EmbeddedRelated.com
Forums
Memfault Beyond the Launch

x86 architecture concepts

Started by Mouarf February 23, 2005
" why it is so different from other
architecture like ARM, PowerPC, alpha....? What are the main markets these
architecture target to?"

Those 2 questions are real questions and if I would have been qualified I 
would certainly not ask questions since I'd know the answers.

"your question appears to indicate you may not be qualified
 for the work you are doing."

According to you, we must work on this area to be allowed to ask 
questions??????????? I'm only curious about what differs from an 
architecture to another, and what an "architecture" exactly means.

Anyway, if I ask unreal questions, consider that you have no need to answer. 
"Google is your friend" is the best answer when someone is able to answer 
the question.

I really wonder what usenet is for...........

"Nicholas O. Lindan" <see@sig.com> schrieb im Newsbeitrag 
news:Eb2Td.5110$Ba3.1246@newsread2.news.atl.earthlink.net...
> "Mouarf" <mouarf@chezmoi.fr> wrote in > >> If you do not have better answers, please don't answer or answer google >> with >> the relevant keywords.... > > No, Mouarf, you have it arse-backwards. It is up to _YOU_ to provide > real questions, not your homework assignment. If this isn't your homework > question, then your question appears to indicate you may not be qualified > for the work you are doing. > > Go google -- what you _need_ to learn is how to limit searches so you > only get the information you want. > > -- > Nicholas O. Lindan, Cleveland, Ohio > Consulting Engineer: Electronics; Informatics; Photonics. > To reply, remove spaces: n o lindan at ix . netcom . com > psst.. want to buy an f-stop timer? nolindan.com/da/fstop/
Mouarf wrote:
> according to you, x86 architecture is only an instruction set?
Please don't toppost. Given that an instruction set defines all of the visible registers and operations, what else do you need to define an architecture? Kelly
"Kelly Hall" <khall@acm.org> schrieb im Newsbeitrag 
news:Un2Td.6505$OU1.1819@newssvr21.news.prodigy.com...
> Mouarf wrote: >> according to you, x86 architecture is only an instruction set? > > Please don't toppost. > > Given that an instruction set defines all of the visible registers and > operations, what else do you need to define an architecture? > > Kelly
visible registers and operations refers to software side, is there a hardware side hidden in the word "architecture"? Comparing to ARM architecture which is known to be lower power than x86 (at equivalent processing power), I really believe that the different instruction set registers are not the cause of this power consumption difference.
On 2005-02-23, Mouarf <mouarf@chezmoi.fr> wrote:

> Comparing to ARM architecture which is known to be lower power > than x86 (at equivalent processing power),
Non-sequitor. Certain ARM _implimentations_ are lower in power consumption than certain equivalent-throughput IA32 _implimentations_.
> I really believe that the different instruction set registers > are not the cause of this power consumption difference.
And I believe I'm rich and famous. Some architecures are simpler to impliment than others and require lower clock rates and fewer gates to acheive a given throughput. Lower clock rates and fewer gates generally means lower power dissipation. Low power, high throughput, and simple implimentations were just not critera for the IA32 architecture. They _were_ criteria for the ARM architecture (and for most other sane ones). Therefore it's much easier for an ARM implimentation to acheive decent throughput with low power consumption. The IA32 architecture is a huge pile of kludges on top of a base of bogus backwards-compatibility requirements. Nobody gave a shit that the 8086 mnemonics and register layout were reminiscent of (but incompatible with) the 8080. But, Intel sacrificed all sorts of practical things to acheive that stupid, useless familiar-but-not-compatible relationship between the 8080 and the 8086. It was a horrible decision for which the personal computer industry has suffered immeasurably for the past 20 years. -- Grant Edwards grante Yow! I'll show you MY at telex number if you show visi.com me YOURS...
On Wed, 23 Feb 2005 17:27:55 +0100, Mouarf <mouarf@chezmoi.fr> wrote:

