Was the P4 an 'engineering failure'?

[Arkaign]

Originally posted by: apoppin
i do agree that P4 taught important lessons; most failures do.

The P4 maintained a massive marketshare lead over AMD's products, even during periods when AMD's products were IMHO superior (before and after the Northwood era). Profits were through the roof. AMD never fielded a Socket A chip that really 'won' after Willamette faded. If that's 'failure', then I'm sure AMD would be ludicrously happy to have some of that 'failure'.

 

[Atechie]

So because Intel was the first to venture into uncharted physics territory the P4 failed?
When i got my P4A 2.4Ghz AMD had nothing to compete.

Hell DDR ram was a PC2100...my RDRAM was PC3200...took about one year before DDR RAM catched up.


But if you call the P4 a failure, then the Phenom 2 must be a DISASTER.
Same diesize as i7, but not in the same leauge...by far.
Needing a 1.2Ghz more just to macht (Phenom II X4 965 BE) the Core i7 920 (3.4Ghz vs 2.66Ghz) is what I call a failure of proportions.

The gap was NEVER that big in performance back in the P4 days...something to think about.

 

[snouter]

I rather think it was a marketing failure that led to engineering misdirection.

At the time a lot of things were not clear. Multi-cpu, multi-core, multi-thread, pure speed. At the time it seemed like the goal was a 10GHz computer. Over time that sort of shifted to performance per watt, performance per cycle, and multi-core, multi-thread.

Also, mobile became a factor, and the P4 was a pretty dreadful mobile chip.

Enter Banias and Pentium M and Intel saw the future and abandoned the pure clock speed race.

Hints were available early on. I remember when the first Pentium 4 chips came out in the 1.6-1.7 range, I remember preferring the Pentium III chips over the Pentium 4 chips. Something seemed wrong. The Pentium 4 chips seemed like they were chugging in a rather wasteful manner. The clock was faster, but they sure did not feel faster.

Anyhow, I don't totally blame Intel, if anything, give them credit for one heck of a lesson learned. They are strong across the board now.

Nobody knew what to do once past the 1GHz level and that had to be worked out. I was not a fan of the P4 and it seems like they were around for a little too long, but, it's hard to say that they did not learn a lot of deep lessons from the Pentium 4 and it shows in the new philosophy (perf per watt, cycle and even some dabbling in low power) and lineup.

 

[dmens]

Originally posted by: apoppin
There is nothing of mine that i have posted here that you have been able to comprehend, nevermind correct. There is nothing of value in any of your posts in this thread at all. Just your own personal opinion and personal attack after attack.

i do agree that P4 taught important lessons; most failures do.

you wouldn't know what kind of lesson P4 taught if it slapped you in the face. besides, you haven't stated anything technical for me to correct because you don't know where to start. if you actually gave specific examples other than the vague "10ghz" soundbite we could have an actual technical discussion, but you can't even tell replay from rename, so why bother.

on the other hand, i did correct your errors in non-technical posts, for example, the P4 team worked on nehalem and westmere after prescott, not "larabeast". architectural features of the P4 still live on, SMT being the biggest one, but others as well which have not yet been disclosed.

 

[snouter]

I did have a hyper-threading 3.0 laptop and always thought it was snappier than the non hyper threading one I had. So. yeah, HT was a good idea and I love it on my 920.

 

[AlgaeEater]

Originally posted by: Jadow
Think of all the millions of P4's that are sitting in offices around the world. It definitely wasn't a financial failure for Intel.

This. THIS COMMENT SO MUCH. Thank you

I absolutely hated the P4 back way then when I had to build systems for my work and friends. The sheer amount of revisions always gave me mixed results across "flat" bulk builds. Going out to buy "20 P4 chips" was like buying a variety pack of potato chips at the supermarket. Always different some-way, some how.

And even now I know and still can reference those machines sitting around certain offices. They haven't moved, and they suck back then as they suck now.

