Was the P4 an 'engineering failure'?

[apoppin]

you have no case

 

[zsdersw]

Originally posted by: apoppin
you have no case

You talk about your own quite well.

 

[apoppin]

of course ... i have shown links, i have quoted wikipedia and there are literally hundreds of links to show P4 was an 'engineering failure'

you otOH, have ... *nothing* ... zero ... no support and no links ... just yourself to quote ... over-and-over
:roll:

instead of giving evidence ... you *nitpick* on the opposing mountain of evidence in an vain attempt to cloud the fact that you still have ... *nothing* ... no case whatsoever

--and Intel Engineer's *cannot* speak for intel [period] ... they sign an 'unbreakable' NDA

 

[zsdersw]

Originally posted by: apoppin
of course ... i have shown links, i have quoted wikipedia and there are literally hundreds of links to show P4 was an 'engineering failure'

you otOH, have ... *nothing* ... zero ... no support and no links ... just yourself to quote ... over-and-over
:roll:

instead of giving evidence ... you *nitpick* on the opposing mountain of evidence in an vain attempt to cloud the fact that you still have ... *nothing* ... no case whatsoever

--and Intel Engineer's *cannot* speak for intel [period] ... they sign an 'unbreakable' NDA

None of your links show anything that demonstrates "engineering failure".. and an article from Wikipedia doesn't mean nearly as much as it used to.

The fact that engineers cannot speak for Intel doesn't necessarily mean the marketing dept. speaks anything backed up by the engineers.

 

[apoppin]

and still you have nothing to say ... no *case*

whatsoever

find something to support your view ... something

anything

all you *prove * is that you can nitpick someone else's evidence

i'd say you're irrelavant

 

[zsdersw]

Originally posted by: apoppin
i'd say you're irrelavant

And yet here you are still typing responses to my posts.

You didn't demonstrate how marketing-speak demonstrates "engineering failure".

 

[apoppin]

and you didn't show *anything*

it's a 'response' just like yours

again ... show something to support your view ... links, please

we have trolls in video who do the exact same thing ... they have nothing to offer or any contributions to add to a thread, so they nitpick other members posts and dispute "sources" though they have none of their own to offer ... we tend to ignore them

you are better than that ... aren't you?


come on, let's see something to support your view

 

[zsdersw]

Originally posted by: apoppin
and you didn't show *anything*

And you didn't answer my question: How does marketing-speak indicate an "engineering failure"?

we have trolls in video who do the exact same thing ... they have nothing to offer or any contributions to add to a thread, so they nitpick other members posts and dispute "sources" though they have none of their own to offer ... we tend to ignore them

If that's true, you should've ignored me a long time ago.. but you didn't.

It doesn't take a "source" to refute irrelevancies and opinions that are totally beside the point. Marketing saying "scale to 10GHz".. and pundits quoting Intel marketing in their opinions does *not* hold any water.

 

[apoppin]

OK ... since you have absolutely nothing ... i'll take your advice belatedly


the resident guys here seem react to you like a troll ... and ignore you also

i just gave you the benefit of any doubt .. since i rarely post here


let us know if you *ever* have anything besides your 'opinion'

:roll:

aloha

 

[zsdersw]

Let me know if you *ever* have anything besides marketing-speak (and pundits quoting or referencing marketing-speak) to make your case.

 

[apoppin]

i couldn't help myself

John Sweeney of UE3 fame:

http://www.nvnews.net/vbulletin/showthread.php?t=70056&highlight=msaa

Conroe is a really fantastic chip. The funny thing is very few people in the industry have been willing to come out and say that the Pentium 4 architecture sucks. It sucked all along. Even at the height of it's sucking, when it was running at 3.6GHz and not performing as well as a 2GHz AMD64... People were reluctant to say it sucked... so IT SUCKS! But Conroe really makes up for that and I am really happy to see that, that Intel is back on this track of extremely high computing performance at reasonable clock rates.

ok ..

p4 architecture sucks ... we get it ... it was an engineering failure

enginering failures suck

i have one

 

[zsdersw]

No, it wasn't.

 

[TurtleCrusher]

It was the first post on failblog.

 

[clarkey01]

I think calling it a engineering failure is untrue, people will tend to call it that due to it now being seen as an inefficient design (and even back then). It was never really left behind in AMD's dust and with the Northwood C's it had the performance crown for most things against the AXP's.

Plus just like Apple, at the time, Intel had Jo average believing that there chips/brand were superior with higher headline clockspeeds, forgetting to mention that this was due to a high pipeline count and less instructions per clock.

Failure, no, inefficient, yes, performance leader , pretty much. However the K8's gave the Prescott's a beating. The Pentium 4 HAD to happen, what Intel learned and discovered from the experience has given us the much more beneficial direction we are heading now.

 

[myocardia]

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

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

 

[TurtleCrusher]

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.

 

[deimos3428]

I don't see how anyone could claim the P4 was anything other than a major success. I've still got one chugging along at home, and you really can't call a chip that works for 10+ years a failure.

Success isn't relative. It doesn't matter if it did better than AMD's offerings, or if it was the best possible design ever. It was a vast improvement over the P3, it sold a ton of units and made money for Intel. Which happened to allow the company to employ a lot of people and keep making even better chips. You wouldn't have the i7 without the P4.

