Why don't AMD and Intel seem to get that the process is dead?

[CandymanCan]

I can tell you one thing tho, if i had the money to offord Intel i would buy a prescott simply because i have never owned a Pentium P4.

It would be nice to see my computer @ 3ghz and above, cuss iv been @ 2ghz since the Palamino 2000+ i had like 2 years ago. But before i do buy a prescott i would prob wait until the problems are fixed with it, but by that time id prob buy an A64.


Altho you say the 120mm Process is more matured then the 90mm process which is true but ithin ksomeone said u cant compare the Mature to the imature, why not ?

Its just Like Ati with its .15 compard to Nvidia .13 process in Video card's.

 

[deathBOB]

But thats old technology going from .15 to .13... Intel and AMD and all the other chip companies did that years ago, its old, proven technology...

The new the 90nm process is new to everyone! Their still working out the kinks.

 

[rudnik68]

Processors speed increases aren't the only reason to change processes. AMD and Intel want to move to a smaller feature size because it decreases their raw material costs per chip. A wafer of silicon costs the same whether you are using 130nm or 90nm technology. Since the die size of the newer process is smaller, you get more chips per wafer, making the production cost for each chip lower.

Intel wants to trasnition the entire P4 line to 90nm tech as fast as possible because their manufacturing cost for each chip will decrease substantially, increasing thier profit margins. Comparing Northwood (134mm^2) to Prescott (100mm^2) on 300mm wafers, you got an extra 180 dies per wafer by moving to the new process. At $200 per die, thats an extra $36,000 of income per wafer.

Edit:
http://www.xbitlabs.com/news/cpu/display/20040522115100.html

If Xbitlabs is right (they usually have good info) and AMD uses 5000 200mm wafers per week @ 20% market share, then Intel uses nearly 8900 wafers per week at 80% market share. Switching to a smaller process allows them to produce an extra 1.6 million processors per week for a potential income increase of 320 million dollars per week!. Paying $2B for a fab is reasonable when it pays for itself in 7 weeks.

 

[strife045]

My 3.0E died after a week at 1.63 V with Mach 2 cooling and took my IC7 MAX 3 with it.

 

[CandymanCan]

But thats old technology going from .15 to .13... Intel and AMD and all the other chip companies did that years ago, its old, proven technology...

The new the 90nm process is new to everyone! Their still working out the kinks.

Umm yea but i was using that as an Exsample ... nm you dont understand what im talking about.

 

[dippy_skoodlez]

Yes, you need to do much more cooling on a Prescott compared to the Northwood. Just like you need to do much more cooling with an AMD system compared to a Northwood.

The higher the Prescott is overclocked, the more it shines, assuming you can control the extra heat. As the Prescott matures and new steppings arrive, they will keep getting better. By later this year, 4 gig O/C's will be common place. The point is 90nm is here to stay and it ain't dead.

Alright.. I was trying not to interfer here, but WHAT? NW cooling < AMD's? WTF are you smoking dude? Im runngin 38C @ 2480mhz 1.85Vcore. Case temp of 33C with a tornado. 90nm isnt dead, but a 3.9ghz presshott= roughly a 3.6-3.7Ghz NW man. Raw numbers are worthless nowadays. Figured of all people, an overclocker would realize that. Not only is it equivilant to a slower CPU, but it also runs HOTTER! Screw presshott. I like my 38C. Compared to most P4's my cooling is much less, I have 2 80mm fans in the back, and an 80mm fan on the window. TONS of cooling? with the fan grills that are on the things right now, its around 20CFM each. As much as I know intellit fanboyz wont admit it, AMD has a very good heat lead in this race. Along with keeping up, if not doing better.

 

[dippy_skoodlez]

Quite an exaggeration. Mind sharing your source of info?

Have you not read [h]ardOCP's melting of the standoffs on their testing?!! Maybe you could giv us your source of where it IS running fine? I have seen heat issues left and right.

 

[Gautam]

Most overclocked Northwoods and Bartons are quite close in terms of how much heat they dissipate. It's rather incorrect to say either is significantly hotter than the other, especially when using BS being spouted out by hopelessly inaccurate motherboard sensors.

But if AMD had a heat race, they've come close to losing it with the A64. The temps I'm getting with this thing are ridiculous, and not surprisingly either, an enormous core, huge cache amount and an on-die memory controller. If Intel's having any difficulty with the .09 process, AMD's only set to have even more.

Not everyone in the world is an overclocker who has 20 degree case temp. The Joe sixpacks and John Smith's may have 90 degree room temps (not everyone can afford A/C), and closed cases, with case temps reaching 55 C, which doesn't help the mosfets one bit.

This is a very interesting and very true point. I know that many are getting very nice overclocks out of the Prescott, but all across the board, this is accomplished with beefy aftermarket cooling. Correct me if I'm wrong, but I've heard that 70C temps at stock with the stock cooler are rather the norm. The problem that this poses is in its scalability. Prescotts were originally designed to run at speeds well in excess of 4 GHz, however this would be extremely difficult to accomplish inside of your average dust-filled Compaq sitting in a Best Buy. The Prescott was a huge leap forward, but by no means is it even close to a mature product, in terms of its potential for the future.

 

[johan851]

Yay, another Intel vs AMD thread, where we can all bash each other until are e-units are just massive. Too graphic? Sorry...

I do agree that the title of this thread is a little rediculous. The process is far from dead - it's just a hurdle. Remember the Williamettes and T-Bred A's? They...well...they sucked, at least for us overclockers. The problems are worse this time around, but I'm sure they'll be overcome. It's just going to take some time on both ends. Intel got 90nm first, and seems to be suffering the worst of the problems. AMD took 90nm a little more slowly, and as a result (I think) they're ending up less severe problems. Not no problems, not little problems...just less severe.

