Was the P4 an 'engineering failure'?

[Conky]

The P4 was and is a good chip. The variations following the Northwood versions were not that great but were still very capable.

To simply call the entire P4 line of cpu's and the millions sold a failure would be ridiculous and is only something someone would suggest if they were some sort of demented fan of another cpu brand.

 

[Acanthus]

Originally posted by: Fox5

Originally posted by: Acanthus
The P4B and P4C were both great.

The P4A, P4D, and pretty much all of the dual cores were low performers.

However, had they been priced more aggressively by Intel, they wouldnt have lost nearly as much marketshare to AMD.

They were pretty aggressively priced, their die sizes were huge compared to AMD's chips.

Yes, and im sure the $5 platter real estate had them reeling in costs over AMD.

 

[tuteja1986]

Prescott was a engineer failure as it failed to beat up Northwood in performance , power consumption and heat. Then every P4 processor after that were failure.

P4 Northwood = going in the right path (i believe northwood came out before AMD 64Bit)
P4 Prescott = Intel totally screwed (AMD 64bit is out and kick its but in almost every way)

 

[Fox5]

Originally posted by: Acanthus

Originally posted by: Fox5

Originally posted by: Acanthus
The P4B and P4C were both great.

The P4A, P4D, and pretty much all of the dual cores were low performers.

However, had they been priced more aggressively by Intel, they wouldnt have lost nearly as much marketshare to AMD.

They were pretty aggressively priced, their die sizes were huge compared to AMD's chips.

Yes, and im sure the $5 platter real estate had them reeling in costs over AMD.

You neglect all the investments they had to put into the research and infrastructure, not to mention failure rates.
Smaller die size = more chips they can produce, it's no coincidence that major price drops tend to happen naturally when a company moves to a small process, and typicall doubles their capacity for that cheap.

 

[Acanthus]

Originally posted by: Fox5

Originally posted by: Acanthus

Originally posted by: Fox5

Originally posted by: Acanthus
The P4B and P4C were both great.

The P4A, P4D, and pretty much all of the dual cores were low performers.

However, had they been priced more aggressively by Intel, they wouldnt have lost nearly as much marketshare to AMD.

They were pretty aggressively priced, their die sizes were huge compared to AMD's chips.

Yes, and im sure the $5 platter real estate had them reeling in costs over AMD.

You neglect all the investments they had to put into the research and infrastructure, not to mention failure rates.
Smaller die size = more chips they can produce, it's no coincidence that major price drops tend to happen naturally when a company moves to a small process, and typicall doubles their capacity for that cheap.

You were talking about die sizes, which is space on the platter. They were on a smaller process, and i dont recall the transistor counts on P4s being double the A64 (which would equate the same die size).

 

[coldpower27]

Originally posted by: Fox5

Originally posted by: Acanthus

Originally posted by: Fox5

Originally posted by: Acanthus
The P4B and P4C were both great.

The P4A, P4D, and pretty much all of the dual cores were low performers.

However, had they been priced more aggressively by Intel, they wouldnt have lost nearly as much marketshare to AMD.

They were pretty aggressively priced, their die sizes were huge compared to AMD's chips.

Yes, and im sure the $5 platter real estate had them reeling in costs over AMD.

You neglect all the investments they had to put into the research and infrastructure, not to mention failure rates.
Smaller die size = more chips they can produce, it's no coincidence that major price drops tend to happen naturally when a company moves to a small process, and typicall doubles their capacity for that cheap.

The only situation where that might have been the case and that I agree with was with the original Willamette P4, once you got to the Northwood stage there were other factors to consider.

Die Size is only one of the cost factors to consider.

 

[coldpower27]

Originally posted by: Acanthus

Originally posted by: Fox5

Originally posted by: Acanthus

Originally posted by: Fox5

Originally posted by: Acanthus
The P4B and P4C were both great.

The P4A, P4D, and pretty much all of the dual cores were low performers.

However, had they been priced more aggressively by Intel, they wouldnt have lost nearly as much marketshare to AMD.

They were pretty aggressively priced, their die sizes were huge compared to AMD's chips.

Yes, and im sure the $5 platter real estate had them reeling in costs over AMD.

You neglect all the investments they had to put into the research and infrastructure, not to mention failure rates.
Smaller die size = more chips they can produce, it's no coincidence that major price drops tend to happen naturally when a company moves to a small process, and typicall doubles their capacity for that cheap.