Sorry if this post seems kinda rambling, I just really hated the P4. Engineering failure or not, it's a failure clean cut in my book.

 

[Keysplayr]

Hmm.. Well, I've had a 2.0a Northy (first P4 for me), 2.4 Northy 2.8 Northy, 3.0 Prescott and a 520J which I think was either 3.73 or 3.8 GHz.

Did these CPU's live up to my expectations? Well, lets see.

I expected to use the PC for work. I worked. All apps ran fast and fine.

I expected to use the PC for gaming. I gamed. GPU mattered more then the CPU it would seem.

I just upgraded my wifes rig from a 2.4 GHz P4 to a Athlon x2 5600 setup last month.

So if anything, the P4 chips were good enough for anybody to do what they needed to do.
A64's were faster at most tasks, but the everyday Joe wouldn't notice. Very reliable and stable.

Oh, forgot about my server. I bought a Celeron 2.6 years ago. Still cranking away to this day.

Ah, and a 1.6a I inherited from upgrading a relatives PC. It's now in my daughters PC for playing her learning games. Six or seven year old CPU. Won't die. hehe.

Failures? No way. Extreme success.

 

[Cookie Monster]

Originally posted by: dmens
snip

I agree with apoppin. There is nothing of value from your posts since it doesn't contribute to the thread at all. Instead all Im reading is personal attacks/insults such as

Originally posted by: dmens
you wouldn't know what kind of lesson P4 taught if it slapped you in the face

if i were a "junior engineer" that still puts my knowledge way above yours don't it.

and yeah you are annoying as fuck and i make it a point to correct your error-laden posts

haha you are too stupid to understand.

List goes on and on. Its quite humorous actually because all the examples above have been from this necro'd thread alone.

Im sure you are quite knowledgeable in the field of micro processors as you seem to boast about it in every single post of yours. So instead of flexing your ego around thinking that all the lesser being around you knows nothing about micro processors except you. Why not explain your points in a constructive manner (like most engineers that I know do)? with maybe evidences to support your claims/points in as to why you disagree or agree with some of the points made by the lesser folks.

Anyway, I dont think the term "Engineering Failure" is appropriate in this situation (unless you were an intel engineer that was part of the team that was designing this chip and knew its exact target specs/goals/objectives) because Pentium4s were still competitive in many ways to the competing product (even if the competing product was faster overall). It wasn't a case where the chip faced disastrous manufacturing problems, nor required a nuclear reactor to run or was so slow performance wise that the only option was to buy an AMD CPU. Id say Intel got "out engineered" sorely based on the performance data that the rest of us has access to in forms of benchmarks. Two excellent engineering solutions that tackled the same problem, except one did better overall.

 

[apoppin]

Originally posted by: dmens

Originally posted by: apoppin
There is nothing of mine that i have posted here that you have been able to comprehend, nevermind correct. There is nothing of value in any of your posts in this thread at all. Just your own personal opinion and personal attack after attack.

i do agree that P4 taught important lessons; most failures do.

you wouldn't know what kind of lesson P4 taught if it slapped you in the face. besides, you haven't stated anything technical for me to correct because you don't know where to start. if you actually gave specific examples other than the vague "10ghz" soundbite we could have an actual technical discussion, but you can't even tell replay from rename, so why bother.

on the other hand, i did correct your errors in non-technical posts, for example, the P4 team worked on nehalem and westmere after prescott, not "larabeast". architectural features of the P4 still live on, SMT being the biggest one, but others as well which have not yet been disclosed.

again, you have zero clue about what i know
:thumbsdown:

Intel *canceled* the P4 abruptly; they stopped all their research and development on it and went back to the drawing board with architecture that became what they have now
- WHY?> because they SAID they could not meet their goals with it

if that isn't FAILURE, you tell me what is
- just because Intel kept some useful "scraps" related to netburst and hyperthreading, does not mean P4 was an "engineering success"
.. what they ended up with was ridiculed as PressHot for years to come
:roll:

Originally posted by: Keysplayr
Hmm.. Well, I've had a 2.0a Northy (first P4 for me), 2.4 Northy 2.8 Northy, 3.0 Prescott and a 520J which I think was either 3.73 or 3.8 GHz.

