Northwood 3.4 GHz review

Toggle sidebar Toggle sidebar
B

blue_heart

Distinguished
Oct 29, 2001
298
0
18,780
<A HREF="http://www.hardwareseeker.com/resourcelink.html?rlid=71362" target="_new">http://www.hardwareseeker.com/resourcelink.html?rlid=71362</A>

wish if there was UnDo in the life
 
Mephistopheles

Mephistopheles

Distinguished
Feb 10, 2003
2,444
0
19,780
I saw that already... It's a great review. It indicates that the Gallatin-based 3.4Ghz EE P4 will probably be quite fast. I hope prescott turns out great too... but that remains to be seen...

:evil: <font color=red><b>M</b></font color=red>ephistopheles
 
Mephistopheles

Mephistopheles

Distinguished
Feb 10, 2003
2,444
0
19,780
Good question... I thought this 3.4Ghz Northwood was under NDA...

Anyway, I still think that a 13x266 Prescott (3.46Ghz, 1066Mhz FSB) can probably still be paired with (the already highly suggestive) DDR-II 533 and compete reasonably with A64 tech by Q2. Intel would then only have to go agressive on AMD by Q3, with a 14x266 Scotty (3.73Ghz), and then a 15x266 (4Ghz) Scotty by Q4. This is probably a hell of a lot easier than going 64-bit, if they don't have it already. But hopefully, Intel will know what to do better than me.

:evil: <font color=red><b>M</b></font color=red>ephistopheles
 
V

Vapor

Distinguished
Jun 11, 2001
2,206
0
19,780
I don't believe that the review is legit--the 3.4C scales too well, IMO.

I personally can't wait for the 1066MT/s Scotties...they should be much better than what we've seen so far.

Maxtor disgraces the six letters that make Matrox.
 
B

blue_heart

Distinguished
Oct 29, 2001
298
0
18,780
do not you guys think the review is biased toward intel?
why they did not use the ahtlon64 3400+ and the 3200+ with non ecc - non reg memory? so they can balance the cpus from both sides?
it is meaningless to use the opteron, and many benchmarks are encoding benchmarks, and it is well known that intel outperforms amd in most of encoding softwares whether audio or video.

i am not saying p4 3.4 is bad, intel's northwood cpus are great and amds 64 as well, but IMO the article is not fair.

wish if there was UnDo in the life
 
V

Vapor

Distinguished
Jun 11, 2001
2,206
0
19,780
Intel doing something silently?! Shocking! They claimed Feb. 2nd before, oh well...I guess they had a need after the great A64 3400+ was released.

Maxtor disgraces the six letters that make Matrox.
 
V

Vapor

Distinguished
Jun 11, 2001
2,206
0
19,780
I read at X-bit that Scotty, the 3.4 Northwood and the 3.4EE would all be released Feb. 2nd. Oh well...still in shock that Intel released it silently...probably in anticipation of Scotty--not wanting to take away any hype.

Maxtor disgraces the six letters that make Matrox.
 
G

Guest

Guest
>I personally can't wait for the 1066MT/s Scotties

Let me guess.. 6 months ago you couldn't wait for prescott, and I predict in 6 months you won't able to wait for Tejas... don't hold you breath mate, 3.4 Northwood is as good as is its going to get for quite some time. Why do you think intel released it in the first place if Prescott is better ? We know prescott is cheaper to make.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
C

castle

Distinguished
Apr 2, 2002
102
0
18,680
Prescott is definitely not cheaper to make.
Given everything else equal, similar part built on 300mm + .09u process should be cheaper than on 200mm + .13u, but the former is new while the latter is an on-going mature technology. There are lots of contributors to the final manufacturing cost such as yield, equipment investment vs. devaluation, etc.
In addition, 3.4G P4 can be just high bin sort accumulated over a long period, rather than a new batch. If Intel has hundred of thousands of P4 that can run 3.4G, why do not sell them on a 3.4G rating? There is no conflict with Prescott.
When AMD announced a bunch of parts with confusing PR rating, especially AXP 3200+ vs. A64 3200+, how come we didnot trolls raises the similar questions, like why AMD bothers to make them?
There are too many speculations on this board.
 