You were talking about die sizes, which is space on the platter. They were on a smaller process, and i dont recall the transistor counts on P4s being double the A64 (which would equate the same die size).

Actually, AMD's processor got alot more expensive to make once they made the switch to the Athlon 64's, the original Clawhammer had 105.9 Mil transistors, with a die size of 193mm2, this was going against the Northwood's with 55 million Transistors, with a die of 131mm2 and later Prescott which was 125 Million on a die size of 112mm2.

 

[coldpower27]

Originally posted by: myocardia

Originally posted by: Viditor
There are 2 metrics here of note, and it's important to consider both.

Marketshare - the number of total processers sold. ColdPower is quite correct...marketshare has been up and down until the K8 was released. Since then it has been mostly a steady upward movement.

Revenue Share - The share of the total money spent on x86. This has been growing quite steadily and more rapidly for AMD over the last 5 years or so...the reason for the difference is that AMD's marketshare has been moving slowly, but their ASP (Average Sale Price) has also been growing. Remember that prior to Opteron, AMD never had more than .05% marketshare in servers...they are now closer to 30%. This is the biggest profit section of x86.

Ahh, you're totally right, Viditor; my mistake. When I said "marketshare", I was in fact referring to "revenue share". I knew you'd be along sooner or later, to clarify. I just didn't think it would take you so long.

Even if for the moment assuming you meant revenue share, which begs the question of why you responded to a statement regarding marketshare. The 5 fold increase values is still way off the mark.

Indeed I can agree with Viditor that AMD's average sale price has grown since the bargain basement Barton days, where the 90US Athlon XP 2500+ was the hot deal. AMD still doesn't have that large a revenue share however compared to it's marketshare.

Here we have the numbers again.

AMD
Q1 2001, Processor Revenue 661 Million
Q4 2006, Processor Revenue 1.344 Billion

Intel
Q1 2001, Processor + Chipset Revenue 5.133 Billion (yeah, that's how Intel recorded back then)
Q4 2006, Processor Revenue 6.523 Billion

Assuming AMD + Intel = 100%, error introduced by Via should indeed be negligible.
Revenue share of AMD would be 11.40% for Q1 2001, you would get a higher revenue share to start with if you subtracted let's say 20% from the 5.133 Billion for Intel's Chipset Division which was included together in the score at the time.

For Q4 2006, AMD's revenue share is 17.08%. So your looking at growth of revenue share due to processors alone of about 50% and this is optimistic as Intel included chipsets in the Q1 2001 Revenue figure which would mean that the initial revenue percentage due to processors would be higher.

For Viditor, AMD's server marketshare is now as of Q4 2006 at 22.3%, though the 30% figure would be close to correct for desktops no doubt due to the fact that Dell now sells AMD desktops.

 

[alkemyst]

If Intel didn't force stuff on Dell not to sell AMD at AMD's peak performance, the company could have been in better shape.

I like AMD, I like Intel, but not their business ethics. They did not have to resort to monopolistic schemes against such a small competitor.

 

[zsdersw]

Originally posted by: alkemyst
If Intel didn't force stuff on Dell not to sell AMD at AMD's peak performance, the company could have been in better shape.

Unlikely. AMD sold every chip it made, even before Dell started selling computers with AMD chips inside. Their marketshare can't rise beyond their production capacity.

 

[alkemyst]

if you put investors behind something, production would increase.

 

[kobymu]

Originally posted by: Hulk
Always blame the engineers...

Not everyone (in the enthusiast crowed) blamed the engineers...

There was THAT rumors that circled about those days...

A rumor that said that Pentium4 (in its earliest generations) is the result of Intel "strategic alliance" with Rambus, they tried to move the entire platform to a high frequency, high bandwidth platform, everything about the early generations of the Pentium4 design (Willamette + 850E chipset + RDRAM; remember?) revolved around high bandwidth/frequency for its arr... efficiency (my wording is probably off).

There are people who will say that that strategy come about from the management level, and there where rumors of some kind of cross shares agreements between Intel and Rambus.

Intel tried to change the rules of the entire CPU industry by going for an all out MHz race but AMD didn?t play by the "rules" and that their own engineers themselves didn?t manage to achieve the golden threshold (in MHz) at which the P4 design should have shined.

After the initial failure of the RD platform (mainly duo to the insane price of memory), well that a different story, but keep in mind that all the other P4 designs after the Willamette chip, where completely based on the Willamette designs.