Did these CPU's live up to my expectations? Well, lets see.

I expected to use the PC for work. I worked. All apps ran fast and fine.

I expected to use the PC for gaming. I gamed. GPU mattered more then the CPU it would seem.

I just upgraded my wifes rig from a 2.4 GHz P4 to a Athlon x2 5600 setup last month.

So if anything, the P4 chips were good enough for anybody to do what they needed to do.
A64's were faster at most tasks, but the everyday Joe wouldn't notice. Very reliable and stable.

Oh, forgot about my server. I bought a Celeron 2.6 years ago. Still cranking away to this day.

Ah, and a 1.6a I inherited from upgrading a relatives PC. It's now in my daughters PC for playing her learning games. Six or seven year old CPU. Won't die. hehe.

Failures? No way. Extreme success.

Engineering failure to meet goals - yes!

i had P4s from the time i traded up my Tualatin Celeron to a P4 Northwood 2.8C that i overclocked to 3.3GHz. Heck, i even ended up with the Extreme Edition that i got for $100 that i got to 3.73 GHz .. and i stayed with Intel right through .. well, until this month when i built my own FIRST AMD PC .. and i expect to get core i5 [or i7] next month
- i think i know a little bit about NetBurst and i believe Intel acknowledged its OWN failure with NetBurst when they DROPPED it like a hot Prescott

sure they "worked" fine .. but when a company CANCELS all research and development - SUDDENLY - as they did with P4; that is not a "engineering success"

 

[zsdersw]

Unrealistic goals (10GHz) do not make the P4 an "engineering failure". A marketing and leadership failure, perhaps, but not an engineering failure.

Show me a 10GHz x86 chip guaranteed to function properly at that speed by its manufacturer. Since there isn't one, by your logic every x86 chip being made today is an engineering failure. Don't kid yourself, every chip firm making x86 chips would love to have a 10GHz x86 chip that works at that speed right out of the box and with the longevity of any other chip, they're just not going to ever state it as a goal until it's feasible from an engineering perspective. That's the only failure with the P4; unrealistic expectations from sales/marketing.

 

[exar333]

Originally posted by: apoppin

Originally posted by: dmens

Originally posted by: apoppin
There is nothing of mine that i have posted here that you have been able to comprehend, nevermind correct. There is nothing of value in any of your posts in this thread at all. Just your own personal opinion and personal attack after attack.

i do agree that P4 taught important lessons; most failures do.

you wouldn't know what kind of lesson P4 taught if it slapped you in the face. besides, you haven't stated anything technical for me to correct because you don't know where to start. if you actually gave specific examples other than the vague "10ghz" soundbite we could have an actual technical discussion, but you can't even tell replay from rename, so why bother.

on the other hand, i did correct your errors in non-technical posts, for example, the P4 team worked on nehalem and westmere after prescott, not "larabeast". architectural features of the P4 still live on, SMT being the biggest one, but others as well which have not yet been disclosed.

again, you have zero clue about what i know
:thumbsdown:

Intel *canceled* the P4 abruptly; they stopped all their research and development on it and went back to the drawing board with architecture that became what they have now
- WHY?> because they SAID they could not meet their goals with it

if that isn't FAILURE, you tell me what is
- just because Intel kept some useful "scraps" related to netburst and hyperthreading, does not mean P4 was an "engineering success"
.. what they ended up with was ridiculed as PressHot for years to come
:roll:

Originally posted by: Keysplayr
Hmm.. Well, I've had a 2.0a Northy (first P4 for me), 2.4 Northy 2.8 Northy, 3.0 Prescott and a 520J which I think was either 3.73 or 3.8 GHz.

Did these CPU's live up to my expectations? Well, lets see.

I expected to use the PC for work. I worked. All apps ran fast and fine.

