactly what Microsoft and IBM intended to happen when they designed OS/2 but for various reasons the first versions of OS/2 were eclipsed by windows 3.only IBM is currently making any effort to develop OS/2 as a challenged to windows 3.OS/2 version 2 was written specifically for the 386 and it will run MS-Dos windows and OS/2 application on a nearly acaule footing Microsoft have more or lees abandoned OS/2 to develop windows NT which is completely new operating system that will run MS-DOS and windows application on a 386 and perhaps even on other  type of prosper. so at the time of writing the composition is between windows 3.1 and OS/2 Version 2 but in the  longer term the battle line -up will changed to be Windows NT versus OS/2.

Intel's 386 ( lesson 1.6 )

Intel's 386?

It is normal practice in the electronics industry to share the design of a chip with another manufacturer so that it can be 'second source'. The argument goes that second sourcing induces confidence that the product has a life beyond the original designer. Intel second sourced the 286 but with the 386,486 and beyond Intel has decided to keep its design to if self by not signing second sourcing deals and by threatening to stop any manufacturer from infringing its patents and copyrights.

For some time this gave Intel a total monopoly on the 386/486 market and resulted in higher prices. Other manufacturers had to work hard to construct their own 386 designs from scratch but eventually they did. The availability of 386 chips from sources other than Intel had to two effects. The first  is that the price of the 386 fell rapidly and the second is that Intel shifted its emphasis onto creating  special purpose and enhanced versions of the 386 and 486 for which other manufacturers had no equivalents.

Some of the marketing ploys used by Intel are bound to confuse and puzzle the onlooker unless they are seen in the context of trying to keep customers loyal to Intel. For example, the 486SX is a lower specification 486DX but Intel only introduced it as a way of drawing customers away from the newly introduced lower cost 386DX chips produced by other manufacturers. In fact the early 486SX chips were simply full 486 chips with the numeric co-processor section disabled! The 486SX quickly settled down to look more like a respectable member of the family but the fact still remains that Intel wouldn't have  produced it if there had been no competition. Another interesting twist in the 486SX story is that in order to add a numeric co-processor, the chip that you install next to it, the 487SX,is actually a full 486DX processor! In this situation the original 486SX sits in its socket doing nothing much at all.

In an affords to turn the 'empty socket' next to the 486SX into a marketing advantage, Intel invented the idea of offering users an upgrade chip which they could fit in place of the 487SX to boost performance still further. The increased performance is achieved by a technique called 'clock doubling'. This is explained In more detail in the next chapter but essentially it succeeds in doubling the rate at which the processor works without the need to make any changes to other parts of the machine. This  is such a good idea that Intel extended it to the other members of the 386 family so, for example, you can now buy the 486DX2 which is identical to the 486DX but runs twice as fast.

Intel's competitors are also producing new and improved versions of the 386 family. There are already faster 386DX and 386SX chips from sources other than Intel. At the time of writing the first non-Intel 486 chips have arrived on the market. There are also likely to be some interesting  innovations. For example, the chip manufacturing company calyx has just announced the 486SLC which is a 486-like processor that can be used to replace a 386SX with a claimed 20% to 40% speed improvement. Of  course many of these innovations are being challenged by Intel as patent infringements and there are many law suits pending.  

From Chips to Systems ( lesson 1.5 )

From Chips to Systems

The 386/486 is a powerful devise but tapping its power when making the transition into complete system poses a number of problem. when the 286 has introduced IBM modified. the existing PC architecture to produce the  At. The modifications were fairly small and mostly upwards commutable with the original PC design. in other words , apart from few  well understood expectations what would work with a PC would with an AT-hardware and software alike. The at quickly became a de factor Stranded to witch other personal computer manufactures conformed.

with the introduction of the 386 similar modification to the AT architecture seemed to be  called for-after all if there was a change in going from 16 to 32 bits .IBM took this apparent  need for changed as a way of trying to gain control of the PC market. Instate of basing their new design on the existing AT the chose to produce something quite  new and different - the PC/2.

Wail the PS/2 range is closely related to the at design in that Nelly all software that will run on an At will run on a PS/2 it is incomparable at the hard wear level. the PS/2 range uses a MCA (MICRO CHANEL ARCHITECTURE) to connect additional hardware instead of the ISA (INDUSTRY STANDARD ARCHIVE in the AT .MCA has lot of performance advantages but IBM chose to make it a proprietary technology and so most  other manufactures were deterred from using it.

 the alterative approach  to the PS/2 was to simply use a 3869 in place of a 286 in a traditional AT design , modified as little as possible. Initially the problem was that different manufacturers made the necessary modifications in slightly different ways. In the main this didn't matter to much and only affected the way that additional memory was added. As memory technology progressed even this difficultly tended to vanish and a 386 design that is from the users point of view, identical to an AT evolved.

