3.06Ghz Pentium 4 to support HYPERTHREADING!!!

[SSXeon5]

http://www.2fastcpu.com/

Prescott 3.20GHz Appears in Intel Roadmaps Posted Tuesday, July 23, 2002 by bloodbot

"THE LATEST INTEL ROADMAPS show the Prescott 90 nanometer processor appearing in the second half of 2003. Prescott will be introduced at 3.20GHz in the second half of next year."

"And Intel will introduce hyperthreading in desktop CPUs at speeds of 3.06GHz and bove. While Intel will also introduce a 3.20GHz Pentium 4 in Q2 of next year."

Looken good already! Granite bay is rumored to show its face end of october with Duel channel DDR266 AGP 8x and possibly ATA150 (SATA)! Look out amd

SSXeon

 

[JeremiahTheGreat]

i'm just wondering, is hyperthreading one of those things you have to code for? Cause if it is, its gonna suck

And wasn't the Athlon (clawhammer) going to Debut @ 3400+ speeds?

Meh, i still want one of those P4's anyway

 

[CrazySaint]

If you don't have to code for hyperthreading (I can't remember), it looks pretty sweet

 

[jiffylube1024]

I thought all P4's (or certainly all after Northwood) have hyperthreading built in, but Intel disabled it because it didn't show that great of a performance increase in most desktop uses. The Xeon P4's I believe have hyperthreading enabled by default because they're used in server environments, which profit from the use of hyperthreading.

 

[Acanthus]

all programs that support multithreading should see a performance increase with hyperthreading. Also hyperthreading should increase multitasking performance. Yes it will have to be coded for to increase performance in single appllications. No it wont have to be coded to be utilized in multitasking.

 

[imgod2u]

That seems amazinly unrealistic. You're gonna tell me that between the release of a 3.0 GHz P4 at the end of this year and the 3.2 GHz Prescott in the SECOND half of next year, Intel will release 1, maybe 2 more chips? A 200MHz jump in half a year? Is Intel going to go back to the 66MHz per model jump again? I mean, I know MHz isn't everything but it's what sells and you'd expect anyone but Intel to not ramp up their chips more than just 200MHz in half a year.

 

[Duvie]

Well who is to say they don't go over 3.2ghz with the p4b chip!!!

Example...

willamette P4 went up to 2.0ghz....

NOrthwood came in and released a 1.6ghz chip....went up to and 2.4ghz 400fsb model....

p4b came in with a 2.26ghz p4b...


There is a little overlapping...


This is all driven by AMD's position...If intel thinks hammer is going to come in and spank current p4, everything could be drastically acelerated....

Who knows the adding of hyperthreading, dual channel ddr, and 666fsb could be quite the performance boost when combined to be equivalent to quite a high p4a or b chip....Maybe they will PR rate their chips!!!

I wonder how they can benchmark hyperthreading capabilities when it may help more in multitasking then in programs that may not be multithreaded capable?? Another thread I guess...

 

[DeepBlueJH]

Intel is holding back on AMD waiting to see what the Hammer does... the better Hammer is, the better Intel will make the P4...

Afterall, the only reason Intel is releasing a 2.8GHz P4 in a week is b/c of "slow sales"... meaning theyve had it ready for a while but have been holding back to increase profits on their 2.53GHz model...

Intel is sandbagging

 

[Circlenaut]

Meh, i still want one of those P4's anyway

JeremiahTheGreat, from whom did you learn "meh". You are the first person besides me thats ever said "meh" around me.:Q
I thought I came up with it.

. Oh well.

 

[SolrFlare24]

Well a couple of things here....first off, Intel's explaination as to why Hyperthreading has been turned off is they want to make sure that everything remains fully compatible once its turned on first. In other words, software won't break with it turned on. Otherwise, yeah anything that supports multithreading will do hyperthreading and conceivably I could see MS releasing something in a service pack, or software form intel enabling Hyperthreading on the OS level as well.

As for the slower increases in speed again...this is probably due to

A) The Northwood around those speeds getting closer to its limit so Intel is milking what they can out of it

B) The really really cool news thats been starting to spread like wildfire. That Intel, in an attempt to completely nail AMD while they are slightly down right now, may start not only selling, but including as a major marketing feature, Unlocked! P4 processors. God Bless Intel if they do that. Imagine what you could do with good RAM, an an Unlocked 3ghz+ Northwood.

 

[imgod2u]

But the top of the line always has a higher clock than the previous model. Case and point, Northwood was first introduced with 2.2 and 2.0 GHz models. The 1.8a and 1.6a came later. The Northwood-B's came in 2.53, 2.4b and 2.26b's. Clockspeed is a very important thing for Intel. I'd be surprised if Prescott was released so late next year and yet would just be at 3.2 GHz initially.

 

[Xtasy]

New type of P4 = New need for mobo! We were lucky that the northwood didn't need a new stepping code in the chipset.

