"P4 Northwood and Prescott Comparison at 4.1 GHz "

[oymd]

http://www.tomshardware.com/cpu/20040517/northwood_prescott-03.html


I previously critisized the article on the dual Xeons....and now....

Will you please browse to the link I provided above....You will find the CPU-Z captured images of BOTH the Prescott and the Northwood CPUs....

The article says BOTH were 3.2GHZ CPUs......however check the CPU-Z data of the NORTHWOOD....:

IT SAYS : 3.4GHZ !!! AND THE MULTIPLIER IS : 16 !!????

What the heck....this cant be just a simple typo...that's a Screen Capture and not typed???

What's going on??

 

[juin]

I you are so good at observation you will see 2 little pack of pixel that form (ES).

i need to change useur name.

 

[Crashman]

Tom's wouldn't have fudged the 3.4 part on the 3.2, and CPU-Z does make mistakes reading codes. 16x is the proper multiplier for a 3.2.

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[endyen]

All in all, the review seemed rather pointless. Oh well, anything to give Intel some advertizing.
For those who would say no, why was there no mention of an Amd chip?

 

[trooper11]

yeah i wonder when we wil see the amd overclcoking article on tom's stil havent seen one since the original athlon fx one, but that was comparing against an intel chip. id like to see one with the athlon 64, fx, or even opterons...

 

[oymd]

Crash....I know 16 is right...but do u see CPU-Z make mistakes often..seriously

 

[juin]

No one have see that a ES sample multiplier can be change.
Toms have use lower multiplier.When you overclock a AXP 2500= to 2.X it still say 2500+.

i need to change useur name.

 

[Crashman]

Yes, whenever the PCU code or stepping is newer than the version of CPU-Z, the chances of getting the right report is around 50%.

If you don't believe that, I can take some screenshots of from my own system using different versions.

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[Crashman]

Yes, it's possible Tom's is using an ES 3.4 with the multiplier set at 16x in BIOS, but I've also seen CPU-Z and WCPUID report some wierd stuff, including a 360MHz bus on my Tualatin Celeron (on a BX board at that!).

An ES 3.4 at 16x would give the same performance numbers as a standard 3.2, but might overclock higher. That's irrelavent in this comparison, which only compares the PERFORMANCE of the Northwood core to the Prescott, and in order to be a fair comparison both cores would have to use the same multiplier.

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[juin]

It write ES just next to the CPU name.Second thing Nw use higher multiplier that why toms drop it to 16 to have equal FSB and mghz with the prescott.

That not a mistake from the software or toms.Tha lack of observation and in some way wisdom.

i need to change useur name.

 

[trooper11]

ouch that was harsh lol

 

[Coop]

So in gaming, the Northwood is a better buy. with other words : The prescott is slower then northwood, and this in every speed !
Xeon, what you have to say about this ? same for you Fugger !


Toms Hardware Site is a joke !

 

[Crashman]

If YOU had wisdom you'd realize who you're talking to and avoid the quips. I just glanced at it, I didn't care either way at the time.

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[oymd]

Ok...so just to put things in perspective....

If the Prescott, according to that article, DOES NOT scale better at higher speeds or perform any better than the Northwood at those speeds, what was the whole point of the Prescott....

Is it to allow higher GHZ speeds upto 5GHZ...which the Northwood cant reach with its .13 manufacture??

If so....what will be the next performance processor to go with the 775 socket??

 

[oymd]

"Yes, it's possible Tom's is using an ES 3.4 with the multiplier set at 16x in BIOS,"

Didnt Tom say they're using TWO 3.2GHZ CPUs??

"We used two conventional 3.2-GHz processors that can be bought at retail channels for this test. They thus reflect performance results that can be achieved with commercially available processors."

 

[ChipDeath]

a Northwood with a 16x multiplier can be called a 3.2Ghz NW. Whether it was bought with that fixed multiplier OR was an engineering sample with it's multiplier set to 16x, it'll perform exactly the same. being an engineering sample it's possible it would perhaps overclock better than some others, but that could be true of any core pretty much. Technically speaking if it was an ES 3.4, and they stated that it was a "Normal 3.2Ghz you could buy in a shop" then that IS a lie, but it's not one that will actually effect the benchies in any way, so doesn't bother me personally.

I'm not remotely suprised that a 4Ghz NW is faster than a 4Ghz Scotty - you're right about the point behind the prescott being to (in theory) allow higher clockspeeds. It's just like the PIII compared to a P4 - a 4Ghz PIII would blow both the NW and Scotty away, but the PIII simply can't reach that speed. If anything, I'm suprised at just how close in performance it is to a NW core at the same speed.

---
Epox 8RDA+ rev1.1 w/ Custom NB HS
XP1700+ @205x11 (~2.26Ghz), 1.575Vcore
2x256Mb Corsair PC3200LL 2-2-2-4
Sapphire 9800Pro 420/744

 

[juin]

What is mean is those are simple Core with unlock multiplier.Those are the same as retail CPU.Intel have use special core for IDF overclocking test often.Toms was refering to this.

i need to change useur name.

 

[Crashman]

Yes, it's also possible that the person who wrote the article wasn't the person who performed the test and there was a comunication problem which resulted in this inaccuracy (whether intentional or mistaken). Or it's also possible CPU-Z was wrong.

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>

 

[Crashman]

Yes, the whole point of the Prescott was to create a new core which could reach 5GHz with negligable performance impact to the core change. The problem is, the people who designed the core probably couldn't forsee the problems Intel is having moving to 90nm.

Personally I would have left the Northwood alone and tried moving that design to 90nm. But Intel likes to have a little breathing room in case things don't go as planned (which they apparently didn't).

<font color=blue>Only a place as big as the internet could be home to a hero as big as Crashman!</font color=blue>
<font color=red>Only a place as big as the internet could be home to an ego as large as Crashman's!</font color=red>