Where AMD leads, Intel follows

[Burning Phoenix]

You mean the first microprocessor, right? There were certainly cpus before.

I guess i should of specified the CPU was a single chip on silicone. "Microprocessor"

There were other 32-bit processors that had been around for a while. Very similar to the current 64-bit situation.

The thread specified 2 companies being AMD and Intel and my post started with these 2. Maybe i should of Highlighted the names in bold?

You mean the Pentium Pro, right?

Pretty sure it's the entire Pentium line.

Sun, Sgi, and IBM also had 64-bit processors available quite a while before Intel, who themselves had one quite a while before AMD.

AMD vs Intel Also check my side note.

 

[Captain Newbie]

I do believe we're talking mainstream consumer x86 processors...so let's leave SGI, Sun, IBM and Motorola out of this.

 

[Burning Phoenix]

Intel didn't invent SMP.

Oh yeh i forgot this one.
Intel produced specifications as a guide for machine hardware designers and OS software writers to produce SMP-capable applications to which just about everyone follows these days.

 

[Hydr0 Styles]

I love this competition.

Keeps the companies on their toes, keeps prices down.
Makes consumers happy, because they get to argue.

If AMD started making crappy CPUs, and Intel started making totally pwning CPUs, then I'd switch to intel.

But for the price I was willing to pay for my CPU, the 2500+ was the only route to go.

IMO this discussion is irrelevant.

 

[OC Noob]

I've always found AMD and Intel solutions to be more then adequate.

Why some people feel the need to raise the banner of one company when it does something well and stomp on those of the other is beyond me.

This list is like a little boy asking for a cookie or gold stars for all correct answers on his spelling test.

Why some people feel the need to flaunt their favorite companies list of accomplishments in other peoples faces is beyond me. Especially when they the need to trash another company (whos list of acomplishments, btw, is probably 5 times as long) in the process.

People complain about this or that about Intel or AMD processors, but as far as I know both have always had somewhat comperable solutions and neither has failed to have a processor that performs all that differently in real world apps or atleast very noticably different since 300 mhz CPUs.

It really doesn't matter how they do it so long as the result are there. So if Intel wants to crank up the mhz or delay 64 bit until a 64 bit OS is released from MS why should that be cause for hatred or disdain. Or if AMD decides to milk x86 for a little longer who cares.

Looking at things from a long term perspective, i really don't see one company really kicking the others @SS as far a CPU performance. So why are we seeing all these posts celebrating AMDs new found and alleged dominance over Intel?

I don't recall seeing people celebrate 800 FSB P4s when they were marinally outperforming Athlons, so why are there so many of these kinds of articles and posts now that the A64 is marginally outperforming those 800 fsb P4s.

And I am talking about noticable real world performance, not a simulated 3dmark score.

A64 is the best overall PC processor, but why in the world do people feel the need to proclaim AMD as the golden boy and Intel as the wicked witch or vica versa, but it seems rather silly and pointless to me.


It would be different if they were talking about how AMD has developed some great technology and if it was accurate, but its basically the prop up AMD by taking shots at Intel that has been typical of AMD fanyboys in this forum lately.


What ever smokes peoples shorts, but I'd prefer a more technical, accurate, non-biased informative article or post vs. this cheap shot, gossipish type of thread/post.

Nothing directed at the poster as he didn't write it. This is directed at the author and posts similar to this article.


Oh well, my point will probably be lost in some pointless fanboy argument, but atleast I got that off my chest flamelessly

 

[XWRed1]

I guess i should of specified the CPU was a single chip on silicone. "Microprocessor"

Right. That *was* the first microprocessor, but not the first cpu.

The thread specified 2 companies being AMD and Intel and my post started with these 2. Maybe i should of Highlighted the names in bold?

Thats fine, just don't phrase statements like they were the first ever. Saying something along the lines of Intel making a 32-bit processor before AMD is fine. I know better than to think that Intel made the first 32-bit cpu ever, but you'll give other people the wrong idea.

Pretty sure it's the entire Pentium line.

I'm pretty sure the PPro is what started the whole internal RISC trend for x86 processors. The PPro was the first i686 implementation, it was supposedly a big break from the Pentium.

I do believe we're talking mainstream consumer x86 processors...so let's leave SGI, Sun, IBM and Motorola out of this.

Thats fine. I just think its bad form to do a bunch of arm waving and declaring Intel or AMD to be the first in major fields where they aren't. Again, it gives people the wrong idea.

Oh yeh i forgot this one.
Intel produced specifications as a guide for machine hardware designers and OS software writers to produce SMP-capable applications to which just about everyone follows these days.

You have a link for this? I'm pretty sure there were smp machines around before Intel made them and I don't think Intel convinced the industry to rethink how things like multithreading work.

 

[I.M.O.G.]

Atleast you mentioned where the article came from... But usually its common courtesy to atleast link to the original article. TheInquirer does depend on traffic somewhat to stay afloat. OC.com wouldn't be happy (or atleast wouldn't encourage it outside of these forums) if its articles were fully ripped and posted elsewhere. Just something for everyone to keep in mind, I don't mean to just pick on 7keys here.