G

Guest

Guest
>We know prescott is cheaper to make.
>explaination please

Sure: it means prescott is less expensive to produce than Northwood, in other words, it doesn't cost as much, or like I said, it is "cheaper". Alternatively, you could say Northwood is more expensive to produce than Prescott, it costs more, or is less cheap. I hope that clarifies things for you.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
G

Guest

Guest
>Prescott is definitely not cheaper to make.
>Given everything else equal, similar part built on 300mm +
>09u process should be cheaper than on 200mm + .13u, but the
>former is new while the latter is an on-going mature
>technology.

Correct, I should have been more precise, and say variable cost of Prescott is less than Northwood. Of course, if you take into account the write offs of new process equipment, you could argue Prescott is more expensive for now, but those write off should be seen over Prescotts life, not just one initial product. If not, there would never be an argument to move to a smaller process.

>When AMD announced a bunch of parts with confusing PR
>rating, especially AXP 3200+ vs. A64 3200+, how come we
>didnot trolls raises the similar questions, like why AMD
>bothers to make them?

Oh lots of questions where raised, and rightfully so. But no one ever claimed A64 was cheaper to produce than Athlon XP, both products are distinctly different (other marketing name, other sockets, vastly different performance characteristics). Lastly, AMD did not release a faster Athlon XP after it introduced Athlon 64. If we had seen a AXP 3400+ launched even before a A64 3200+, rest assured I would have asked questions.

Anyway, we'll know for sure on monday, but I'm pretty sure it will be a while before Prescott is intel's fastest cpu. It will probably take >>3.4 GHz to overtake Northwood in most benches, and a whole lot more to overtake the 3.4 P4EE.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
Kanavit

Kanavit

Distinguished
Jan 6, 2004
390
0
18,780
LOL AMD got raped in that review. good thing the A64 3400+ or 3200+ wasn't included in the review, it would have been a disaster for AMD.

------
:evil: <b><font color=blue>Intel rules</b></font color=blue><A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=1817959409" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=1817959409</A>
 
V

Vapor

Distinguished
Jun 11, 2001
2,206
0
19,780
You're right, I can't wait for all of them :smile: . I also can't wait for Troy, which will probably be the CPU(s) I'll end up upgrading to.

Anyway, Intel released the 3.4GHz Northwood because they can make money off of that too, without taking away from Scotty's market. Think about it this way--if the Scotty doesn't do well, enthusiasts and those in the know won't buy it over a cheaper Northwood. So those people buy 3.2GHz Northwoods, but Intel doesn't make as much money off of them because of the pricecuts...so they release the 3.4C (which shouldn't cost any extra to make) and have a rediculously high MSRP, meaning they see more money in the end. If Prescott is a champ, the 3.4C still has a market by being a cheaper 3.4GHz processor (remember, most people only care about GHz or PR rating), which also will help the sales of Scotty, with the too often used mentality of: "It's more expensive therefore it's better."

In the end it will most likely become the fastest Socket 478 chip made, which, as I said before, enthusiasts will lean toward.

Anyway, I can't wait for Troy!

Maxtor disgraces the six letters that make Matrox.
 
J

juin

Distinguished
May 19, 2001
3,323
0
20,780
Simple math for you

Northwood

55 million transistor 300MM waffer 130 nm 2.1 SRAM cell about 85% yield with some peak around 90%

Prescott

120 million tansistor more tha twice 300MM wafer 1.0 SRAM cell yield my guess around 45% by now.

I dont see how that cost less to intel.

I dont like french test
 
G

Guest

Guest
So you know intel's yields ? and 45% ? Last time I listened to Pat telling us how healthy 90nm process was, he assured me 90nm yields and defect density where as good as with 130nm. He had some nice graph to back this up too. But maybe Pat Gelsinger doesnt know as much as you do ?

Transistor count is a moot point btw, die size matters. Prescott is smaller, yields are similar (according to intel), so prescott is cheaper unless binsplits are a lot worse at the same clock (which should be quite surprising, given prescott has a longer pipeline, and is built on a smaller, more advanced process.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
G

Guest

Guest
>In the end it will most likely become the fastest Socket
>478 chip made, which, as I said before, enthusiasts will
>lean toward.

Yep, and my guess is, it will be Northwood (at same clockspeed). The 3.4 NW is for the benchmarks and enthousiasts "in the know", 3.4 Prescott will later be the cheap cpu to sell to oems and to unwarry custumors that think they get a better deal, especially with the 1 MB cache and "SSE3". Since northwood and prescott will cost the same for the same clockspeed, I really don't see what point there would be in releasing a 3.4 NW now, unless either Prescott is slower, or intel has difficulty scaling prescott that 'high', or both.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom