PDA

View Full Version : Next Pentium 5 Tejas info


Pages : [1] 2

-=DVS=-
07-02-03, 12:09 PM
"Looks like its gonna be a monster" :D

90nm/65nm Micron Process
120-140mm2 Die Size (90nm) / 80-100mm2 Die Size (65nm)
Extended Enhanced Hyper-Threading
8 Tejas New Instructions (TNI) : New high-performance audio standard, code-named Azalia supporting Dolby digital, multistreaming and improved speech recognition.
1MB L2 Cache
24KB L1 Cache
16K uOps Trace Cache
800/1066Mhz FSB
775 LGA (Land Grid Array) Package
Clock Speed From 4.4Ghz & Above
Supports Socket T, Dual Channel DDR-II 400/533Mhz, PCI Express x16 & ICH6 with Grantsdale chipsets
Availability H2 2004

Source 1 (http://www.theinquirer.net/?article=10286)
Source 2 (http://www.vr-zone.com/#3099)
http://www.vr-zone.com/Home/news201/Tejas-2-s1.jpg

-=DVS=-
07-02-03, 12:13 PM
"WOW Intel is pushing Megahertz fast :eek:"

10.20GHz Intel Nehalem slated for 2005

Future Desktop Roadmaps Tejas to reach 9.20GHz, Prescott 5.20GHz

The contents of the note appear to reveal future plans for future Intel desktop processors right up until 2005.

By then, according to the note, Intel will be able to deliver 10.20GHz desktop CPUs codenamed "Nehalem" and produced using 65 nanometer technology.

If Intel manages to migrate away from the 90 nanometer technology it will introduce towards the end of this year, by then the "Prescott" core will deliver at least 5.20GHz using the 800MHz system bus.

The immediate successor to Prescott after it tops out at 5.20GHz will be the "Tejas" core, also produced on a 90 nanometer process and delivering 5.60GHz using a 1066MHz system bus. That's slated to start appearing towards the end of 2004.

Tejas will increase in steady increments which appear to be 6GHz, 6.40GHz, 6.80GHz, 7.20GHz, 7.60GHz, 7GHz, 8.40GHz, 8.80GHz and topping out at 9.20GHz.

The first Nehalem is supposed to appear at 9.60GHz before Intel succeeds in its goal to produce a 10GHz+ chip, the Nehalem, and using a 1200MHz front side bus.

Source 1 (http://www.theinquirer.net/?article=7481)

budd_wm
07-02-03, 06:41 PM
Just wait until we start overclocking those things...

B&R
07-02-03, 07:52 PM
Man 10 Ghz..wow. I would rather get em to make an efficient architecture than jus pushing Mhz(or in this case Ghz).

StealthHawk
07-02-03, 08:01 PM
Originally posted by -=DVS=-
"WOW Intel is pushing Megahertz fast :eek:"

10.20GHz Intel Nehalem slated for 2005

Yeah, given Intel's past strategies, I have a hard time believing we will see 10.2GHz in 2005. And if we do, it most likely won't be 3x the speed of today's P4 3.2GHz.

St Lobus
07-02-03, 08:24 PM
That's amazingly fast, but why?

I am a major gamer, and every time I read about 3 or 4 GHz processors, I start drooling over the thoughts of running games like HL2 and Doom 3 on those. But 10 GHz? Will there be hardware that can keep up with it?

But then again, playing any of the major games that come out at the end of this year at resolutions 4x higher than my current monitor can handle with everything maxed...that does sound pretty nice. :)

Of course, there are other applications for computers, but what they are is beyond me. ;)

StealthHawk
07-02-03, 08:32 PM
Originally posted by St Lobus
That's amazingly fast, but why?

I am a major gamer, and every time I read about 3 or 4 GHz processors, I start drooling over the thoughts of running games like HL2 and Doom 3 on those. But 10 GHz? Will there be hardware that can keep up with it?

The latest cards are CPU limited at all but the highest resolutions. Most games are CPU limited.

St Lobus
07-02-03, 10:00 PM
Bah...I was gonna try to say something smart, but I've played WAY too much Warcraft 3 today. I'll fill this space with something intelligent later. :rolleyes:

sebazve
07-02-03, 10:49 PM
well thats a lot of freakin mhz:angel: but let see how good they are..

