Prescott Questions

[bigdawgfoxx]

Ok, Intel has released Prescott, but why?

Read Here (http://secure.newegg.com/app/CustratingReview.asp?item=19-116-174)

Those are some customer reviews off of newegg talking about how even though it has double the cache, the pipeline is much longer so it doesnt run ANY faster.

They also say it runs ALOT hotter.

The CPU is NOT 64bit, so why did they release it when the 32bit P4 C is doing great? Why did they not save money on research and make their 64bit cpu INCREADIBLE. Just a few questions im having... :unsure:

[Smurfette]

Originally posted by bigdawgfoxx@15 February 2004 - 18:39
Why did they not save money on research and make their 64bit cpu INCREADIBLE.
Perhaps they don't know how to make anything INCREADIBLE. :)

[bigdawgfoxx]

lol maybe...but still

they just realeased this chip that is hotter and slower...whats the freaking point?

[callum]

Originally posted by bigdawgfoxx@15 February 2004 - 17:39
Those are some customer reviews off of newegg talking about how even though it has double the cache, the pipeline is much longer so it doesnt run ANY faster.

Have you read any other reviews? I wouldn't base my opinion of something on a couple of customer reviews.

http://www.anandtech.com/cpu/showdoc.html?i=1956

[Mad Cat]

The Prescott has a 90nm process compared to the 120nm of current Northwoods.

Look here (http://www20.tomshardware.com/cpu/20040201/prescott-10.html#opengl).

The ones with a single E are Prescotts, double E (EE) are Extreme Edition Northwoods, no Es are Northwoods. You'll see that in some places they do beat Northwoods.

This slowdown on lower chips is also needed for higher clockspeeds later. Sounds like they've released a flawed product, doesn't it.

EDIT: I also heard something about Prescott being a cheaper alternative and Intel trying to reclaim the budget market, although I can't remember the source and whether there was any truth behind the comment.


Originally posted by Tom&#39;s Hardware+--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (Tom&#39;s Hardware)</td></tr><tr><td id='QUOTE'>
In our opinion, the debut of a processor based on a cutting-edge process technology and architecture that only offers the same performance level as the predecessor is unanticipated. Despite 1 MB L2 cache and some optimizations, Prescott is slower than Northwood in roughly a third of our benchmarks. Software, like many 3D shooters and more serious applications like Lame, MS Movie Maker 2, Mathematica, Cinema 4D or even 3DStudio perform worse than before.
On the other hand, there are a similar amount of applications that run faster on a Prescott CPU. These are DivX encoding with Xmpeg, file archiving with WinRAR, video authoring using Pinnacle Studio 9 and the remolded SYSmark 2004. [/b]

Faster or similar on majority (2/3 of the benchmarks) doesn&#39;t sound bad to me.

Now I understand:
<!--QuoteBegin-Anandtech
Intel wants to shift all Pentium 4s over to Prescott as soon as possible, mostly because once production ramps up it will be cheaper for Intel to make a 112 mm^2 Prescott than it is for them to make a 131 mm^2 Northwood. Therefore Prescott launches at clock speeds that are equivalent to currently available Northwoods. [/quote]

[atiVidia]

i play games. i like intel


Originally posted by Tom&#39;s Hardware
Despite 1 MB L2 cache and some optimizations, Prescott is slower than Northwood in roughly a third of our benchmarks. Software, like many 3D shooters and more serious applications like Lame, MS Movie Maker 2, Mathematica, Cinema 4D or even 3DStudio perform worse than before.

i dont like them anymore

amd, you got urself another loyal customer




im sure alot of ppl will agree with me on this front

[adamp2p]

Well bigdawg, you should do a little bit of reading and research to find out a little bit about the Prescott processor. I have done a little bit every day, and now I have a pretty good understanding exactly what makes it tick...

For one thing, Intel extended the pipeline from 21 stages to 31 stages.

Imagine a CPU processing data. The fewer stages, the more "work" each stage has to make, thus the hotter the CPU gets. Keep in mind that the CPU&#39;s operating frequency is, to a certain extent, limited by the heat it generates. Therefore, theoretically, extending the pipeline by 10 stages should cool down the CPU, because each stage peforms less "work." However, modern CPU&#39;s are not that "simple," so there are several other elements that come into play when measuring effectiveness of a CPU in processing data.

When Intel first introduced the Pentium 4 CPU, it was slower than the Pentium 3 at equal clockspeeds, and the entire world cried "fowl." The P3 had only 11 stages in its pipeline and the P4 had 21. One thing that you need to understand about extending the pipeline of the CPU is that it enables, ultimately, the CPU to operate at higher frequencies. However, if one of those stages is unable to complete its "work," then the data that was being processed at that level must return all the way to the beginning and start over, thus wasting precious clock cycles and effectively hindering the processor&#39;s ability to process data efficiently.

This is a quick and incomplete summary; you should really do some reading at Anandtech or HardOCP to gain a full understanding.

;)

[atiVidia]

Originally posted by adamp2p@15 February 2004 - 15:36
For one thing, Intel extended the pipeline from 21 stages to 31 stages.
HOLY FUCK&#33;&#33;&#33;

31 stages??? that could mean that an aXP would only need to run at 2400 to achieve 3400+

god damn...

[bigdawgfoxx]

Originally posted by atiVidia+15 February 2004 - 15:36--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (atiVidia @ 15 February 2004 - 15:36)</td></tr><tr><td id='QUOTE'> <!--QuoteBegin-adamp2p@15 February 2004 - 15:36
For one thing, Intel extended the pipeline from 21 stages to 31 stages.
HOLY FUCK&#33;&#33;&#33;

31 stages??? that could mean that an aXP would only need to run at 2400 to achieve 3400+

god damn... [/b][/quote]
what are you talking bout ati? dont understand that lol

Thanks adam that helped me understand it. But those people said it ran hotter, not cooler as more stages should. also arent the pipelines longer so it takes longer to get data from point a to point b? I guess its kinda like you said that the P4 was slower then the P3. I guess it will get alot faster as they develop it..but why release it if its not ready? And is this actually pentium 5? or just a diff core P4 like the barton of amd?

[atiVidia]

Originally posted by bigdawgfoxx+15 February 2004 - 17:53--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (bigdawgfoxx @ 15 February 2004 - 17:53)</td></tr><tr><td id='QUOTE'>
Originally posted by atiVidia@15 February 2004 - 15:36
<!--QuoteBegin-adamp2p@15 February 2004 - 15:36
For one thing, Intel extended the pipeline from 21 stages to 31 stages.
HOLY FUCK&#33;&#33;&#33;

31 stages??? that could mean that an aXP would only need to run at 2400 to achieve 3400+

god damn...
what are you talking bout ati? dont understand that lol

Thanks adam that helped me understand it. But those people said it ran hotter, not cooler as more stages should. also arent the pipelines longer so it takes longer to get data from point a to point b? I guess its kinda like you said that the P4 was slower then the P3. I guess it will get alot faster as they develop it..but why release it if its not ready? And is this actually pentium 5? or just a diff core P4 like the barton of amd? [/b][/quote]
the reason AMD&#39;s slower chips run as fast as some of intels faster chips is bcuz AMD has less stages per clock cycle. the clock is slower, but less time is spent on clock anyway so...

also, since the intel pipes are now LONGER, it means that if there is a data transfer error, there could be a bigger penalty:

what i mean by that is that a processor guesses what comes next. if a pipe is longer, the guesses are harder, and thus, if the chip screws a guess, it has to make it up with a speed penalization. the longer the pipe, the bigger the penalty

therefore: AMD, with less stages per clock, has 2 advantages:
1) slower clocks get very decent performance
2) shorter pipe means less penalization on an error


now that intel has added 10 stages, AMD chips are likely to need even less speed and STILL tie up with intel, unless AMD raises the number of stages as well

[bigdawgfoxx]

oh alright, cool.

But if making more stages slows it down...why is intel making more stages??

[atiVidia]

Originally posted by bigdawgfoxx@15 February 2004 - 18:04
oh alright, cool.

But if making more stages slows it down...why is intel making more stages??
im sure that theres some sort of advantage

got no clue as to what it is tho :lol:

[Keikan]

I still would buy an AMD

[Mad Cat]

Originally posted by bigdawgfoxx@15 February 2004 - 23:04
oh alright, cool.

But if making more stages slows it down...why is intel making more stages??
The Prescotts allow for higher frequencies, hence better performance in the long run.

[bigdawgfoxx]

So will they have those things at like 4Ghz soon? And with that many stages will it be that much diff then a 3.2Ghz with 21 stages and shorter pipelines?

1. 21 Stage P4C with shorter pipelines at 3.2Ghz
2. 31 Stage P4E with longer pipelines at 4.0Ghz

Would number 2 win?

[Mad Cat]

Originally posted by bigdawgfoxx@15 February 2004 - 23:40
So will they have those things at like 4Ghz soon? And with that many stages will it be that much diff then a 3.2Ghz with 21 stages and shorter pipelines?

1. 21 Stage P4C with shorter pipelines at 3.2Ghz
2. 31 Stage P4E with longer pipelines at 4.0Ghz

Would number 2 win?
Easily. The Prescott only just loses to the Northwood anyway.


Read that stuff on Anandtech.

[bigdawgfoxx]

so at the same speeds more stages will be slower...but more stages allow for greater speeds in the longrun which will be faster...i think i got it lol

[adamp2p]

Originally posted by bigdawgfoxx+15 February 2004 - 14:53--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td>QUOTE (bigdawgfoxx @ 15 February 2004 - 14:53)</td></tr><tr><td id='QUOTE'>
Originally posted by atiVidia@15 February 2004 - 15:36
<!--QuoteBegin-adamp2p@15 February 2004 - 15:36
For one thing, Intel extended the pipeline from 21 stages to 31 stages.
HOLY FUCK&#33;&#33;&#33;

31 stages??? that could mean that an aXP would only need to run at 2400 to achieve 3400+

god damn...
what are you talking bout ati? dont understand that lol

Thanks adam that helped me understand it. But those people said it ran hotter, not cooler as more stages should. also arent the pipelines longer so it takes longer to get data from point a to point b? I guess its kinda like you said that the P4 was slower then the P3. I guess it will get alot faster as they develop it..but why release it if its not ready? And is this actually pentium 5? or just a diff core P4 like the barton of amd? [/b][/quote]
Good question big dawg. The chip "should" run cooler, but it does not.

Keep in mind that the engineers who helped develop the Prescott were under IMMENSE pressure from Intel to release the Prescott, largely due to the incredible success of AMD&#39;s Athlon 64 line of CPU&#39;s. In fact, the Prescott was orginally slated to be released on the second or third quarter of 2003. It was subsequently delayed until the forth quarter of 2003, and then again until February 2004. This implies that Intel was having trouble with Prescott; however, this can be expected because Intel, we must remember and realize, are the first to release a chip that is based on the 90 nanometer process.

You have to understand that developing a microprocessor is not something you and a couple of friends could do in your free time; the folks who are behind these technologies are not your average joes, okay...

There are several reasons why these chips may run hotter and have higher voltage requirements than we would expect at this juncture. However, I do not have a degree in tenser equations of the third degree; we are talking about the speed of light here, big dawg. There are all kinds of physics you need to know to understand what is happening with transistors at the molecular level. If this excites you, and you want to learn more about it, there are volumes of information available on the internet. However, you will not understand it without a few years of advanced physics and engineering.
;)

[DWk]

Yeps, if you are REALLY interested, maybe you could find out about taking some MIT or Berkeley courses, or just some local tech college courses.