P4 Extreme

[Tahir2]

Everyones favourite website The Inquirer has posted some information about a new revision to the P4 CPU tentatively called Pentium 4 Extreme which has 2MB of L2 Cache. It has been designed to compete with the Athlon64 FX CPU's until 'Prescott' arrives or maybe it is actually 'Prescott?'

Anyway here is the linkage:

http://www.theinquirer.net/?article=11576

IMHO the likelihood of this CPU being released is very low.

 

[Slides]

I think it depends on how well the Athlon64 FX does.

 

[zurich]

Everyones favourite website The Inquirer has posted some information about a new revision to the P4 CPU tentatively called Pentium 4 Extreme which has 2MB of L2 Cache. It has been designed to compete with the Athlon64 FX CPU's until 'Prescott' arrives or maybe it is actually 'Prescott?'

Anyway here is the linkage:

http://www.theinquirer.net/?article=11576

IMHO the likelihood of this CPU being released is very low.

How much do you want to wager on that?

http://www.anandtech.com/IT/showdoc.html?i=1870

 

[pcchen]

Some preliminary benchmarks at Aces Hardware. Most are games and show about 5% ~ 15% improvement over P4 of the same clock speed.

Cinebench shows almost no improvement, however.

 

[Accord1999]

Cinebench shows almost no improvement, however.

Doesn't rely on cache or memory bandwidth very much... Doesn't matter anyways, since the P4's score of 380+ is a 100 points better than an Opteron 146.

 

[jvd]

prescott isn't fast enough to compete with the athlon 64 fx . Amd 64 is overclocking to 2900mhz already. So amd can easily scale it to 2600 or 2700 on .13 micron. With prescott delayed there is no way intel can keep up with it . Oh btw the 2900mhz athlon gets a pie test score of 30 the slowest p4 that can do that is 4.8ghz

 

[Saem]

Overclocking shouldn't be used to judge the headroom of a chip.

Overclockers don't have the ability to appropriately validate the chip. Without that you can't really tell if there is headroom.

In anycase, we'll see. Having 3 full decoders in the critical path and the seemingly short pipeline don't seem like a good idea.

 

[zurich]

Doh, Ace's pulled the review.

 

[jvd]

Overclocking shouldn't be used to judge the headroom of a chip.

Overclockers don't have the ability to appropriately validate the chip. Without that you can't really tell if there is headroom.

In anycase, we'll see. Having 3 full decoders in the critical path and the seemingly short pipeline don't seem like a good idea.

Well in all fairness I know amd has two surpise chips for this year. Both will be more than a paper launch.

 

[AAlcHemY]

Well in all fairness I know amd has two surpise chips for this year. Both will be more than a paper launch.

Can you give more info on that one ?

Anyway, here is the Fx compared to the EE.

 

[jvd]

Well in all fairness I know amd has two surpise chips for this year. Both will be more than a paper launch.

Can you give more info on that one ?

Anyway, here is the Fx compared to the EE.

From what I hear (don't take this to seriously) Is that amd has 2.8ghz .13 chips air cooled. The yields are very bad but the yields on the 2.4 and 2.6 are very very good and the 2.4 might see a launch tommorow and the 2.6 in nov to december depending on how intel reacts. There is also one more flavor of the fx. Remember all fx chips have been tested on the nforce 3 chipset which has ht at 600 and not the 800 of the via and ali chipsets. So the tests at aces are not vaild . IF u happened to catch them .

 

[3dilettante]

Just how invalid exactly would the benchmarks be with a reduced HT frequency?
How close to saturating the 600mhz bus do any of the benchmarks come?