Nemesis
07-03-03, 06:56 AM
Originally posted by StealthHawk
The latest cards are CPU limited at all but the highest resolutions. Most games are CPU limited.
This could be completely wrong, and if so I apologise, but this review of an Abit IC7-G at [H]ard | OCP showed a Radeon 9700Pro to be the limiting factor in 3DMark2001. If true, this is good, as it shows there is more of a balance in hardware performance under certain configurations

StealthHawk
07-03-03, 07:28 AM
Originally posted by Nemesis
This could be completely wrong, and if so I apologise, but this review of an Abit IC7-G at [H]ard | OCP showed a Radeon 9700Pro to be the limiting factor in 3DMark2001. If true, this is good, as it shows there is more of a balance in hardware performance under certain configurations

I'm not sure how you come to that conclusion, as it is well known that 3dmark2001 is just as CPU limited as it is GPU limited, if not more.

In the review in question, you can see score differences between different CPUs. There is a 1000 point difference between 2.4GHz and 3.0GHz on the same motherboard.

vampireuk
07-03-03, 07:42 AM
A 10ghz Intel cpu would most likely be matched by a 5ghz AMD cpu:D

Nemesis
07-03-03, 08:21 AM
Originally posted by StealthHawk
I'm not sure how you come to that conclusion, as it is well known that 3dmark2001 is just as CPU limited as it is GPU limited, if not more.

In the review in question, you can see score differences between different CPUs. There is a 1000 point difference between 2.4GHz and 3.0GHz on the same motherboard.

I should have posted the url, sorry - it's [H]ere (http://www.hardocp.com/article.html?art=NDY5LDU=) - the 3.0GHz P4 delivered virtually the same score at 3.0 and 3.3GHz. They implied this was because the 9700Pro was the limiting factor. As I said, I could be wrong.

With this kind of setup, wouldn't the GPU be the limiting factor when playing with AA and AF, which most owners of high end cards would be doing?

You are bang on about CPU limitations - my new setup gave me nearly 6000 more in 3dMark2001, on the same 64Mb Ti4200 - 14,300.

GlowStick
07-03-03, 10:52 AM
Are you sure the Tejas will be called Pentium 5? I was understanding that it was still a p4!!?

And i think intel is being coy with their processors, they probly wont release anything faster untill amd forces them too.

Nutty
07-03-03, 10:55 AM
Those specs look crap!

8 Tejas New Instructions (TNI) : New high-performance audio standard, code-named Azalia supporting Dolby digital, multistreaming and improved speech recognition.

Why in gods name would you want to waste cpu power doing dolby digital encoding? Thats what we have nforce2 chipsets for..


24KB L1 Cache

24k?? Sheese.. do they ever realise just how important decent cache size is? How about 1/4meg of L1 cache?

Seems Intel are just megahertz pushers.. I expect AMD's next chip will be half the speed, and still be of similar performance. Why dont they start making smarter chips, instead of just cranking the speed up continually. Seems to work for AMD.

GlowStick
07-03-03, 11:16 AM
Originally posted by Nutty
Those specs look crap!

8 Tejas New Instructions (TNI) : New high-performance audio standard, code-named Azalia supporting Dolby digital, multistreaming and improved speech recognition.

Why in gods name would you want to waste cpu power doing dolby digital encoding? Thats what we have nforce2 chipsets for..


24KB L1 Cache

24k?? Sheese.. do they ever realise just how important decent cache size is? How about 1/4meg of L1 cache?

Seems Intel are just megahertz pushers.. I expect AMD's next chip will be half the speed, and still be of similar performance. Why dont they start making smarter chips, instead of just cranking the speed up continually. Seems to work for AMD.

Well i belvie intel dose it because they play the "lets just release low clocked versions if we have no compatition" but once amd turned the heat on em they had running, and intel went to their backup plan, ship processors at clock speeds at the top end insted of playing lame duck.

budd_wm
07-03-03, 11:47 AM
Originally posted by Nutty
Those specs look crap!

8 Tejas New Instructions (TNI) : New high-performance audio standard, code-named Azalia supporting Dolby digital, multistreaming and improved speech recognition.

Why in gods name would you want to waste cpu power doing dolby digital encoding? Thats what we have nforce2 chipsets for..


24KB L1 Cache

24k?? Sheese.. do they ever realise just how important decent cache size is? How about 1/4meg of L1 cache?