I expected to use the PC for gaming. I gamed. GPU mattered more then the CPU it would seem.

I just upgraded my wifes rig from a 2.4 GHz P4 to a Athlon x2 5600 setup last month.

So if anything, the P4 chips were good enough for anybody to do what they needed to do.
A64's were faster at most tasks, but the everyday Joe wouldn't notice. Very reliable and stable.

Oh, forgot about my server. I bought a Celeron 2.6 years ago. Still cranking away to this day.

Ah, and a 1.6a I inherited from upgrading a relatives PC. It's now in my daughters PC for playing her learning games. Six or seven year old CPU. Won't die. hehe.

Failures? No way. Extreme success.

Engineering failure to meet goals - yes!

i had P4s from the time i traded up my Tualatin Celeron to a P4 Northwood 2.8C that i overclocked to 3.3GHz. Heck, i even ended up with the Extreme Edition that i got for $100 that i got to 3.73 GHz .. and i stayed with Intel right through .. well, until this month when i built my own FIRST AMD PC .. and i expect to get core i5 [or i7] next month
- i think i know a little bit about NetBurst and i believe Intel acknowledged its OWN failure with NetBurst when they DROPPED it like a hot Prescott

sure they "worked" fine .. but when a company CANCELS all research and development - SUDDENLY - as they did with P4; that is not a "engineering success"

Maybe this isn't so black and white. Although it might not have been a complete success, it succeded in some areas, and obviously fell short in longevity and reaching 10ghz.

The failure could be more of unrealistic expectations on the pat of Intel, and a moderately successful platform that peaked with the Northwood variants, and really showed it's issues with power usage with Prescott.

IMHO, I would say OVERALL, that the P4 was neither a failure or a success in engineering. Intel learned some hard lessons (power usage) and had some successes (HT, etc). It met some expectations and failed others, I don't think this categorizes it as fully an engineering failure.

Apoppin - One point I do disagree with you is that just because Intel dropped the P4 architecture, that doesn't inherently mean it's a failure. Companies (and people) drop something useful and successful for the next "big thing" all the time, it doesn't mean the previous was a failure, but that the new thing is BETTER.

Short Version:

P4A - Crap
P4B - Meh
P4C - Great
Prescott - meh

p4 overall - meh

Edit: SP

 

[apoppin]

Originally posted by: ExarKun333

Originally posted by: apoppin

Originally posted by: dmens

Originally posted by: apoppin
There is nothing of mine that i have posted here that you have been able to comprehend, nevermind correct. There is nothing of value in any of your posts in this thread at all. Just your own personal opinion and personal attack after attack.

i do agree that P4 taught important lessons; most failures do.

you wouldn't know what kind of lesson P4 taught if it slapped you in the face. besides, you haven't stated anything technical for me to correct because you don't know where to start. if you actually gave specific examples other than the vague "10ghz" soundbite we could have an actual technical discussion, but you can't even tell replay from rename, so why bother.

on the other hand, i did correct your errors in non-technical posts, for example, the P4 team worked on nehalem and westmere after prescott, not "larabeast". architectural features of the P4 still live on, SMT being the biggest one, but others as well which have not yet been disclosed.

again, you have zero clue about what i know
:thumbsdown:

Intel *canceled* the P4 abruptly; they stopped all their research and development on it and went back to the drawing board with architecture that became what they have now
- WHY?> because they SAID they could not meet their goals with it

if that isn't FAILURE, you tell me what is
- just because Intel kept some useful "scraps" related to netburst and hyperthreading, does not mean P4 was an "engineering success"
.. what they ended up with was ridiculed as PressHot for years to come
:roll:

Originally posted by: Keysplayr
Hmm.. Well, I've had a 2.0a Northy (first P4 for me), 2.4 Northy 2.8 Northy, 3.0 Prescott and a 520J which I think was either 3.73 or 3.8 GHz.

Did these CPU's live up to my expectations? Well, lets see.