 

[Kell]

Originally posted by: imgod2u
But the top of the line always has a higher clock than the previous model. Case and point, Northwood was first introduced with 2.2 and 2.0 GHz models. The 1.8a and 1.6a came later. The Northwood-B's came in 2.53, 2.4b and 2.26b's. Clockspeed is a very important thing for Intel. I'd be surprised if Prescott was released so late next year and yet would just be at 3.2 GHz initially.

This seems odd to me too. Intel's going to release the 3.06GHz P4 in Q4, then is just going to sit on that for what, 7-8 months? It already sounded odd enough when the 3.2GHz was "unofficially scheduled" for Q203, and now this?

Stranger still that after pushing forward the release schedule for Northwoods, Intel now pushes back the release schedule for Prescotts (although unofficially).

 

[Rainsford]

Hmm, I don't know about this. I agree with the general thoughts around here, why would Intel wait so long between frequency jumps? Unless AMD does the same thing or really drops the ball before the end of this year (and I don't think companies like Intel bet on that), won't AMD blow past Intel until Intel get's Prescott out the door?

 

[Duvie]

won't AMD blow past Intel until Intel get's Prescott out the door?

With what??? I know many of you guys are counting on the hammer being the next big thing since sliced bread, but I still don't think this is really going to lay the smack down...

AMD states 20-30percent better then its xp counterpart....I also keep hearing 1800mhz as possible starting speed...Even at 20-30 percent I don't think it will be that far ahead of the 3.06ghz which should be hitting street at same time. If they enable the hyperthreading maybe that can have some improvement (at least in multitasking which may not be seen in benchmarks but for many may be a heavy selling point) as well as the introduction of dual ddr mobos. Also anytime a company states percentage you can count on those never being reach...haven't seen many reach their hype....

Also you need to re-review the numbers on toms hardwares last amd 2200+ review...According to all but 3 test where the 2200+ actually bested the 2.53ghz p4 the 1.8ghz xp usually is behind an average 16%....Look at the tested 2.8ghz and though it is fsb boosted a bit it still wont be as fast as the 3.06ghz p4 to be released at years end...p4 2.8ghz is and an average 27% ahead....3.06ghz p4 could be 34-36% ahead of 1800mhz xp by then....

So hopefully the hammer will debut higher then 1800mhz, meet amds 20-30 percent improvement, and get here on time....I think it only tightens the race, but doesn't take a commanding lead...

 

[Kell]

Originally posted by: Duvie

won't AMD blow past Intel until Intel get's Prescott out the door?

With what??? I know many of you guys are counting on the hammer being the next big thing since sliced bread, but I still don't think this is really going to lay the smack down...

AMD states 20-30percent better then its xp counterpart....I also keep hearing 1800mhz as possible starting speed...Even at 20-30 percent I don't think it will be that far ahead of the 3.06ghz which should be hitting street at same time. If they enable the hyperthreading maybe that can have some improvement (at least in multitasking which may not be seen in benchmarks but for many may be a heavy selling point) as well as the introduction of dual ddr mobos. Also anytime a company states percentage you can count on those never being reach...haven't seen many reach their hype....

Also you need to re-review the numbers on toms hardwares last amd 2200+ review...According to all but 3 test where the 2200+ actually bested the 2.53ghz p4 the 1.8ghz xp usually is behind an average 16%....Look at the tested 2.8ghz and though it is fsb boosted a bit it still wont be as fast as the 3.06ghz p4 to be released at years end...p4 2.8ghz is and an average 27% ahead....3.06ghz p4 could be 34-36% ahead of 1800mhz xp by then....

So hopefully the hammer will debut higher then 1800mhz, meet amds 20-30 percent improvement, and get here on time....I think it only tightens the race, but doesn't take a commanding lead...

No, of course it won't take a commanding lead at the start. In fact, I'm already fairly certain it won't debut at M3400+, but something more like M3000+.

Even with such a "low" PR, it's reasonable to expect it to take a slight lead over the Northwood--then we'll be back to the Athlon-vs-Willamette days, when AMD was really only slightly in the lead, but became so popular by being cheaper as well as faster. Then, if AMD can get to M3400+ by 2Q03, it'll be a while before even the Prescott can defeat it--unless Intel can ramp 90nm Prescotts way, way ahead of schedule. This strange twist in the release schedules makes me wonder if Prescott will even meet the current release deadline, much less arrive early.

And while a lot of people (primarily Intel fans) might jeer at the Athlon's PR ratings, the fact remains that the PR ratings are surprisingly accurate--even conservative. If AMD gives the Hammer an M4000+ rating, we can reasonably expect it to at least slightly edge out a 4GHz P4.

 

[dexvx]

Why is SSXeon is interested in Prescott? Is that your name or something? You've posted a lot of "news" about Prescott and about 90% of them happened to be wrong.