Seems Intel are just megahertz pushers.. I expect AMD's next chip will be half the speed, and still be of similar performance. Why dont they start making smarter chips, instead of just cranking the speed up continually. Seems to work for AMD.

Because the average consumer doesn't look at the L1 cache, he/she looks at the Megahertz.

StealthHawk
07-03-03, 07:31 PM
Originally posted by Nemesis
With this kind of setup, wouldn't the GPU be the limiting factor when playing with AA and AF, which most owners of high end cards would be doing?

Well, yes. FSAA/AF make games much more GPU limited. However, these options are forced on in the driver control panel. Lots of games support AF natively, but most games don't support FSAA in-game. Hence why I say that games are CPU limited. So yes, you can make them GPU limited, just not with in-game settings for the most part :p

reever2
07-03-03, 07:52 PM
Originally posted by -=DVS=-
"WOW Intel is pushing Megahertz fast :eek:"

10.20GHz Intel Nehalem slated for 2005

Future Desktop Roadmaps Tejas to reach 9.20GHz, Prescott 5.20GHz

The contents of the note appear to reveal future plans for future Intel desktop processors right up until 2005.

By then, according to the note, Intel will be able to deliver 10.20GHz desktop CPUs codenamed "Nehalem" and produced using 65 nanometer technology.

If Intel manages to migrate away from the 90 nanometer technology it will introduce towards the end of this year, by then the "Prescott" core will deliver at least 5.20GHz using the 800MHz system bus.

The immediate successor to Prescott after it tops out at 5.20GHz will be the "Tejas" core, also produced on a 90 nanometer process and delivering 5.60GHz using a 1066MHz system bus. That's slated to start appearing towards the end of 2004.

Tejas will increase in steady increments which appear to be 6GHz, 6.40GHz, 6.80GHz, 7.20GHz, 7.60GHz, 7GHz, 8.40GHz, 8.80GHz and topping out at 9.20GHz.

The first Nehalem is supposed to appear at 9.60GHz before Intel succeeds in its goal to produce a 10GHz+ chip, the Nehalem, and using a 1200MHz front side bus.

Source 1 (http://www.theinquirer.net/?article=7481)

Too bad most of those specs are wrong. Intel always does that, releases information on cpu's years ahead of their actual appearance giving unrealistic speeds and expectations for their consumers. Take for example the Prescott, first articles talking about it said it would go to 6 ghz, then the next said 5.5, then the next ~5, and now this says it stops at 4.4 and tejas takes over. Then while they said 5ghz, they said it would reach that in 2005, now it looks like they will reach that even earlier, except the speed is less than 5 ghz. And the 10ghz by 2005 is an absolute joke, clock speeds dont move that fast, and in microprocessor time, the move from 3ghz to 10 in less than 3 years would be too quick and would flood the market with tons of cpus, which is NOT intels strategy, they keep atleast 2 speed bumps on deck in case their competitors have a speed bump, then they release the 1 on deck and get another one ready so they are always ahead of their competition and only release things when it is absolutely necessary to make your competition look bad

Geforce4ti4200
07-03-03, 11:33 PM
Originally posted by StealthHawk
The latest cards are CPU limited at all but the highest resolutions. Most games are CPU limited.


exactly. I sometimes wonder if I wasted $100 on a ti4200 when I could have just kept my ti200 :rolleyes: of course I like 1600x1200 so maybe I can justify a ti4200, also she got me 13k marks, another good thing about her :D


anyway wow exciting new cpus! amd's 5GHz will match Intel's 9GHz LOL

Steppy
07-03-03, 11:56 PM
Originally posted by Nutty
Those specs look crap!

8 Tejas New Instructions (TNI) : New high-performance audio standard, code-named Azalia supporting Dolby digital, multistreaming and improved speech recognition.

Why in gods name would you want to waste cpu power doing dolby digital encoding? Thats what we have nforce2 chipsets for..


24KB L1 Cache

24k?? Sheese.. do they ever realise just how important decent cache size is? How about 1/4meg of L1 cache?

Seems Intel are just megahertz pushers.. I expect AMD's next chip will be half the speed, and still be of similar performance. Why dont they start making smarter chips, instead of just cranking the speed up continually. Seems to work for AMD. Cache size is one of many factors that are equally important. Latency, bit-depth, associativity, are all just as important. You do know that Intel's L1 is WAY faster than the athlon's right? I don't have the numbers in front of me, but last time I checked I believe Intel's small cache delivered something like 5 or 6 times the bandwidth to the CPU that the K7's did.