I expected to use the PC for work. I worked. All apps ran fast and fine.

I expected to use the PC for gaming. I gamed. GPU mattered more then the CPU it would seem.

I just upgraded my wifes rig from a 2.4 GHz P4 to a Athlon x2 5600 setup last month.

So if anything, the P4 chips were good enough for anybody to do what they needed to do.
A64's were faster at most tasks, but the everyday Joe wouldn't notice. Very reliable and stable.

Oh, forgot about my server. I bought a Celeron 2.6 years ago. Still cranking away to this day.

Ah, and a 1.6a I inherited from upgrading a relatives PC. It's now in my daughters PC for playing her learning games. Six or seven year old CPU. Won't die. hehe.

Failures? No way. Extreme success.

Engineering failure to meet goals - yes!

i had P4s from the time i traded up my Tualatin Celeron to a P4 Northwood 2.8C that i overclocked to 3.3GHz. Heck, i even ended up with the Extreme Edition that i got for $100 that i got to 3.73 GHz .. and i stayed with Intel right through .. well, until this month when i built my own FIRST AMD PC .. and i expect to get core i5 [or i7] next month
- i think i know a little bit about NetBurst and i believe Intel acknowledged its OWN failure with NetBurst when they DROPPED it like a hot Prescott

sure they "worked" fine .. but when a company CANCELS all research and development - SUDDENLY - as they did with P4; that is not a "engineering success"

[snip]

Apoppin - One point I do disagree with you is that just because Intel dropped the P4 architecture, that doesn't inherently mean it's a failure. Companies (and people) drop something useful and successful for the next "big thing" all the time, it doesn't mean the previous was a failure, but that the new thing is BETTER.

[snip]

i would agree with you except for the FACT that intel had NOTHING BETTER at the time - just Bananas and "M"

Intel *stopped* research and development COLD
- and started in a new area they did NOT INTEND to continue; they went to their mobile line and resurrected it for Desktop

it was an engineering failure to intel - at least they treated it as such

.. i will also agree that some good came of it; but when they only take "parts" of it and leave the majority behind, that IS *failure*
.. now IF Intel had NEVER started with P4, but instead continued on with Tualatin and "M" - originally - i would say that they would be much further along at this point in time - P4 was a detour and a dead end for intel

 

[Arkaign]

Apoppin, by your overly simplistic logic, any technical architecture that is ever replaced by another, is a failure. That's grossly ignorant. The P4 traded blows for years with the best of AMD until it came time for a replacement. The P4 served its purpose perfectly.

Designing CPUs is a difficult task, it often takes years to go from a general concept to a retail product, and for the most part, the P4's that hit retail were either very competitive or class leading. When it became clear that the competition was at last able to field a truly superior product with AMD64, they shifted focus to another project that seemed promising. Notice they never stopped the P3 architecture development that led into Conroe/etc, Intel is large enough to be able to manage a multitude of development projects at once.

P4 served its purpose, and was not a 'failure'. Get over it.

 

[aigomorla]

OK Dmens, and POPin, take a time out please....

I dont mind u taking swings at the topic, but please not each other...

Anyhow... the P4 u can agrue all you want but it was a design failure because the project was scrapped.

The entire netburst timetable, and entire netburst advances were ALL CANCELED.

As i said C2D is nothing like a P4. Its an adaptation and maturity of a P3.

So C2D should of came after P3 instead of the whole fiasco we went though with P4.

So now compare C2D vs a P4.... umm yes P4 was a complete and utter failure.

The only thing we learned from a P4 was HyperThreading.

And we use it on an i7. So if u want to consider hyperthreading a good thing, because it itself has mix opinions, then i guess u can say p4 was okey.

 

[zsdersw]

Intel didn't stop anything "cold". They put it on the shelf and took quite a bit from the P4 and incorporated it with the Banias core: quad-pumped FSB (now replaced by QPI), SMT, and a huge one--branch prediction. The nature of the P4 demanded improvements in branch prediction to avoid huge performance penalties in the event of branch prediction errors. As a result, Conroe and its successors/relatives have incredibly good branch predictors.