This at / 386 design has all the advantages of being hardware commutable with the original AT but has the disadvantage of not begins able  to deal with high speed add-on cards. in other words ISA is not as fast or as flexible as MCA. Whether this is a real problem or not depend very much on the application your interested in. there is an argument that ISA is a  used to its full capacity in most system anyway! Even so this Christmas resulted in a group of manufacturers defining a new improved version of ISA-EISA (EXTENDED INDUSTRY STANDARD ARCHITECTURE).EISA offers many of the advantages of MCA but it is upward comparable with ISA and so you can use existing AT style  hardware with it.

Much of the fuss concerning which type mo machine to buy is irrelevant. The key fact to remember is that the effect of ISA, MCA or EISA depend on what add-on you require and in ,many cases it isn't critical see chapter 3.For example , if you are plugging in a card to provide an additional printer port then  from a performance point of view it doesn't matter which bus your are using because they are all fast enough to cope with the date rate of printer. On the other hand if you want to contact a high speed disk drive then you might need MCA or EISA to achieve the performance you desire.

The situation isn't quite as clear cut as if you need high speed disk  you need MCA or EISA'.In practice the speed disk using ISA can be improved by selecting IDE , SCSI or ESDI drives (Chapter 6) and this might be all that is necessary for the application.   Another confusing    factor is that one of the main task for which    MCA was designed that connecting additional memory that will work at the full speed of the processor has largely become redundant. part of the reason has been  the introduction of spacial add-on brad with high speed  connection to the processor just for memory  extinction but the main reason is that the amount  of memory that can be fitted to the main system bored has increased to the point where memory  extensions seam an unnecessary concept!

The pros and cons of the different architecture can be summaries as.

ISA 386-AT - low cost, lots of exiting add-on cards but limited in its maximum performance.

EISA 386 AT- Higher cost not many add-on cards as yet but upward commutable with ISA, higher potential preference, not backed bu IBM.

MCA PS/2            - Higher cost higher potential performance backed by IBM.

As thing have turned out the dominate architecture has proved to be the ISA -386 The raisin is that for most application this provides more than enough power and flexibility. in some  specialist striation the EISA 386-AT is an advantaged and this are discussed latter in this book. the MCA PS/2 has proved to be of intrust only to dedicated IBM coustemers It is to early to say that the MCA PS/2 detained is unless there the reticule and unforeseen changed in the PC market the MCA bus a and the PS/2  have to be seen as another failed attempts at forcing a revolution on the PC using community .This isn't to say that the EISA 386AT can be hailed as the victory of evolution over revolution. As early day stated the plain facts is that the majority of 386-AT designs use and need nothing more sophisticated than the ISA bus. all of these topics are taken up in more detail in the next chapter.       

The Legacy Of Upward Compatibility (lesson 1.4)

The Legacy Of Upward Compatibility

The Evaluation Of The PC has certainly made the transition from the under powered mashies of the 80s ton the desktop mainframes of the 90s but it has its disadvantages as well. The truth of the matter is that the 386/486 provides facilities that take into the realm of the work station but for most of the time it is begins treated as faster version of the original 8088!In the other words , most of its power is simply igorned. This is Legacy of  upward compatibility. Each new generation of processor has add features and facilities but not at the expense of compatibility with earlier processor.  

The first improvement on the 8088 can be  considered to be the 8086 even though  they were crate in the sane periods of time. the 8086 is a 16 bit version of the 8-bit 8088 and , from a software writer's points of view, it is identical but faster .As a result program that run of the 8088 will, without modification, run on the 8086 only faster. what is more there are mor modifications that can be made that will make a program designed for the 8088 run any faster on an 8086. Another way of looking at this is that programe written for 8086.In of looking at this is that a program written for the 8086 processors could be moved back down to the less powerful 8088 without making any changes. In this sense the 8086 is also downward, or backward , compatibility.  

In the Case of 80286,the next prossecer that Intel produced, the story is very different. Although the 80286 is a 16-bit prosier like the 8086,it contains many advanced modes of apparition not sherd by the 8086.You can Take an 8086 program and run without modification on the 8028 but the program will not make any use of the 80286's new futures .As you might gussied and programed that is written for the 80286that does make use of its special futures cannot be moved down to and 8086 processors without extensive changes. In other words , the 80286 upward but not dawn wards computable with the 8088 and 8086.The trouble with this situation is that very few software producer took the tuber or risk pf producing software specially 80286 prospector and so much of the power of this devise has remained well and truly hidden.