 

[apoppin]

Originally posted by: deimos3428
I don't see how anyone could claim the P4 was anything other than a major success. I've still got one chugging along at home, and you really can't call a chip that works for 10+ years a failure.

Success isn't relative. It doesn't matter if it did better than AMD's offerings, or if it was the best possible design ever. It was a vast improvement over the P3, it sold a ton of units and made money for Intel. Which happened to allow the company to employ a lot of people and keep making even better chips. You wouldn't have the i7 without the P4.

it was a flop even though it made intel money

P4 was NO improvement over PIII
- in fact, they pretty much went back to it with their "M" design

intel viewed it as an engineering failure .. however, it looks like the P4 team is back to work

... on Larrabeast
:Q

 

[drizek]

Originally posted by: deimos3428
I don't see how anyone could claim the P4 was anything other than a major success. I've still got one chugging along at home, and you really can't call a chip that works for 10+ years a failure.

Success isn't relative. It doesn't matter if it did better than AMD's offerings, or if it was the best possible design ever. It was a vast improvement over the P3, it sold a ton of units and made money for Intel. Which happened to allow the company to employ a lot of people and keep making even better chips. You wouldn't have the i7 without the P4.

Do you know how much that thing cost you over the past 10 years in electricity?

Isn't the P4 team the one did Nehalem?

 

[deputc26]

I would say that it was a failure of strategy rather than a failure of engineering. Intel's strategy of bettering clockspeed rather than u-architecture was a failed strategy. The engineers successfully designed a chip that was (to a degree) unsuccessful due to the strategy it was based on.

 

[alyarb]

their position in the industry allowed them to sell just as many netburst products as they would have sold on a better or worse design. it was a success in many respects, but it was an inefficient architecture throughout its life. i doubt intel foresaw the thermal challenges ahead of them with the high frequency 90nm parts. they went from 80 watts with 130 nm to what, 115 watts with prescott? for a little single core, that is a gargantuan increase in power and it's all attributable to the frequencies they were gunning for. all this during a time when heatsink design was still very 1990's and uninspired. usually the die shrink is intended to save power, but they had to scale to get away from K8. they really painted themselves into a corner. in the mean time it was pretty interesting to see 2 GHz K8's performing like 3 GHz pentiums. we'll never see it again, that's for sure.

 

[apoppin]

Originally posted by: alyarb
their position in the industry allowed them to sell just as many netburst products as they would have sold on a better or worse design. it was a success in many respects, but it was an inefficient architecture throughout its life. i doubt intel foresaw the thermal challenges ahead of them with the high frequency 90nm parts. they went from 80 watts with 130 nm to what, 115 watts with prescott? for a little single core, that is a gargantuan increase in power and it's all attributable to the frequencies they were gunning for. all this during a time when heatsink design was still very 1990's and uninspired. usually the die shrink is intended to save power, but they had to scale to get away from K8. they really painted themselves into a corner. in the mean time it was pretty interesting to see 2 GHz K8's performing like 3 GHz pentiums. we'll never see it again, that's for sure.

sure we will; the NetBust guys are back and evidently they have designed Larrabeast
- what, next year we may get to see a repeat of failure

at any rate it looks like the same marketing team was rebooted

 

[alyarb]

Larrabee isn't gunning for high freqs, and it's incredibly wide. it's like the anti-netburst when you think about it. plus they have HKMG now. If you could spruce up a prescott with 45 nm HKMG, give it QPI to MCH (or iMCH) and 32MB of L2 (so its about the size of a 800 M transistor chip), i would look a lot prettier than it did in 2004. netburst was not the fault of the engineers (in fact, SSE2 performance was astounding initially); AMD would not have done a better job at the same approach (at least not during the same historical era of available techniques). the guys were given a job to do and they did their best with their crude 20th century implements. larrabee could have a lot of potential depending on how efficient the software renderer is, and it certainly will help scaling wide x86 apps.

 

[apoppin]

Originally posted by: alyarb
Larrabee isn't gunning for high freqs, and it's incredibly wide. it's like the anti-netburst when you think about it. plus they have HKMG now. If you could spruce up a prescott with 45 nm HKMG, give it QPI to MCH (or iMCH) and 32MB of L2 (so its about the size of a 800 M transistor chip), i would look a lot prettier than it did in 2004. netburst was not the fault of the engineers (in fact, SSE2 performance was astounding initially); AMD would not have done a better job at the same approach (at least not during the same historical era of available techniques). the guys were given a job to do and they did their best with their crude 20th century implements. larrabee could have a lot of potential depending on how efficient the software renderer is, and it certainly will help scaling wide x86 apps.

i am saying the *potential for failure* is similar; the later it is released, the more chance the competition has to make it irrelevant

Intel's engineers were given an IMPOSSIBLE task by marketing
i.e. - "10 GHz or Bust"

in that sense it was an engineering failure

 

[alyarb]

well that was the point i was making with my original post. but yeah, i've heard that "10ghz or bust" story before. pretty funny with y2k technology.

thankfully the timeliness of larrabee's introduction is not critical to its success. larrabee's first impressions, however, may be critical. i would rather intel take the time to make an efficient software renderer to back up their architecture than to rush it and disappoint everybody. if they stall long enough, they may be able to introduce it on 32nm with 64 cores rather than 45nm with 40 or 48. larrabee's approach is totally different than geforce or radeon, so if they "fall behind" in their product cycle it will not force them into a redesign.