If/when there comes a time when a long/narrow CPU design is feasible or desired Intel can take the P4 off the shelf and start on it with many design elements already done.

 

[Nemesis 1]

I begg to differ. P4 didn't not turn out to be what intel invisioned that true . But a failure hardly . What intel learned from P4 carries forward with next generation . Netburst moved uo to nehalem family with hyper threading . I consider HT part of the netburst family . I look at it on nehalem and I don't see failure. Was P4 all intel invisioned . no it wasn't . But it made billions for intel . SO failure hardly unless ya use a red herring argument.

 

[Arkaign]

Originally posted by: aigomorla
OK Dmens, and POPin, take a time out please....

I dont mind u taking swings at the topic, but please not each other...

Anyhow... the P4 u can agrue all you want but it was a design failure because the project was scrapped.

The entire netburst timetable, and entire netburst advances were ALL CANCELED.

As i said C2D is nothing like a P4. Its an adaptation and maturity of a P3.

So C2D should of came after P3 instead of the whole fiasco we went though with P4.

So now compare C2D vs a P4.... umm yes P4 was a complete and utter failure.

The only thing we learned from a P4 was HyperThreading.

And we use it on an i7. So if u want to consider hyperthreading a good thing, because it itself has mix opinions, then i guess u can say p4 was okey.

?? You too?

The engineering aspects of creating stable processors in the 3ghz+ range were learned during the P4 era.

The P4 architecture lasted for YEARS against the competition, and during the Northwood era of roughly ~18 months or so, was dominantly faster.

The P4 served its purpose adequately until it was time to replace it. It's certainly true that they rode the P4 train for about a year too long, but they had to ramp up production of a viable replacement project ahead of schedule. Remember, Athlon XP had run out of steam completely, and the 3200+ was hardly a match for the 3.2Ghz Northwood. AMD had a lot of trouble getting their AMD64 to market, early samples couldn't even break 1ghz. When the consumer-level Althon64 finally hit, it still wasn't compellingly faster than the Pentium 4, as they only premiered the FX and 3200+. Certainly better chips IMHO, but not dominant by any means. It took the 3400+, 3500+, etc, to finally put the nail in the coffin for P4.

Would you say Athlon XP / Socket 462 were a failure because they eventually needed replacement against P4? How about the short-lived 'Slot-A'?

If the product competes well, sells well, and serves the purpose for the company, it's incredibly ignorant to label them a failure. Sure competition may require replacement, but that doesn't rate something a failure. Socket 775 and the C2D are going away slowly but surely, I guess they are total failures too, as Core i7 is a significant departure?

 

[Fox5]

Originally posted by: Atechie
So because Intel was the first to venture into uncharted physics territory the P4 failed?
When i got my P4A 2.4Ghz AMD had nothing to compete.

Hell DDR ram was a PC2100...my RDRAM was PC3200...took about one year before DDR RAM catched up.


But if you call the P4 a failure, then the Phenom 2 must be a DISASTER.
Same diesize as i7, but not in the same leauge...by far.
Needing a 1.2Ghz more just to macht (Phenom II X4 965 BE) the Core i7 920 (3.4Ghz vs 2.66Ghz) is what I call a failure of proportions.

The gap was NEVER that big in performance back in the P4 days...something to think about.

Math check, that's a bit less than 800Mhz more, not 1.2Ghz. Unless you were saying the 965BE still doesn't match the i7 920 at that speed.

But I'd say the gap in performance is about the same as it was back in the P4 days. AMD eventually had 3 or 3.2Ghz Athlon 64s, Intel maxed out at 3.8Ghz Prescotts, and even those were beat by near 2Ghz Athlon 64s. Depending on the application, the 3.8Ghz Prescott was beat by the 2.2Ghz Athlon 64.