 

[mithrandir2001]

My work PC is a 3.2GHz Northwood C. It's a respectable CPU and certainly fast enough for software development tasks today. It does throw out too much heat under full load but such is the nature of NetBurst CPUs.

P4 became an engineering failure when Intel realized they could not get more performance out of the architecture without sending heat dissipation through the roof.

It's pretty amazing how much Core 2 Duo changed things. When running Oggenc (audio encoder) the 3.2Ghz Northwood C runs at about 15x real-time. My 1.8GHz E6300 runs just slightly faster at about 16x real-time. So it eeks out more performance despite its 44% slower clock. And since Oggenc is single-threaded, I can run two copies simultaneously on my C2D for an effective 32x real-time.

 

[jiffylube1024]

I voted yes, but for specific reasons.

Was the P4 an engineering failure? Yes and no. Some really good ideas came out of it, and Intel innovated a lot of good stuff in that time (quad pumped FSB, the trace cache, integrating over meg of full-speed 2+ GHz cache memory was a big step in the right direction, hyperthreading (which became much less important with the introduction of multiple CPU's on a single die), etc). Some stuff, like the quad pumped FSB, the Israeli team responsible for the Pentium M just "borrowed" directly for use in their own architecture.

I would say the P4 was ultimately a failure because:

a.) it veered off the roadmap into oblivion after Prescott came out [ie. it failed to meet expectations once it hit the 90nm die shrink)

and

b.) using Intel's MASSIVE engineering resources, they were ultimately overtaken by their Israeli counterpart (the Pentium M team), who, using 1/10th of the resources and building primarily on the Pentium 3 architecture, created what led to the Core/Core2 architecture that Intel is using today.

Intel's design for the Pentium 4 was based on marketing, not engineering. Their idea was to smash the competition into the ground by winning a clockspeed war, IPC (instructions per clock) be damned. AMD took the reverse route, and, upon releasing the Athlon64, gained a large performance lead on Intel in both the desktop and server market.

Right out of the gate, the P4 was a piece of crap; the original P4 with 256K of cache was released at 1.3 and 1.4 GHz, and performed similarly to the 1 GHz Coppermine P3 CPU. It was subsequently tweaked to go up to 1.9 GHz.

The Northwood P4, hower was a brilliant revision, carrying the P4 from 1.8 GHz to 3.2 GHz and beyond. Northwood also took the Pentium 4 from 180nm (0.18um) to 130nm.

Northwood bumped the FSB to 200MHz (800MHz effective), doubled the cache to 512K, and pulled away from AMD's Athlon through sheer clockspeed (Intel's original intention). Even AMD's model numbers were defeated by the Northwood; a 3 GHz Pentium 4 outperformed an AMD Athlon 3200+ at 2.2 GHz.

But that's where nature reared its ugly head, and power leakage (among other factors) made the 90nm die shrink (called Prescott) run ridiculously hot, and barely outperformed the Northwood core. And they needed to add length of the pipeline on Prescott, I believe going from 20 to 30. The end result was that per MHz, Northwood was marginally faster, even though Prescott had double the L2 cache. The Prescott dual cores were such a knee-jerk reaction to AMD that they barely deserve mention. They simply ran too hot, and consumed an inordinate amount of power. Intel designed the ill-fated BTX platform around the fact that Prescott CPU's were producing so much head, that they needed a better way to exhaust it out of the system. We have AMD's sucessful Athlon64 to thank for not letting BTX gain much foothold, except in the micro-case market.

If Intel didn't have the massive lead in manufacturing capacity, Prescott could have almost put them out of business, but Intel took the step to 65nm so much earlier than AMD (almost 2 years) that they were able to get away with putting so much cache onto their CPU's to stay somewhat competitive, versus the much larger 90nm chips AMD has been producing until just recently.

Prescott was supposed to take the P4 from 3+ GHz to 5+ GHz, and Intel was planning on hitting 8-10 GHz by 2008/2009. Needless to say, that didn't happen... The Pentium 4 was a 10-year architecture design that lasted about 5 years before it became unviable to proceed.
-------

So was the pentium 4 a failure? From an engineering perspective, ultimately yes.

But if you own a Pentium 4 Northwood CPU, it's still a decent CPU today for most applications (hardcore gaming aside). It's better than an AMD Athlon XP, or anything that came before the P4.

