P4 Isn't Ready for AMD

[Rigoletto]

sure Havoc. I said that I thought the P4 was actually competing pretty well considering its clock was the product of using a 20 stage pipeline compared to 12 of the PIII... but nobody really got the point did they. I wonder how many on these forums actually have understanding of what they are using.

 

[Noriaki]

That's all good and well but these benches are comparing what? A 1Ghz P3 and a 1.4 Ghz P4?
That's fine, but look at the very first P6 core, the PPro at 150Mhz.
Now they have a P3 at 1Ghz
Apply a similar scale to the P4 and you have it topping out a 10Ghz.
That's an exageration, but really the P4 is maybe not much better than the P3 clock/clock, but when you consider how much trouble intel is having getting anything about 1Ghz out, and how much higher the P4 can go, in the long run (with p7 core revisions) the pentium4 will slaughter the P3. It's hardly fair to compare a beta p7 at it's lowest clockspeed to a super tweaked 6th revison 8th stepping p6 at it maxxed out clock speed. Sure they are pretty close, but it's the potentional that's in the P4 that intel is going for here.
I have many beefs with the P4, it's performance comparison to the P3 is not really one of them though.
- It's price
- It makes the Athlon look conservative in power use.
- It needs heatsinks so big and heavy they need to be mounted to the case
- Only Rambus so far
- The next revision is going to have a different socket.

If you want to pick on the P4, there are lots of good reasons, it's minimal performance gain over the P3 isn't really a good one though.

To answer your question though, i don't really think the P4 can top the Mustang, it seems the Mustang is going to ramp up in speed quite nicely and I really doubt the P4 will be able to keep up with the Mustang at the same speed, hopefully intel will be able to keep a good Mhz lead or the Mustang will clobber it.



<< Maximum PC is substandard single-ply toilet paper. >>

- Modus

Damn right.

 

[GT1999]

I could start a new thread with this one, but it's closely related to this thread. Also NOTE: the P4 probably won't be out 'next month' due to some other problems Intel is having.


CNET Reports about P4 (Oct 11):



<< Although Intel will release the Pentium 4 next month, the chip won't become the breadwinner of the company's microprocessor family until at least 2002. Because of a number of marketing and technical issues, the more technologically advanced Pentium 4 will continue to be overshadowed by the Pentium III, at least in terms of revenue and units produced, for some time. In 2002, the Pentium 4 will start to overtake the Pentium III in terms of revenue, but &quot;in terms of units, it will be longer than that.&quot;

The length of the timetable derives from the size of the Pentium 4. The Pentium 4, when it debuts, will cover approximately 217 square millimeters, more than twice the area of current Pentium IIIs and larger than the 180 square millimeter chip expected earlier by analysts. The principal problem surrounding the chip's size is that it greatly reduces the number of chips that can be produced form a single silicon wafer. The larger size means the Pentium 4 will cost around $80 to $90 to manufacture, more than double the $40 manufacturing cost of the Pentium III. Intel will shrink the Pentium 4 and the Pentium III to a more financially attractive size when the company converts from the current 0.18-micron manufacturing process to the more advanced 0.13-micron process.

Shrinking the processors will increase the number of chips that can be produced per wafer. In addition, it will cut the manufacturing cost of each chip and allow Intel to speed them both up. These smaller chips made on the 0.13-micron process, however, won't appear until the first part of the second half of 2001. When it debuts, the Pentium 4 will work only with computers containing Rambus memory. Intel is coming out with a chipset, the piece of silicon that connects the microprocessor to the rest of the computer, that will allow PC makers to use less expensive DDR DRAM memory. >>



Source: CNet

 

[JCholewa]

OuterSquare: Intel's (not publicly) published numbers for final product are not very compelling. Except for specfp, which is largely memory throughput dependent.

boot Magazine, I never really got into until its last few months. It kicked ass.

MaximumPC kicked ass for its first few months. Then it started sucking, quite bigtime. Yes, I do think the Celeron/Duron review was the last straw on their reliability marker.

-JC

 

[KarsinTheHutt]

Tom's Hardware seems fairly optimistic about P4 - but this attitude is conditional on the availability of a P4/DDR solution. The P3-Tualatin seems to be quite promising too.

 

[ElFenix]

i thought when i bought my pro back in 1997 that i wouldn't be buying another p6 design... then i bought a p2 400 and a p3 700... maybe tualatin will kick enough ass that i'll buy one of those instead of an athlon or p4.

btw- the team that designed the p6 also designed the p7. the team that designed the p5 designed ia-64. and yet, the p6-p7 team is intel's second stringers...