Oh, and don't underestimate what the P4 may have done for Intel's knowledge of process engineering. Having to deal with tough problems early could have been a large help for them.

 

[Denithor]

Originally posted by: ExarKun333
P4A - Crap
P4B - Meh
P4C - Great
Prescott - Meh

Pretty much truth right there. I went from a P3-733 to a P4-2.4B (533FSB, no HT) and while it was noticeably faster it just didn't impress. Then I popped in a 2.4C and WOW everything went faster, especially when I was running multiple things simultaneously. Next up was an A64 3000+ and I honestly felt like it was a downgrade - simply wasn't as responsive when multitasking. Then I bumped up to an X2 3800+ and the WOW factor was back - everything was speedy. Next upgrade was to an e6400 and it was like WOW WOW! Now I'm on an e8400 and it's not noticeably better than the e6400 (except for lower thermals which is nice).

Next up will probably be an i5 quad (not sure if I will do a 4/4 or a 2/4HT).

 

[Arkaign]

Originally posted by: Fox5

Originally posted by: Atechie
So because Intel was the first to venture into uncharted physics territory the P4 failed?
When i got my P4A 2.4Ghz AMD had nothing to compete.

Hell DDR ram was a PC2100...my RDRAM was PC3200...took about one year before DDR RAM catched up.


But if you call the P4 a failure, then the Phenom 2 must be a DISASTER.
Same diesize as i7, but not in the same leauge...by far.
Needing a 1.2Ghz more just to macht (Phenom II X4 965 BE) the Core i7 920 (3.4Ghz vs 2.66Ghz) is what I call a failure of proportions.

The gap was NEVER that big in performance back in the P4 days...something to think about.

Math check, that's a bit less than 800Mhz more, not 1.2Ghz. Unless you were saying the 965BE still doesn't match the i7 920 at that speed.

But I'd say the gap in performance is about the same as it was back in the P4 days. AMD eventually had 3 or 3.2Ghz Athlon 64s, Intel maxed out at 3.8Ghz Prescotts, and even those were beat by near 2Ghz Athlon 64s. Depending on the application, the 3.8Ghz Prescott was beat by the 2.2Ghz Athlon 64.

Oh, and don't underestimate what the P4 may have done for Intel's knowledge of process engineering. Having to deal with tough problems early could have been a large help for them.

I think that's exactly what he's saying, as it takes the Phenom II X4 at around 3.8-4.0ghz to start matching the stock i7 920 performance in some apps, while in others it still lags even then. My brand-new X4 @ 3.4ghz gets beaten in many things by stock Q6600s that are a couple of years old.

 

[myocardia]

Originally posted by: zsdersw
Unrealistic goals (10GHz) do not make the P4 an "engineering failure".

You obviously are neither an engineer, nor even an engineering student. Unrealistic (i.e., unattainable) goals is the very definition of an engineering failure.

 

[apoppin]

Originally posted by: Arkaign

Originally posted by: aigomorla
OK Dmens, and POPin, take a time out please....

I dont mind u taking swings at the topic, but please not each other...

Anyhow... the P4 u can agrue all you want but it was a design failure because the project was scrapped.

The entire netburst timetable, and entire netburst advances were ALL CANCELED.

As i said C2D is nothing like a P4. Its an adaptation and maturity of a P3.

So C2D should of came after P3 instead of the whole fiasco we went though with P4.

So now compare C2D vs a P4.... umm yes P4 was a complete and utter failure.

The only thing we learned from a P4 was HyperThreading.

And we use it on an i7. So if u want to consider hyperthreading a good thing, because it itself has mix opinions, then i guess u can say p4 was okey.

?? You too?

The engineering aspects of creating stable processors in the 3ghz+ range were learned during the P4 era.

The P4 architecture lasted for YEARS against the competition, and during the Northwood era of roughly ~18 months or so, was dominantly faster.