Opinion One can't be a follower and expect to lead the industry

The INQUIRER
By Mario Rodrigues: Monday 29 March 2004, 10:34

SNIP

So for both performance and technology leadership, will Intel turn out to be the follower in 2004? If the chip giant can't get a handle on its challenges, and AMD executes smoothly to 90 nm, then yes, from AMD's perspective, Intel will most definitely be behind.

 

[capmuffin]

well with AMD and INTEL for who is in the driver seat,even tho its a different part its ATI and invidia in the front fighting over who gets shotgun, AMD and INTEL in the back yelling in screaming,while microsoft drives along in full control,doesnt microsoft make software than companies build for them,dx9 64bit OS idea...

 

[Cow2kie]

I really don't care who was first, or who used whos idea. If it makes their cpus better... GOOD.. I wish they would share more ideas.

 

[sappo]

It isnt really a question of getting a product out FIRST or "winning" as some would put it. With 64 bit, AMD made a bold move by trying to standardize it. This relatively small corporation was able to force intel to make 64 bit standardized, essentially forcing them to eat their words about when the general public would "need" x86-64. And they were also able to get Microsoft to go along with them.

Even just a few months ago, the world questioned if PC's needed 64 bits. Now AMD has forced the two largest players in the PC industry (Intel and microsoft) to go along with them. Standardization is no longer a question of IF, but WHEN.

Gone are the days where people can legitimately question AMD's ability to influence the industry.

 

[Burning Phoenix]

You have a link for this? I'm pretty sure there were smp machines around before Intel made them and I don't think Intel convinced the industry to rethink how things like multithreading work.

http://www.developerpit.com/intel-smp-24201606.pdf

AS for the Pentium class being a Superscalar RISC I think the correct term would be a Superscalar CISC. Superscalar refers to a "microprocessor" being able to execute more than 1 instruction per clock cycle. This can be done by either introducing more ALU's or pipelines. The Pentium class started with 2 pipelines while the Pentium Pro had 5. CISC is more like an enhanced or compound version of RISC. I have read the Pentium class can still use subset of instructions more resembling a typical RISC instruction set.

 

[XWRed1]

dx9 64bit OS idea...

No, 64-bit AMD and Intel processors were around before Microsoft released operating systems for them.

http://www.developerpit.com/intel-smp-24201606.pdf

All this looks like is Intel documenting mps 1.4, their smp implementation on their own systems. You think non-x86 systems use this?

Heck, in the document their mention existing multiprocessor systems and software.

CISC is more like an enhanced or compound version of RISC.

No. It is a different design philosophy for an instruction set. Its pretty opposite of RISC actually.

CISC is alot of specialized, variable-length instructions while RISC is a smaller set of more generalized, fixed-length instructions.

I have read the Pentium class can still use subset of instructions more resembling a typical RISC instruction set.

Everything since the Pentium Pro is supposed to use RISC-like microops internally, which another part of the processor translates normal x86 CISC code to.

 

[Burning Phoenix]

All this looks like is Intel documenting mps 1.4, their smp implementation on their own systems. You think non-x86 systems use this?

Again stick to the topic being AMD and Intel. Why do you feel like changing the subject. Aren't AMD and Intel x86 based?
Are you trying to say Intel didn't make these specifications which the industry has been following for years? Are you trying to not see that OSes like Windows and Linux support it? Are you trying to say MPS 1.4 and SMP have nothing to do with one another? Why do dual AMD and dual Intel MB's have bios settings for it?

 

[OC Noob]

No, 64-bit AMD and Intel processors were around before Microsoft released operating systems for them.




All this looks like is Intel documenting mps 1.4, their smp implementation on their own systems. You think non-x86 systems use this?

Heck, in the document their mention existing multiprocessor systems and software.




No. It is a different design philosophy for an instruction set. Its pretty opposite of RISC actually.

CISC is alot of specialized, variable-length instructions while RISC is a smaller set of more generalized, fixed-length instructions.




Everything since the Pentium Pro is supposed to use RISC-like microops internally, which another part of the processor translates normal x86 CISC code to.

Are RISC and CISC instruction sets like SSE?

 

[KOXC2003]

Sure, it may be a cheap, and some may consider an unfair way to beat the competitor but, it works for intel. AMD does need a much better marketing scheme to keep up with Intels.

But its hard for the general public to understand how AMD can even be equivallent to an intel, not to mention possible superior. Because the average person believes clock speed is everything. Cache and architecture and even 64 bit vs 32 bit reall mean nothing towards an average consumer.

I remember reading in PCworld a few months back that AMD may have to change its whole idea in making chips, and change to slower architectures but a higher clock speed.

But take a look at benchmarks, even the 64 bit 3400+ only beats Intel's 3.4C in only the majority of games. But in just about every other test, Intel wins.

I think I'll be staying with Intel for awhile.