Steppy
07-04-03, 12:20 AM
Originally posted by reever2
Too bad most of those specs are wrong. Intel always does that, releases information on cpu's years ahead of their actual appearance giving unrealistic speeds and expectations for their consumers. Take for example the Prescott, first articles talking about it said it would go to 6 ghz, then the next said 5.5, then the next ~5, and now this says it stops at 4.4 and tejas takes over. Then while they said 5ghz, they said it would reach that in 2005, now it looks like they will reach that even earlier, except the speed is less than 5 ghz. And the 10ghz by 2005 is an absolute joke, clock speeds dont move that fast, and in microprocessor time, the move from 3ghz to 10 in less than 3 years would be too quick and would flood the market with tons of cpus, which is NOT intels strategy, they keep atleast 2 speed bumps on deck in case their competitors have a speed bump, then they release the 1 on deck and get another one ready so they are always ahead of their competition and only release things when it is absolutely necessary to make your competition look bad How about we look at the time it took to go from 100 Mhz to 1000 vs how long it took to get from 1000 to 2000. You're looking at a 3x increase in clock over 3 years which has happened consistantly in the past(usually much less than 3 years) It's really not all that much different than going from 100 Mhz in 300Mhz. Intel's expected to increase the difference between processors up to 400 Mhz per increment. Intel stated 10Ghz by 2005 back at the P4 release, this would seem to back that up a bit, and it actually fits with the rate they're going at now. It should be roughly 24 months from 2.0-4.0, and 24 months from then would bring you up to about 8.0 with a little bit of time left in 2005. A slightly accelerated release schedule could put them at 10 at some point in 2005.

Steppy
07-04-03, 12:26 AM
Originally posted by GlowStick
Are you sure the Tejas will be called Pentium 5? I was understanding that it was still a p4!!?

And i think intel is being coy with their processors, they probly wont release anything faster untill amd forces them too. There seems to be a lot of confusion as to the difference between being a P4 core based processor and being named the P4. The p4 core(aka the netburst core, aka the P7 core) could go through several processor name changes. It's predecessor the P6 core went through no less than THREE names(the PPro, The P2, and the P3) but they were all based on primarily the same core. If I was a betting man I'd say there's a 15% chance of the P4 name being used for Prescot much less Tejas which I'd wager has less than a 1% of being still called the p4.

As far as AMD dictating what Intel does, Intel has stuck to pretty much the same path for the past 20 years regardless of what AMD does or did, and really only deviated from that at times where the publics perception of AMD could drastically rise(ie the 1.13 P3, and the 1.0 P3).

StealthHawk
07-04-03, 04:49 AM
Originally posted by Steppy
Cache size is one of many factors that are equally important. Latency, bit-depth, associativity, are all just as important. You do know that Intel's L1 is WAY faster than the athlon's right? I don't have the numbers in front of me, but last time I checked I believe Intel's small cache delivered something like 5 or 6 times the bandwidth to the CPU that the K7's did.

I have heard that Intel's L2 cache is much faster and more efficient than AMD's L2 cache. Haven't heard that about the L1 cache though. Haven't heard anything about the L1 cache as far as I can remember, actually. Any linkage would be much appreciated.

GlowStick
07-04-03, 09:56 AM
Originally posted by Steppy
There seems to be a lot of confusion as to the difference between being a P4 core based processor and being named the P4. The p4 core(aka the netburst core, aka the P7 core) could go through several processor name changes. It's predecessor the P6 core went through no less than THREE names(the PPro, The P2, and the P3) but they were all based on primarily the same core. If I was a betting man I'd say there's a 15% chance of the P4 name being used for Prescot much less Tejas which I'd wager has less than a 1% of being still called the p4.

As far as AMD dictating what Intel does, Intel has stuck to pretty much the same path for the past 20 years regardless of what AMD does or did, and really only deviated from that at times where the publics perception of AMD could drastically rise(ie the 1.13 P3, and the 1.0 P3).

Hm, i really feel that strongly that the Tejas will be called a Pentium 4. And im posative that the Prescott will be called P4 because its socket 479 and 'may/should' work in 875P boards, MAYBE 865PE boards. thats what intel claimed to everyone who boght a 875P though heh.