The P4 served its purpose adequately until it was time to replace it. It's certainly true that they rode the P4 train for about a year too long, but they had to ramp up production of a viable replacement project ahead of schedule. Remember, Athlon XP had run out of steam completely, and the 3200+ was hardly a match for the 3.2Ghz Northwood. AMD had a lot of trouble getting their AMD64 to market, early samples couldn't even break 1ghz. When the consumer-level Althon64 finally hit, it still wasn't compellingly faster than the Pentium 4, as they only premiered the FX and 3200+. Certainly better chips IMHO, but not dominant by any means. It took the 3400+, 3500+, etc, to finally put the nail in the coffin for P4.

Would you say Athlon XP / Socket 462 were a failure because they eventually needed replacement against P4? How about the short-lived 'Slot-A'?

If the product competes well, sells well, and serves the purpose for the company, it's incredibly ignorant to label them a failure. Sure competition may require replacement, but that doesn't rate something a failure. Socket 775 and the C2D are going away slowly but surely, I guess they are total failures too, as Core i7 is a significant departure?

Yes, him also.

A *failure* in every way

1. it did not scale very well after 3.3GHz ... they expected WAAY the hell better with PressHot
2. They *canceled* its production SUDDENLY and pulled people off the project to start them on a new previously UNplanned direction
3. Intel made an announcement that they we finished with NetBurst - like never before .. and never since

Just because Intel "made money" does not P4 an *engineering success* make
. . . P4 is the very definition of failure

 

[Arkaign]

I hate to say it, but you're batsh*t crazy lol. Intel dominated AMD in the market with P4, made untold billions, and led in performance for a good deal of time. If that's failure, like I said, AMD would have loved to have had a fraction of that 'failure'. You only want to look at certain aspects, and utterly ignore all the other evidence and history, just so you can state something false.

Engineering success = creating a viable and competitive product, that was either class-leading, or neck and neck for most of it's timeframe.

BTW, they didn't suddenly pull people off out of nowhere to go work on projects they pulled out of thin air, the Banias/Merom/Conroe/etc team was separate, and had been continually working on development for years. It was just that the P4 architecture as it stood had reached the end of it's line in terms of being a viable product.

Prescott was admittedly a very poor link in the chain, and so was Willamette, but both of those were somewhat limited in the overall picture.

A product ending it's useful life cycle after years, even if the hopes were that it would last even LONGER, does not equal failure. That's a patently idiotic way to look at things. Legitimate failures are things like Microsoft BOB, the AMD K5, the Intel i740, and so on, where the product never achieves any intended performance benchmark or viable spot in the marketplace.

I'm guessing the vast majority of us here look back at the Athlon XP and Athlon 64 as the better values, and the better performers in many cases, but to think that the P4 was a failure because it didn't ramp up to a marketing target for ridiculous Ghz, or that it was a failure for being EOL'd after several years, that's ignorant.

 

[SickBeast]

Originally posted by: Scholzpdx

Originally posted by: myocardia

Originally posted by: Scholzpdx
It was the first post on failblog.

You didn't notice that this thread was 2½ years old?:laugh:

I did but the CPU forum has been dead all night. Might as well spice it up with some retro necro threaded goodness.

Should I be flattered?

Unfortuately, this thread never did answer my original question, and really just led to a bunch of personal flame wars and pissing contests. :evil:

 

[apoppin]

Originally posted by: Arkaign
I hate to say it, but you're batsh*t crazy lol. Intel dominated AMD in the market with P4, made untold billions, and led in performance for a good deal of time. If that's failure, like I said, AMD would have loved to have had a fraction of that 'failure'. You only want to look at certain aspects, and utterly ignore all the other evidence and history, just so you can state something false.

snip

Wasn't that about the time that the EU fined Intel for cheating in business and using their monopoly to cut AMD out of the market?


Intel was not successful because of the P4 - rather in spite of it

ANY time there is a major change of architecture, there is a transition
- Not with P4

Everything - *canceled at ONCE!*
- complete change of direction to a "Fall back" Plan B

that has never happened before or since at intel