64bit Prescott?

Its not enough for me to believe yet, but it sure does make me wonder.

Linkage -

http://www.theinquirer.net/?article=11668

Seems curious . . .

Rumors fly all the time, so there's no reason to believe this one over any other. On the other hand, Intel likely knows a LOT more about Hammer than any of us, so they know the potential threat that is Hammer like the back of their hand (the front, or palm, is much more complex . . .). Thus, if Intel was worried, I guess it's not THAT unreasonable to think that they included the CAPABILITY in Prescott, just in case AMD succeeded with Hammer, as they appear to be doing, based on Intel's P4 with HT EE. On the OTHER hand, it seems like a HUGE HUGE HUGE sum of money to spend "just in case". So it's my opinion that:

1.) Intel didn't do it, and it's just a rumor or
2.) Intel DID do it and will enable it regardless, and never intended NOT to enable it, or
3.) Intel was taking a MONUMENTAL risk in developing a 64-bit processor "just in case". Whether or not it was monumentally stupid, or monumentally smart remains to be seen.

I don't think Intel is stupid. God knows they've got at least 50 people sitting around with calculators telling the HPIC (head people in charge) exactly how much EVERY option will cost them, and what the chances are that it will be good or bad. I'm sure they were the ones that had the final say in the P4 with HT EE . . . Intel wouldn't have spent that much money developing a chip unless they thought they were going to need it. And if they DID think they were going to need it, then we can already guess they were afraid of Hammer. And if they were afraid of Hammer enough to make an EE version of the P4, then there's no reason to think that they weren't afraid enough to make a 64-bit Prescott, though I really don't think Intel developed two concurrent Prescott chips, one 32-bit, one 64-bit. I guess they have the resources, but that sounds a little too spendy for any smart company.

On the other hand, I'm just an 18 year old college student, so what the he** do I know about Intel's business practices or development?

Z

i think its a hoax.. until we intel announce it officially.. even if its in the core.. im sure no one can really use it until intel decides to relaease it to the world

I absolutely agree that this won't be one of those "turn a Duron into an XP" moments. Intel's way to smart to allow that to happen. It's one thing to enable cache or multipliers (which is, by itself, a pretty big deal). It's quite another to give every chip a completely new lease on life with 64-bit for the price of a 32-bit chip. We'll only get to use it when Intel gives say so.

Z

i bleive it 100%. Yamhill anyone?
It takes up less than 5% of the die space on a Hammer to extend the ISA to 64bits.. so that would be less than what, 3% of intels die size... very cheap insurance option.

Just rember all P4s have HT built into the core, yes even Willies. So something like this is not a new idea for intel at all.
If they turn it on or not is a completly different story

so you are saying the northwoods are 64 bit but disabled? man this reminds me of the 486sx with the co processor disabled but still there and possible some of the celerons started out with 256kb cache

Overclocker550 said:

so you are saying the northwoods are 64 bit but disabled? man this reminds me of the 486sx with the co processor disabled but still there and possible some of the celerons started out with 256kb cache

Click to expand...

I am not saying Northwoods are 64Bit enabled or have it built into the core...
But I saying this, they all do have Hypertreading built in, as DO ALL CELERONS based on the P4 core...

Now the PRESCOTT die is very likly to have 64bit built in.
http://www.chip-architect.com/news/2003_04_20_Looking_at_Intels_Prescott_part2.html
the write up is a little old...
but a very good read non the less

and its not like intel has never done something like this before

Reminds me of a game of checkers. Except this go-around, Intel is baked into a corner with an option of only two moves....and no matter which move Intel makes, their gonna get jumped.

If Intel enables 64-bit extensions in Prescott, it will be at the expense of Itanium sales. If so, this just means more room for Opteron in the workstation/server segment- which has the most potential for profitability.



-PC

no one said intel is going to enable 64-bit extentions.
and i doubt they will becuase of Itanium....

ed seems to think so too
http://www.overclockers.com/articles834/

I wouldn't doubt that it's in there, but I don't think it's any more likely that it is than that it isn't. I think Intel likes the expensive insurance. The kind that has no deductable. Of course, if they enable 64-bit, Itanium sales will go down, but what's the alternative? Most of their market is the consumer market, not Itanium. They'd much rather keep money in their pocket. And considering Intel's PR prowess, I'd wager they can reap a LOT of benefits from enabling it. They'll get the whole computer world to upgrade to 64-bit and dump their old platforms. That'll more than make up for Itanium losses.

Z

Itanium is moving Multicore, we should see a 3+ core version withen the next year.. and a 8core version in about 3 years...
Itanium is made for SERVER. and it makes sence.

Putting out a 64bit measly desktop still couldnt compete with a Itanium with 3+ cores.
Also Itanium has efficent code. unlike the dirty unefficent code x86 runs on.

I'd say more than not the workings are in the Prescott. If it works 100% or not, is another question. But i dont doupt its like HT in willies, jut not 100% working, but present.
Why i mention this?
Rember, Intel doesnt think a Hammer is equal to an Itanium.
They think Hammer is equal to an Xeon. Tejas Xeons with 64bit, that might not be too far of a strech seeing how Itaniums by then will have MANY cores