But the Core/Core2 (in single and multiple CPU configurations) as well as the Athlon64 are smarter, cooler-running, higher-IPC, better designed CPUs.

 

[jiffylube1024]

Another thing Intel tried to do throughout the P4 architecture was cram proprietary technology down our throats. The original P4 had to be run with expensive RAMBUS RAM, and there were even versions of the Pentium 4 Northwood that used RAMBUS.

Fortunately, due to competition from AMD, they were forced to adopt to the much more popular (and cheaper) DDR memory technology.

Intel supported RAMBUS on their i820 and i850 (top of the line chipsets), and put DDR support in i845, which became their mass-market chipset. They ultimately abandoned RAMBUS, and stuck with DDR support on their i865, i875 chipsets and beyond.

 

[VirtualLarry]

The P4 was specifically engineered for "advertising Mhz", because "Mhz sells". Not for IPC. So in that regard, I would say that it was in fact very sucessfully engineered. Whether or not it is efficient or useful at those clock speeds is another question entirely, really.

 

[evolucion8]

Originally posted by: Cookie Monster
Northwood wasnt cosidered tobe better than prescotts and its other derivatives because it was "faster" in performance. It was "better" because it had a very good performance/power consumption ratio compared to the prescott. They also produced less heat, so it was a better overall complete package. If you look closely at old benches, prescotts exceeded the northwood in overall performance (e.g 3.6ghz perscott would be faster than 3.6ghz northwood although the difference isnt breath taking).

The Prescott outperformed the Northwood by a very small margin only on Hyper Threading tests and Video Encoding/Audio encoding when SSE3 was used, otherwise, the Northwood outperformed the Prescott by a good margin in games, and by a small margin in everything else, I have the last Northwood breed and I have to admit that in theorical benchmarks like Sandra, my CPU even outperforms a Pentium 4 E 570 which runs at 3.8GHz in Multimedia which performs almost identical as a Pentium 4 3.0Ghz Northwood, and in Arithmetic benchmark my Northwood still outperform the P4 E 570 but by a small margin, so Northwood was the best Hyper Threaded Single core Pentium 4 in a power consumption/price ratio. Even today I'm able to play games while encoding movies with little effect in performance, but it runs hot, up to 55C at max load with stock cooling. While my friend's P4 3.2GHz Prescott would idle at that temperature at stock fan lol

 

[evolucion8]

Originally posted by: jiffylube1024
I voted yes, but for specific reasons.

Was the P4 an engineering failure? Yes and no. Some really good ideas came out of it, and Intel innovated a lot of good stuff in that time (quad pumped FSB, the trace cache, integrating over meg of full-speed 2+ GHz cache memory was a big step in the right direction, hyperthreading (which became much less important with the introduction of multiple CPU's on a single die), etc). Some stuff, like the quad pumped FSB, the Israeli team responsible for the Pentium M just "borrowed" directly for use in their own architecture.

I would say the P4 was ultimately a failure because:

a.) it veered off the roadmap into oblivion after Prescott came out [ie. it failed to meet expectations once it hit the 90nm die shrink)

and

b.) using Intel's MASSIVE engineering resources, they were ultimately overtaken by their Israeli counterpart (the Pentium M team), who, using 1/10th of the resources and building primarily on the Pentium 3 architecture, created what led to the Core/Core2 architecture that Intel is using today.

Intel's design for the Pentium 4 was based on marketing, not engineering. Their idea was to smash the competition into the ground by winning a clockspeed war, IPC (instructions per clock) be damned. AMD took the reverse route, and, upon releasing the Athlon64, gained a large performance lead on Intel in both the desktop and server market.

Right out of the gate, the P4 was a piece of crap; the original P4 with 256K of cache was released at 1.3 and 1.4 GHz, and performed similarly to the 1 GHz Coppermine P3 CPU. It was subsequently tweaked to go up to 1.9 GHz.

The Northwood P4, hower was a brilliant revision, carrying the P4 from 1.8 GHz to 3.2 GHz and beyond. Northwood also took the Pentium 4 from 180nm (0.18um) to 130nm.

Northwood bumped the FSB to 200MHz (800MHz effective), doubled the cache to 512K, and pulled away from AMD's Athlon through sheer clockspeed (Intel's original intention). Even AMD's model numbers were defeated by the Northwood; a 3 GHz Pentium 4 outperformed an AMD Athlon 3200+ at 2.2 GHz.