> I'm not sure, see the PXA processor from Intel that use ARM > architecture........
Absolutely. PXA is based on XScale (micro)architecture, which is in turn "ARM Architecture 5TE", by ARM nomenclature. Your original question was about x86, which is completely different can of worms. It all has started the day Intel decided to ..ermm.. "extend" their i8080 processor and make it 16-bit. The new i8086 processor had inherited the odd registers functionality and names. Subsequent family members (286, 386, 486, 586 aka Pentium, 686 aka PentiumII) just added to the mess. Due to the numbering convention this whole line is referred to as "x86". As Intel produces Itanium (or do they still?), the "x86" get another name, "IA32". Itanium, for it's 64-bit nature, is named "IA64". And now ask somebody about i8035 or i8051 :) HTH, Vadim
*** Rude topposting fixed ***
Mouarf wrote:
> "David" schrieb im Newsbeitrag >> Mouarf wrote: >>>
... snip ...
>>> >>> If you also could give me a good link to these architectures too >>> I would be really pleased. >> >> Try: www.google.com >> >> Once you have more specific questions, try again, but in a relevant >> newsgroup. >> >> And if it is homework (as this appears to be), say so - there are >> plenty of people on usenet willing to help you learn, but very >> few willing to do your homework for you. > > If you do not have better answers, please don't answer or answer > google with the relevant keywords.... > > If you feed google with words around x86 you get a jungle of > irrelevant pages. My last homeworks date very far away and I'm > sorry to inform you that the question is not trivial. Even if I > was a student, of course all answers are available on the web, > in this case, why usenet exists? > > I asked about the concepts that characterize the x86 > architecture (and also ARM, alpha...) not the instruction set > and not the vendors of x86 stuff like google like to give. > > I other terms, what x86 means? Idem for PowerPC, alpha. > > Is this just a brand of Intel? > > Is it a core that use the same instructions as the Intel 86 > processors? > > When we look for low power 32bit &#4294967295;C, why don't we say first x86 > rather than ARM (what makes ARM architecture low power comparing > x86)?
Please don't toppost. Your answer goes AFTER (or intermixed with) the quote, with irrelevant material snipped out. This makes the article a coherent whole. If and when you are accused of being off-topic, it is better to explain the misunderstanding (if any) and accept the resultant consensus than to come back with a snippy answer which is sure to offend someone. I already gave you an answer in the light of the history of the X86 chips. The X86 chip line has done a remarkable job of expanding capabilities in a downward compatible manner. This is not really conducive to power consumption reduction. Chips such as the ARM and the TI 430 are not saddled with these requirements, and can concentrate much better on the power problem. Todays X86 chips are routinely dissipating in the order of 100 Watts, while other systems are in the milliwatt region. Not too long ago the idea of dissipating 1/2 watt in a CPU was already ridiculous. -- Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net) Available for consulting/temporary embedded and systems. <http://cbfalconer.home.att.net> USE worldnet address!
> Some architecures are simpler to impliment than others and > require lower clock rates and fewer gates to acheive a given > throughput. Lower clock rates and fewer gates generally means > lower power dissipation.
OK I had this idea before (but I wanted to know, I what the x86 architecture is more complex, because there are much more instructions?). Do you have an idea how National Semiconductor succeeded in strongly reducing the power consumption of their Geode processor (now AMD) which is called to be a x86 processor, architecture complexity is the same as Intel's one and the silicon technology not that different.? One other thing, I don't really understand how the Crusoe processors from Transmetta that emulates a x86 processor by a software in RAM are the x86 lowest power consumers. The clock speed is up to 1GHz (I don't know the equivalent frequency for non emulated x86 processors) and the power consumption ridiculous although there is an overhead consumption caused by the software emulation. Do you have an idea about that? Best regards
>Subsequent family members (286, 386, 486, 586 aka Pentium, 686 aka >PentiumII) just added to the mess. Due to the numbering convention this >whole line is referred to as "x86".
OK. You mean that the only way for non Intel processors (AMD's one, Crusoe, Geode) to be called x86 is to have the same instruction set and main registers as Intel's x86? Wasn't there any patent on the Intel x86 instruction set that did not allow any other company to build x86 processors?
On 2005-02-23, Mouarf <moi@chezmoi.fr> wrote:

>> Some architecures are simpler to impliment than others and >> require lower clock rates and fewer gates to acheive a given >> throughput. Lower clock rates and fewer gates generally means >> lower power dissipation. > > OK I had this idea before (but I wanted to know, I what the > x86 architecture is more complex, because there are much more > instructions?).
More instructions, more complex instructions, registers with special meanings, etc.
> Do you have an idea how National Semiconductor succeeded in > strongly reducing the power consumption of their Geode > processor (now AMD) which is called to be a x86 processor, > architecture complexity is the same as Intel's one and the > silicon technology not that different.?
IIRC, the Geode uses fewer transisters and a simpler microcode engine that's optimized for the frequently used IA32 instructions. Less frequently used instructions take a bit of a performance hit. For a given clock speed the Geode is quite a bit slower than a Pentium (maybe half the throughput), but the simpler design results in far lower power. National cared more about power than raw throughput. Intel cares more about raw throughput than power. Intel could probably make a lower power IA32 if that's what the market wanted. However, it's much easier to make a low-power ARM, so if you want low power Intel will justifiably try to sell you a StrongARM part rather than an IA32 part.
> One other thing, I don't really understand how the Crusoe > processors from Transmetta that emulates a x86 processor by a > software in RAM are the x86 lowest power consumers.
Mostly because they actually care about being low power.
> The clock speed is up to 1GHz (I don't know the equivalent > frequency for non emulated x86 processors) and the power > consumption ridiculous although there is an overhead > consumption caused by the software emulation. Do you have an > idea about that?
Have you read the papers Transmeta put out on the topic? -- Grant Edwards grante Yow! Mr and Mrs PED, can I at borrow 26.7% of the RAYON visi.com TEXTILE production of the INDONESIAN archipelago?
On 2005-02-23, Mouarf <moi@chezmoi.fr> wrote:

>>Subsequent family members (286, 386, 486, 586 aka Pentium, 686 aka >>PentiumII) just added to the mess. Due to the numbering convention this >>whole line is referred to as "x86". > > OK. > > You mean that the only way for non Intel processors (AMD's one, Crusoe, > Geode) to be called x86 is to have the same instruction set and main > registers as Intel's x86?
Yes. The proper term for the 32-bit group of the x86 family is IA32. If that's what you're talking about (rather than the 16 bit or 64 bit parts), you should say IA32.
> Wasn't there any patent on the Intel x86 instruction set that > did not allow any other company to build x86 processors?
Other companies built x86 processors. That pretty much answers your question about the existence of _any_ X where X would not allow other companies to build x86 processors. -- Grant Edwards grante Yow! Don't hit me!! I'm in at the Twilight Zone!!! visi.com

Memfault Beyond the Launch