Both companies will overcome, and I think both companies will, after time, produce some very good products.

 

[Dukemurmur]

Each company is doing there best in there own ways or ideas on how they think that they should handle the way to make there chips.....Prescott is a rather hot running chip at 90nm but u also take the 3.4EE (Extreme Northwood) at 130nm and the 3.4E (Prescott) and they run around the same temp! so when u think about it the Prescott is hot yes but 4 new tech it is a HUGE step ahead. Intel wanted to go 90nm AMD wanted 64bit.....as of now in the respects of "CURRENT" tech(tech that can be bought from newegg not like the Donthon(sp?) that is only in OEM laptops) I would have to say AMD is ahead as there 64 bit is up and working and is selling in OEM comps left and right(nothing like XP's or P4 but selling) unlike the Prescotts wich seem to "hot" 4 the OEMs to "touch" right now.

I am an AMD man all the way and probably allways will b but i am looking forward to Intel "perfecting" the Prescott as believe it or not i would LOVE to own one....But the MHz race is dead and it seems that AMD has known this 4 sometime and intel has also they where just trying to push it as fare as they could and it has worked how man average Joes REALLY need a P4-c at 3.2? Or a 3200 barton or a 3400 A64? not many maybe the video editors and the game makers but that is a small percent. Then on the same not how many need a FX-53 or a 3.4EE almost none but it is the thing about "o hey this one is faster and i *think* that i need it to run word" that in my mind is the illusion that ppl get when they go to buy a comp from a large OEM company.

To sum it all up each company is doing what they think that they ned to do to get ahead of the other. They BOTH relize the MHz race is over and dead. And that the ppl that buy in oure MHz will now have to learn that the MHz where just a "cover" up 4 "poor" eficiency on both parts and that hey now have to ofcus on efeincy instead of pure MHz. They will both in the end come out with WONDERFULLY working chips at more likly then not ANY costs to do so.

Just my 2 cents feel free to correct ANYTHING that is wrong or dosn't seem right........

 

[aNTiChRisT]

What a degredation of the forums. Please people, take a step back and look what your saying, its quite pathetic. The AMD vs INtel war will never be won, so just quit now, be friends and enjoy overclocking.

AMD and Intel have both served me well, they are both pushing the boundaries of processors; AMD with 64/32bit cores and Intel going to the 90nm proccess. It all takes time to straighten out the kinks -- and you shouldn't be bashing you should be discussing.

~t0m

 

[{Epsilon}Stalin]

I know i sure would..... i think we all would.... hell i'd like to see THz CPU's

 

[Voltron]

you mean you wouldn't want a 5ghz p4?

..Nope. I would rather have a 3 gig dual core. Unless U giving them away, then I'll take two.

 

[{Epsilon}Stalin]

Personally i dont think that process minuturization is dead..... like has been said before its just an engineering hurdle...

but i hear all this talk of dual cores, and have read some on them...i read thta they are much slower than single cores of the same frequency...so can someone describe to me the benefits of a dual core?

 

[PunkRawk911]

The benefits of Dual core processors can already be seen in pentium4 processors with hyper-threading. Basically, if you want to run two programs at the same time, one core will handle one, and the second core will handle the other.

 

[Sunburn]

Well welll.... looks like AMD is leaking more power than a Prescott. LoL It's official. Now what do you fanbois have to say for yourself?

Socket 939:







http://www.x86-secret.com

 

[{Epsilon}Stalin]

WOOOOOOOOOOOO HOOOOOOOOOOOOOOO

thats some nice work

although prescott trails by one volt

 

[batboy]

Have you not read [h]ardOCP's melting of the standoffs on their testing?!! Maybe you could giv us your source of where it IS running fine? I have seen heat issues left and right.

In my original post I said this: "These Prescotts are working fine for the typical consumer at their rated speeds."

Falkentyne responded with this false statement: "The prescotts are *NOT* working "Fine" at their rated speeds. Motherboard mosfets are blowing out left and right."

Dippy, now you point to one article where the were overclocking the snot out of a Prescott. No surprise to me they had a meltdown. But, please tell me how this relates to the masses of people running at stock speeds?

Sheesh, if everyone that has a Prescott is frying motherboards (even the Joe Sixpacks that don't overclock) how come nobody can provide a link or the data to prove it? Thousands of Prescotts are going into OEM computers, so you'd think there would be some record of all these dead computers that are supposedly littering the countryside.

There are la whole bunch of people in the Intel CPU section of the forum that have overclocked Prescotts. I had a 2.8E Prescott myself for a few months. Most of the time I had it watercooled and ran it stable up to 3.8 gig, but I also ran it air cooled for a few weeks at 3.4 gig with no problems. The only time you'll find a meltdown is from extreme overclocking using high voltages. Please get your facts straight before making outragiously silly claims.

The 90nm Intel processors are here to stay for a while. They are hardly dead. A new stepping will be released soon and they will keep improving and getting better as they mature.

 

[Sentential]

I think most of us would rather see a 3 GHz Conroe.

Amen to that....thats my next system if all goes well At this point they need to increase memory rates and lower latency. Clock frequency is becoming a thing of the past

 

[xtrmeocr]

Well welll.... looks like AMD is leaking more power than a Prescott. LoL It's official. Now what do you fanbois have to say for yourself?

Socket 939:







http://www.x86-secret.com

i have to say you dont know what your talking about

AMD gives the theoretical maximum power output for the entire line of chips. so while if every single transistor was on it would be 104watts.(an impossible situation basically)

intel gives a normal operating wattage on a per chip basis. so while prescotts normal operating range is 103watts it could very well be MORE then that.

in short. AMD's TDP != Intel's TDP.