But that's where nature reared its ugly head, and power leakage (among other factors) made the 90nm die shrink (called Prescott) run ridiculously hot, and barely outperformed the Northwood core. And they needed to add length of the pipeline on Prescott, I believe going from 20 to 30. The end result was that per MHz, Northwood was marginally faster, even though Prescott had double the L2 cache. The Prescott dual cores were such a knee-jerk reaction to AMD that they barely deserve mention. They simply ran too hot, and consumed an inordinate amount of power. Intel designed the ill-fated BTX platform around the fact that Prescott CPU's were producing so much head, that they needed a better way to exhaust it out of the system. We have AMD's sucessful Athlon64 to thank for not letting BTX gain much foothold, except in the micro-case market.

If Intel didn't have the massive lead in manufacturing capacity, Prescott could have almost put them out of business, but Intel took the step to 65nm so much earlier than AMD (almost 2 years) that they were able to get away with putting so much cache onto their CPU's to stay somewhat competitive, versus the much larger 90nm chips AMD has been producing until just recently.

Prescott was supposed to take the P4 from 3+ GHz to 5+ GHz, and Intel was planning on hitting 8-10 GHz by 2008/2009. Needless to say, that didn't happen... The Pentium 4 was a 10-year architecture design that lasted about 5 years before it became unviable to proceed.
-------

So was the pentium 4 a failure? From an engineering perspective, ultimately yes.

But if you own a Pentium 4 Northwood CPU, it's still a decent CPU today for most applications (hardcore gaming aside). It's better than an AMD Athlon XP, or anything that came before the P4.

But the Core/Core2 (in single and multiple CPU configurations) as well as the Athlon64 are smarter, cooler-running, higher-IPC, better designed CPUs.

I couldn't say it better. The Athlon 64 3200+ and up was the only CPU that was able to trade blows with the Northwood core at 3.2GHz and more and outperforms it in general applications and games, the only thing that the Athlon 64 lagged a bit was in encoding audio/video and multi threaded application due to the high speed clocks and HT of the P4 Northwood, but considering the clockspeed difference, it was by 20% average, sometimes even smaller the difference.

 

[Cookie Monster]

I see you own a P4 3.4C northwood, i got a 3.0C, northwood (yes its OCable to 3.6 and above ). Been using it for 4 years and its still going strong. Northwoods got to be the best of the P4s, because it terms of price/power consumption/performance its so much better in that regard than its prescott and its other derivatives.

What jiffylube said is correct. They tried everything they can to at least reach not 10ghz as they originally planned but 4ghz, using all sorts of methods like die shrinks to 65nm, using various methods to get the heat/power consumption in check, add more L2 cache etc but at the end of the day they failed to pass over the 3.8ghz limit.

What chip i think was good after prescott was the cedar mill chips. But too little and too late by intel, and by then core 2 duo was around the corner.

 

[Shlong]

I was happy with the Northwood C. I bought a Barton 2500 (I think it was $130 at the time) and overclocked that to 2.4 ghz which was faster than the XP3200 and much cheaper. I also bought at the same time a Pentium4 2.4C (I think it was $165 at the time) and overclocked that to 3.4 ghz which was faster than the 3.2 and much cheaper. Both were good machines, the Barton was faster in the games but I preferred the 2.4c because it felt faster using it on the desktop (maybe because of hyper-threading) and it was also faster at multimedia encoding & photoshop. I still use the 2.4c as a secondary machine today, AMD had better performance for the buck but I don't see it as a failure.

 

[tno]

Was the Pentium 4 a failure? To answer that question, all you need to do is look at the Core 2 Duos we love so much. With the Pentium III's Intel knew it had a good thing going. But they were stuck at a certain clock speed and with their smaller process (.13 micron) still a ways away, they needed to do something different to break the clock wall. Have no doubt, the Intel engineers knew that a few years down the road short-pipelined, multicore, nm process CPUs would rule the roost. But that would take lots of development and investment in the smaller processes. So, they created the two team approach. One team was tasked with a singular goal, produce faster clock speeds. The other team, was to perfect the Pentium III's approach. The first team, produced what we know as the Pentium 4. Read Anand's 2001 coverage of the P4 premiere and its earliest 1.7Ghz revision. They were horribly unimpressed with the processors per Ghz performance, as the early P4's were outpaced by similarly clocked Athlons AND PIII's. But the long pipeline and turbocharged APU made the chip easy to clock much higher than the PIII or its near succesor the Pentium M ever would. That is until the Core 2 series.

