The History Of Intel CPUs (Archive)

[bit_user]

there are details about essentially every processor on the list that isn't in this piece. This is due to the type of article this is. It is designed to be a brief overview of Intel's processors, give you a look at them, and address the key points and changes from one CPU to another.

I understand that there will always be details people think you should've included, but there are some significant factual errors you should correct, in order to avoid misinforming readers:

• 
  ■ 8086 could not "simultaneously work on two eight-bit instructions".
  ■ 8088 was a full 16-bit CPU with a half-width bus. It split 16-bit bus transactions into two 8-bit operations. It was simply a lower-cost, lower-performance version of the 8086, and fully instruction-compatible.
  ■ 80386 SX was the same thing - a lower-cost, lower-performance version of the 386 DX. For all practical purposes it was a fully 32-bit CPU, and instruction compatible with the 386 DX. 386 SL was a laptop-oriented version of the 386 SX.
  ■ Prescott Pentium 4 wasn't 64-bit. Rumored to have support in hardware. Even if it did, it was never enabled. Pentium D was their first x86-64 desktop CPU.

I don't ever want to see any of these show up in wikipedia entries, citing your article as the source.

IMO, these are the most important details omitted, which support the narrative of how x86 evolved:

• 
  ■ 286 introduced memory protection (i.e. protecting one program's memory from others).
  ■ 486 was the first pipelined x86 CPU.
  ■ Pentium was the first super-scalar x86 CPU.
  ■ Original Atom was hyperthreaded, in an attempt to compensate for being in-order.
  ■ Itanium & IA64.

For further details, see my posts.

 

[IInuyasha74]

• 
  ■ 8086 could not "simultaneously work on two eight-bit instructions".
  ■ 8088 was a full 16-bit CPU with a half-width bus. It split 16-bit bus transactions into two 8-bit operations. It was simply a lower-cost, lower-performance version of the 8086, and fully instruction-compatible.
  ■ 80386 SX was the same thing - a lower-cost, lower-performance version of the 386 DX. For all practical purposes it was a fully 32-bit CPU, and instruction compatible with the 386 DX. 386 SL was a laptop-oriented version of the 386 SX.
  ■ Prescott Pentium 4 wasn't 64-bit. Rumored to have support in hardware. Even if it did, it was never enabled. Pentium D was their first x86-64 desktop CPU.

1. 8086 - I will check into this further.
2. 8088 - It is an 8-bit processor. Yes, it could execute 16-bit instructions by breaking them and executing them as two 8-bit operations. The 8-bit 8080 was capable of this also, but no one questions that the 8080 is an 8-bit processor. That is because we refer to a processor as "8-bit" or "16-bit" etc. by the width of the execution unit. Technically it contains a 16-bit execution unit, but if the execution unit is only capable of executing code in 8-bit chunks and is thus limited to 8-bit operations, then it is an 8-bit processor.

For a similar example we could look at other processors from both AMD and Intel that use a 64-bit core, but for marketing purposes were limited to 32-bit operations. We still called these processors 32-bit processors. There are dozens of examples of this.
3. 80386: Same as above.
4. There were Prescott Pentium 4 processors that supported 64-bit operations.:
http://ark.intel.com/Products/Spec/SL88G
http://ark.intel.com/Products/Spec/SL8U5
http://ark.intel.com/Products/Spec/SL8JA

As evidence, here are three links to Prescott Pentium 4 products listed on Intel's website that indicate they are Intel 64 compliant. Now it is true that the bulk of Prescott Pentium 4 CPUs did not. These 64-bit compliant versions use a special E0 core stepping.

 

[shrapnel_indie]

1. 8086 - I will check into this further.
2. 8088 - It is an 8-bit processor. Yes, it could execute 16-bit instructions by breaking them and executing them as two 8-bit operations. The 8-bit 8080 was capable of this also, but no one questions that the 8080 is an 8-bit processor. That is because we refer to a processor as "8-bit" or "16-bit" etc. by the width of the execution unit. Technically it contains a 16-bit execution unit, but if the execution unit is only capable of executing code in 8-bit chunks and is thus limited to 8-bit operations, then it is an 8-bit processor.