 

[OC Noob]

Sure, it may be a cheap, and some may consider an unfair way to beat the competitor but, it works for intel. AMD does need a much better marketing scheme to keep up with Intels.

But its hard for the general public to understand how AMD can even be equivallent to an intel, not to mention possible superior. Because the average person believes clock speed is everything. Cache and architecture and even 64 bit vs 32 bit reall mean nothing towards an average consumer.

I remember reading in PCworld a few months back that AMD may have to change its whole idea in making chips, and change to slower architectures but a higher clock speed.

But take a look at benchmarks, even the 64 bit 3400+ only beats Intel's 3.4C in only the majority of games. But in just about every other test, Intel wins.

I think I'll be staying with Intel for awhile.

Xbitlabs.com did a CPU roundup with a ton of tests. While the P4 at 3.2 won a fair share of non-gaming apps the 3200+ still won more of them.

Its a very nice CPU.

 

[Alacritan]

Unfortunately, Intel seems to have hit a wall. They can't ramp up the speed of Prescott at all. They probably struggled to get it to run cool enough to make a production version, and some would argue it still runs too hot to be sold as a mainstream consumer product. Dell knows this, and that's why they don't sell them. And BTX, while potentially an improvment, was only developed because Intel knew they had an extreme problem with heat production. You won't see a faster Prescott for quite a while, I'm willing to wager.

AMD doesn't seem to have this problem. They've always had the superior architecture, (when you can run a chip at 1GHz slower clock speed and get similar results, it's better IMO). Intel is scraping to make their CPUs any faster, where AMD, while not exactly widening the performance gap too quickly (mostly due to financial constraints) is making constant forward progress. It won't be much longer until we see ramped up Hammers and eventually, Newcastles. What will Intel produce to compete with this? Prescott can't get any faster (and consequently hotter). There's been talk of using the Xeon, but what does that gain Intel, using old technology that faces the same problems all P4's have. AMD is on track to switch to 90nm and Intel is dead in the water. As soon as AMD has a solid performance advantage over Intel they would be wise to invest in marketing and really cut prices on CPUs that directly compete with Northwood and Prescott chips.

But who says Intel will even bother to compete with AMD much longer? Intel has actually indicated they're not going to be focusing on speed much longer, but instead, focusing on power consumption and size of their processors so they can dominate the consumer electronics market. Which leaves AMD alone in the consumer CPU market. Intel is wrong to think speed is no longer going to be a driving demand for consumers. Until we have true photorealism and highly advanced AI in games, speed will remain a huge issue. But at least they're smart enough not to beat their heads against the wall and try to make Prescott faster. If they have to move into the consumer electronics market, so be it.

 

[Como]

i agree. with most of you, actually. prescott isnt looking too useful to me, and AMD is producing similar results at a lower cost (and a lower speed...)

now i've slowly come to realise that the general public isnt that bright. sratings are everything, and nothing else matters. even my A+ teacher hates AMD's, because tehy dont run as fast. he never even looks at cache, pipelines, archatecture (i killed that...sp?), or anything else...exept for the # of MHZ. Doesnt read reviews, doesnt care. he will be one of the first to own a BTX system and he will gloat about it over and over, while the 64 bit AMDs are walking all over him, and he wont care because he has more MHZ.

"you know why AMD has those stupid ratings? Bcause they're slower and cant admit it." (quote directly from teacher)

i'd like ot get proof that intel is doing a similar rating to rub in his face...i havnt had time to really get one.

Anyways, why is it that Intel doesnt seem to realise that POWER is what matters to the users? step back and take a look at the last 5 years of computing. a machine from '99 is almost at 100% usage all the time, IF they're still online at all. another 5 years down the road, these fancy new PIV's will be too.

well, supper, g2g!

 

[Captain Newbie]

I find it quite incredible that the prices for the 130-nanometer Athlon XP processors are bottom-basement and that the 130nm Athlon64 processors are still cheaper than their Intel counterparts, and yet deliver similar or better performance. If Intel decides to cut their 130nm prices, I would consider purchasing some of their units for comparisons and to see if hyperthreading actually does anything worthwhile.

A shorter pipeline and a lower clock speed will give you better, more effecient performance. Until recently the PowerPC pipelines had only five stages (approx). Damn shame they weren't x86 compatible.

Between AMD and Intel, however, AMD has the technological upper hand. They have solved the heat problems that plagued their previous flagship processors (Athlon XPs) and are set to move to 64-bit and 90-nanometer. The staggering 32-bit performance of their 64-bit processors is incredible and may well help turn them around. Additionally, their OEM deals with Compaq allow them to sell some truly incredible systems for great prices and will also help AMD assert some OEM might.

AMD me!

 

[man_utd]

You guys are forgetting something though. Efficiency is great and all as long as it is still scalable, as in it can still pound for pound beat other things competing with it. You could feasibly use 5 stage pipeline, but how well would it scale? Would it beat a 21 stage at x faster? If not what is the point?

As for technologically, AMD is kicking the crap out of intel technologically, but AMD is shooting themselves in the foot with the current way they sell hammers.