Don't consider the Pentium 4 a failure because it didn't perform as efficiently as its competition, nor do as well at gaming. It was not made to do either of these things. It was made so that we wouldn't have to suffer through anemic clockspeeds and performance for two years, instead we'd get to see 3.0Ghz way faster than we expected AND they made enough money off the damned things to keep the low yield Israeli chip lab working on the incredible chips we have now.

I'm not an "Elite" indeed this is my first post. But in hindsight, those six year old P4 reviews I've been reading clue us in to the brilliance of those engineers.

Jason

 

[magreen]

Good first post!

But I'm not so sure what you said fits with the interviews of Mooly Eden I read. He made it sound like Intel made a management mistake to go with Netburst, and only years later they realized they'd goofed and switched to the Pentium M concept for the desktop/server. That they didn't have it in mind all along. Of course, being a rep. of the Israeli team he might have a natural inclination to say they should have gone with their concept sooner

 

[apoppin]

yes P4 was a *compelete utter failure*

it *failed* to reach intel's target of 10Ghz ... by over 6Ghz
that's a *missfire* in any language

how many other ways can you spell f-a-i-l-u-r-e

and intel's sudden *dumping* of NetBurst to head in another direction speaks volumes about their original choice of the P4

IF they had *stuck* with the P3 architecture, they'd be WAY ahead of AMD
:Q

even Further then they are now
:brokenheart:

P4 was an "engineering triumph" ... for AMD

imo

 

[Neurodog]

Originally posted by: apoppin
yes P4 was a *compelete utter failure*

it *failed* to reach intel's target of 10Ghz ... by over 6Ghz
that's a *missfire* in any language

how many other ways can you spell f-a-i-l-u-r-e

and intel's sudden *dumping* of NetBurst to head in another direction speaks volumes about their original choice of the P4

IF they had *stuck* with the P3 architecture, they'd be WAY ahead of AMD
:Q

even Further then they are now
:brokenheart:

P4 was an "engineering triumph" ... for AMD

imo

Well said!

 

[apoppin]

Originally posted by: Neurodog

Originally posted by: apoppin
yes P4 was a *compelete utter failure*

it *failed* to reach intel's target of 10Ghz ... by over 6Ghz
that's a *missfire* in any language

how many other ways can you spell f-a-i-l-u-r-e

and intel's sudden *dumping* of NetBurst to head in another direction speaks volumes about their original choice of the P4

IF they had *stuck* with the P3 architecture, they'd be WAY ahead of AMD
:Q

even Further then they are now
:brokenheart:

P4 was an "engineering triumph" ... for AMD

imo

Well said!

thank-you

and that was just Part One

shall i go on?


even when the Williamette P4 was *introduced* it was horrible performance ... i didn't realize the River it was named for was so long, shallow, muddy and clogged with debris

i had a 1.2Ghz P3 Tualiatin Celeron at 1.6 Ghz and it just blew-away the early P4s ...well over 2Ghz .. it wasn't until the 533FSB 2.4Ghz came out that it was even 'competitive' and only massive o/cs make it a great buy ... it DID however spank the XP series.

the first decent P4 was the 800FSB 2.80c - it soundly defeated XP and it put up a pretty good 'fight' against the early a64 when the P4 was ONLY well-o/c'd ..... that's the one i got and running it at 3.3-3.31Ghz allowed it to compete with a stock a64 3000+ in games.

from then on it was DOWNhill for intel ... they realized that they went a *wrong direction* with Prescott ... the pipelines kept getting longer and *debugging* was getting to be a matter of "luck" ... and you could cook your case with the early ones ... in desperation to keep up with AMD, they went to the 1000[+] dollar Extreme edition P4 ... and my 3.4EE gets it's ass handed to it at stock speeds by a cheaper FX53.

so Intel wisely *dumped* it ... and returning to P3 via the M architecture has put them on top again [while AMD snoozed]

the P4 was very good for AMD

anyone think P4 was a *success*?

... of course, you have to remember that hindsight is always 20/20 ... IF intel kept going with the budget they wasted on P4 and instead placed it on refining the PIII-M, they'd be way ahead ... and AMD might be a minor player now
:Q

i always though P4 was a "marketing-driven" decision ... to dupe the public about clockspeed ... and one i think intel regrets and one they are trying to play that down now

now it's "all about" the number of cores