The 8088 came after the 8086. the 8088 was internally exactly the same (registers, instruction set, I/O, address bus) other than the 8088 had its data bus pared down to 8-bits instead of 16 bits. ALso the pre-fetch instruction queue was two bytes smaller on the 8088 (4 vs 6.) They run the exact same byte code, running practically the same support chips/hardware, except the 8088 used an 8-bit data bus instead of 16, and so there were less lines to demultiplex on the 8088 compared to the 8086.

Although the 8086 was a great chip, it was expensive at the time and more importantly required an
expensive 16-bit support chip and board design. To help bring costs down, in 1979, Intel released a
crippled version of the 8086 called the 8088. The 8088 processor used the same internal core as the
8086, had the same 16-bit registers, and could address the same 1MB of memory, but the external
data bus was reduced to 8 bits. This allowed support chips from the older 8-bit 8085 to be used, and
far less expensive boards and systems could be made. It is for these reasons that IBM chose the
crippled chip, the 8088, for the first PC

- source: http://vig.prenhall.com/samplechapter/0789725363.pdf , page 3 of that sample chapter (Chapter 3).

Other sources:
http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-823-computer-system-architecture-fall-2005/lecture-notes/l15_micro_evlutn.pdf
http://ece-research.unm.edu/jimp/310/slides/8086_chipset.html

 

[epobirs]

Worth noting that the Pentium III and P-III based Celerons did not use the same cache structure. The difference was more than just the size of the cache. This was important as it further reduced performance on the Celerons to reduce cost. The Xbox used the same cache as the P-III but only had 128K. Thus the Xbox CPU was NOT a Celeron but rather a slightly customized Pentium III.

 

[bit_user]

2. 8088 - It is an 8-bit processor. Yes, it could execute 16-bit instructions by breaking them and executing them as two 8-bit operations. The 8-bit 8080 was capable of this also, but no one questions that the 8080 is an 8-bit processor. That is because we refer to a processor as "8-bit" or "16-bit" etc. by the width of the execution unit. Technically it contains a 16-bit execution unit, but if the execution unit is only capable of executing code in 8-bit chunks and is thus limited to 8-bit operations, then it is an 8-bit processor.

I don't have any knowledge of the 8080, so I won't comment on that.

I think you're missing a key concept, which is the distinction between instruction execution and memory bus transactions. Not every x86 instruction involves memory. If an instruction operates on 16-bit register-only operands, then the 8088 can execute the whole thing in one go. And if one or both of the operands is in memory, then only the the memory read/write stage gets split - not the full instruction.

Furthermore, if one were trying to operate on 16-bit data, using 8-bit instructions, then it often takes more than 2 instructions to do the same thing. So, it wouldn't even be correct to say that a 16-bit instruction is split into two 8-bit ones.

But I don't care what size data bus you put on a CPU - if all its instructions can natively support 16-bit operands and it supports at least 16-bit addressing, then it's a 16-bit architecture. Conversely, a CPU that has a dual channel memory interface doesn't suddenly become 128-bit.

 

[JonDol]

Otherwise, I like the new version of the THW web site: there is a lot less crap to be blocked by my script/add blocker but I greatly miss the print feature which could put on a single page a whole article.

You know they make their money from ads, right? And more page views = more ads. But some of their articles (not this slide show, I think) are using a more mobile-friendly format.

I understand they make a living using the ads but this is way too much: it simply kills the website which otherwise I greatly appreciate.

Yesterday again I tried to read a review on this site without the add/script blocker and it was IMPOSSIBLE: between the animated ads which slow down my PC to a snail and the refresh of the add at the top of the page which also makes the page to automatically jump at its start while you are nearly at the end it was too much for me just to read some poor 20 lines of the article. This is to say that for a second of delay I'm able to quit a website and go read the same news on another faster websites so I let you imagine what I do when a website makes me waste minutes...

THW tried a few years ago to make us ingurgitate the ads using a Fossil Networks service which roughly pretended that add blockers are malware and refused to show the content and then I stopped reading THW for a few months. When I came back the Fossil Networks was gone...

Cheers

 

[bit_user]

Yesterday again I tried to read a review on this site without the add/script blocker and it was IMPOSSIBLE: between the animated ads which slow down my PC to a snail and the refresh of the add at the top of the page which also makes the page to automatically jump at its start while you are nearly at the end

I'm definitely with you, on that. I don't know how much they can do, but there's got to be a way to use ads that aren't so obnoxious they either drive users away or to use blockers.

 

[Redmatt]

No reference to Itanium ???