The same situation holds with the 80386/486.it is upward compatible with the original 8088 and 8086 processors and that is how much current software treats it is as a fast 8088/86.The only difference between this and the 80286 is that now hidden power is very great indeed.  If you buy small motor boat  and insist on getting about by peddling it from one place to another then people will think that your are eccentric , but if you buy a high speed jet foil and treat it in the same way people we deiced that your are made .the same is true of the current states of the computing. while it might have been fine to paddle the 80286, the 80386/486 deserves to be use fully. in latter chapters the topic of how the extra power can be released is discussed in detail.    

The false start ( lesson 1.3 )

The 80286 – The false start 

When inlet produced the 80286 it was to have been the first processor to put the power of a mainframe on your desktop . IBM even emphasized the point by calling the machine they designed to follow the PC the AT – standing foe Advanced Technology . Although the design of the AT took the existing PC architecture and adapted it more towards a true 16 bit form, from the user’s point of view it made little difference. Most, if not all, of the software available until recently treated the AT as if it was a faster version of the PC and there was no software designed to take advantage of its additional powers.
Even so the AT- proved popular and with falling prices became the standard machine for serious work. The irony is that just at the time that the 386/486 became available at reasonable prices, the first software designed for the specially for 286 began to make an impact – foe example OS/2 , Lotus 1-2-3 Release 3 and Windows 3. Of course this software will run perfectly well on a386/486 but it doesn’t use them to the best effect.
In many ways it is best to see the 286 as a false start on the road to a desktop mainframe. It had many flaws that forced software designers to have to compromise and ‘think small’ . Foe example, the 286 still retains some of the 64Kbyte (Kilobyte) Imitations of the original 8088. In a286 system you 64Kbyte chunks or segments. In a 386 system each segment can be as large as 4 Gbytes (Gigabytes) and this allows programmers to work without restriction . It als0 makes it mush easier to move existing mainframe programs to 386’s abilities quickly leads to the notion that the 386 is the processor that the 286 should have been and the sooner the 286 dies out the better
Unfortunately the 286 still has a large , although no longer growing , user base . Even today manufactures are still tempting the occasional user to buy a 286by very low prices. There is also still a luxury and only necessary foe the most demanding application or user . In support of this notion it is frequently said that a 286 will run some applications as fast as a 386 and is cheaper. The error in this argument is that the performance comparisons are based on treating both machines. As if they were just fast 8088/ 86 machines. In other words, if you don’t use any of the special features of the 386 but make them both pretend they are the older 8088/ 86 chip, there isn’t much to chose between them !
IT is already quite clear that the 386 is the dominant processor and will remain so for some years in one from or another . At the moment the penalty for owning a 286 is mainly to be excluded form taking advantage of the 386 –specific modes of some software that will refuse to run on a 286, there is an increasing amount of software that doesn’t perform to its best on a 286 . In the very near future you can expect both operating systems and application software to abandon the need to be 286 compatible and move to a ‘true 32bit – 386’ from . when this happens the 386 will no longer be treated as just a fast 286 with one or two special extras by the software that runs on it but as the full 32-bit powerful processor that it really is. This will reveal the gulf between the 286 once and for all.
To makes this more concrete simply consider Windows . Windows 3 and even Windows 3.1 were both written for the 286 but with extra capabilities to make use of some of the features of the 386. The next version of Windows on the other hand, Windows NT, is being designed specifically fo the 386 family and will not make any attempt to work with the 286. You can see that the very existence of the 286 and the need to be compatible with it has held back progress.

Evolution not revolution ( lesson 1.2 )

 Evolution not revolution

The introduction of the 386 family has brought about , and is still bringing about , and is still bringing about , revolutionary changes in the ways that we use PCs , but much of its success is the way that it softened this revolution. At many stages in the history of computing a revolution in computing power has come about as users have abandoned their existing hardware and software and adopted new. The first generation of almost practical personal computers were based on the Z80 and an operating system called CP/M . At the time it looked as though the status Quo would last a thousand years but this entire generation of machines was swept away almost overnight when IBM announced the IBM PC in 1981. This was based on a completely different processor , the intel 8088, and a completely new operating system , Ms – Dos . while not perfect this really did offer a performance that allowed personal computing to be taken seriously. Even at the time of its introduction the design of the PC was considered old fashioned and at best conservative – still it resulted in millions of machines being scrapped.
This successful revolutionary change gave rise to the idea that users could be made to change their hardware and software if the benefits were large enough. In practice , and despite the best attempts of IBM and others, this has not force them to throw out their existing hardware and software. now it looks as though the size of the commitment to the PC is so great that all future changes will be a gradual step-by-step evolution . Although it has been extended, the original PC design is still very evident in the highest performance 386/486 based machines . obtainable . There is the argument that we would have reached the current level of computing power much sooner if the Pc architecture had been abandoned and a completely new start made.