 

[Kell]

Originally posted by: dexvx
You've posted a lot of "news" about Prescott and about 90% of them happened to be wrong.

Now that's rather harsh. Sure he's posted a lot of Prescott speculation, but most of it we don't actually know if it's right or wrong. Kinda the same with Hammer.

 

[Rainsford]

Originally posted by: Duvie

won't AMD blow past Intel until Intel get's Prescott out the door?

With what??? I know many of you guys are counting on the hammer being the next big thing since sliced bread, but I still don't think this is really going to lay the smack down...

AMD states 20-30percent better then its xp counterpart....I also keep hearing 1800mhz as possible starting speed...Even at 20-30 percent I don't think it will be that far ahead of the 3.06ghz which should be hitting street at same time. If they enable the hyperthreading maybe that can have some improvement (at least in multitasking which may not be seen in benchmarks but for many may be a heavy selling point) as well as the introduction of dual ddr mobos. Also anytime a company states percentage you can count on those never being reach...haven't seen many reach their hype....

Also you need to re-review the numbers on toms hardwares last amd 2200+ review...According to all but 3 test where the 2200+ actually bested the 2.53ghz p4 the 1.8ghz xp usually is behind an average 16%....Look at the tested 2.8ghz and though it is fsb boosted a bit it still wont be as fast as the 3.06ghz p4 to be released at years end...p4 2.8ghz is and an average 27% ahead....3.06ghz p4 could be 34-36% ahead of 1800mhz xp by then....

So hopefully the hammer will debut higher then 1800mhz, meet amds 20-30 percent improvement, and get here on time....I think it only tightens the race, but doesn't take a commanding lead...

What I meant was that if Intel releases the 3.06 ghz and then sits around for 6 months until their next chip release (3.20 ghz), during that time it seems like AMD would keep improving their chips and eventually pass Intel. You can't just sit around and not expect that to happen, even if you are Intel

 

[Duvie]

I agree.....I think the timetable is a liitle speculative, anyways...I think they will show their hand when it comes time to address the hammer and how well it does...

Remember the 3.2ghz will be a prescott going to 666fsb, which should be a good boost in itself when coupled with say dual channel pc2700ddr...the p4 666fsb will produce 5.4gb of bandwidth and dual pc2700ddr should fit nicely with that....


Also if intel decides to integrate larger l2 cache as well could be a boost...The changing to the p4c chip may be a large jump due to the enhancements alone....


It is all specualtive and we can talk until we are blue in the face...We just have to wait and see!!!!

 

[dexvx]

I think you guys are taking clockspeed too seriously. Willamette to Northwood got about a 5-20% boost in performance across the board at the same clock. now you move to 533Mhz FSB, thats another 5-15% boost. Then you move to PC1066 RDram, thats another 5-15% boost. Those things stack up. You guys keep saying that Mhz doesnt matter and point to AMD. But it also applies to Intel (McKinley).

People make hammer speculations, but at least they dont see a picture of a computer and claim it to be a Prescott box (one of his threads).

 

[sonoran]

Originally posted by: Duvie


Also you need to re-review the numbers on toms hardwares last amd 2200+ review...According to all but 3 test where the 2200+ actually bested the 2.53ghz p4 the 1.8ghz xp usually is behind an average 16%....Look at the tested 2.8ghz and though it is fsb boosted a bit it still wont be as fast as the 3.06ghz p4 to be released at years end...p4 2.8ghz is and an average 27% ahead....3.06ghz p4 could be 34-36% ahead of 1800mhz xp by then....

Duvie, the one thing that's going to throw a monkey wrench into your numbers is hyperthreading. You might want to take a look at Paul Otellini's keynote address from WinHEC 2002. One particular quite of interest from the guy who did the demo (using a 3GHz processor):

What you saw is in a relatively simple application, the first video encoding thing, about a 20 percent performance improvement. And that's significant. It's sort of two-to-three bin splits, so in the current parlance, sort of 500 more megahertz capability being brought to the machine.

Like an extra 500MHz of performance. If you think the benchmark wars are ugly now, just wait until hyperthreading is thrown into the fray. Maybe we'll have to start benchmarking two copies of an app at once to get meaningful results? Should be great fun.

BTW, there's a pretty cool intro to hyperthreading here for anyone who's interested in what it's all about.

 

[SSXeon5]

Originally posted by: dexvx
Why is SSXeon is interested in Prescott? Is that your name or something? You've posted a lot of "news" about Prescott and about 90% of them happened to be wrong.

Way to be an ass

And I do have facts that they were running the prescott @ 4ghz if you want me to bring the thread back up...

SSXeon

 

[dexvx]

Just like you made the thread that said that was a picture of a Prescott when it was in fact a Tidewater (or a new form factor P4, i forget)?

 

[Duvie]

sonoran...thanks for the link...

I am ignorant about hyperthreading and its effects...I was being conservative with the numbers only factoring in the features as they stand today in p4b northwoods...

I will read though...