10.2GHz Intel "Nehalem" with a 1200MHz FSB by 2005

Toggle sidebar Toggle sidebar
Random Variable

Random Variable

Lifer
Aug 10, 2001
10,420
2
0
"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."

http://www.theinquirer.net/?article=7481
 
S

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
5.2 GHz on a .09 process?

Doesn't that sound a tad optimistic?

pm? Wingz? Sochan?
 
dullard

dullard

Elite Member
May 21, 2001
25,741
4,264
126
Originally posted by: Sunner
5.2 GHz on a .09 process?

Doesn't that sound a tad optimistic?

pm? Wingz? Sochan?
Click to expand...
I wouldn't say so. There is a tremendous difference between 0.13 and 0.09. Much bigger difference than the jump from 0.18 to 0.13. I know it looks the opposite (0.18 - 0.13 = 0.05, while 0.13 - 0.09 = 0.04). But you have to remember that chips are basically 2D and not 1D. So you have to square the dimensions. Going from 0.18 to 0.13 gives you theoretically 0.18^2 / 0.13^2 => 92% more transistors in the same space. Going from 0.13 to 0.09 gives you theoretically 0.13^2/0.09^2 => 109% more transistors in the same space. Now there are issues that pop up, so you won't ever get a full 109% perfectly functioning transistor gain.

Now lets pretend transistor count scales roughly linearly with speed - to keep things simple. The P4 went from 2.0 GHz on a mature 0.18 process and will probably top out at about 3.4 GHz on its final 0.13 process (a 70% frequency gain). With 109% more transistors we could hope for even more than an 70% frequency gain. But lets pretend that they run into issues that cannot be resolved and just get a 60% frequency gain. 3.4*160% = 5.4 GHz.

Yes these are very rough calculations. But they show that this same chip without any major modifications (other than die size) should reach their claimed 5.2 GHz.
 
B

buleyb

Golden Member
Aug 12, 2002
1,301
0
0
I'll be impressed when I wake up and my computer has made me breakfast...
 
S

Sunner

Elite Member
Oct 9, 1999
11,641
0
76
Originally posted by: dullard
Originally posted by: Sunner
5.2 GHz on a .09 process?

Doesn't that sound a tad optimistic?

pm? Wingz? Sochan?
Click to expand...
I wouldn't say so. There is a tremendous difference between 0.13 and 0.09. Much bigger difference than the jump from 0.18 to 0.13. I know it looks the opposite (0.18 - 0.13 = 0.05, while 0.13 - 0.09 = 0.04). But you have to remember that chips are basically 2D and not 1D. So you have to square the dimensions. Going from 0.18 to 0.13 gives you theoretically 0.18^2 / 0.13^2 => 92% more transistors in the same space. Going from 0.13 to 0.09 gives you theoretically 0.13^2/0.09^2 => 109% more transistors in the same space. Now there are issues that pop up, so you won't ever get a full 109% perfectly functioning transistor gain.

Now lets pretend transistor count scales roughly linearly with speed - to keep things simple. The P4 went from 2.0 GHz on a mature 0.18 process and will probably top out at about 3.4 GHz on its final 0.13 process (a 70% frequency gain). With 109% more transistors we could hope for even more than an 70% frequency gain. But lets pretend that they run into issues that cannot be resolved and just get a 60% frequency gain. 3.4*160% = 5.4 GHz.

Yes these are very rough calculations. But they show that this same chip without any major modifications (other than die size) should reach their claimed 5.2 GHz.
Click to expand...

Yeah true, makes alot more sense with a bit of math thrown in, I was just too lazy to do it myself...of well some sucker always does it anyway ;)

Though one can't help but wonder if Northwood will reach 3.4 GHz, considdering how much power the 3.06 P4 draws, especially with HT enabled.
 
Megatomic

Megatomic

Lifer
Nov 9, 2000
20,127
6
81
Holy cow... :Q

What application is going to be driving that performance increase? I am hoping it is 3D gaming. :)
 
D

Darien

Platinum Member
Feb 27, 2002
2,817
1
0
10GHz...



Makes my TBred XP 1700+ seem slow.



I wonder what the heck the casual user will need all that speed for. But for us gamers, editors, etc...:D
 
B

bgeh

Platinum Member
Nov 16, 2001
2,946
0
0
to satisfy the speed of the FSB, it think we'll need DCDDR-II
or yellowstone
 
dullard

dullard

Elite Member
May 21, 2001
25,741
4,264
126
Originally posted by: Czar
sooooo hows this hammer doing?
Click to expand...
In my opinion (which has been the same for well over a year if you do a search):
1) The initial Hammer (3400+ and slower) will be a flop.
2) Hammer will quickly scale to 4000+ speed and be a great success.
3) Hammer will stall around 4400+ and Intel will come charging back into the lead.
4) Past that and it is too far into the future to make good estimates.
 
Soulkeeper

Soulkeeper

Diamond Member
Nov 23, 2001
6,731
155
106
they could prob hit 10ghz now if they just added another 50 or so more stages.......... hehe
maybe call it webburst or blanketburst or something stupid


i'd just get me a 8ghz amd solution and watch it kill the benchmarks, thank you hehe



sarcasm and speculation intended brace yurself

 
D

DanFungus

Diamond Member
Jul 27, 2001
5,857
0
0
Originally posted by: Darien
10GHz...



Makes my TBred XP 1700+ seem slow.



I wonder what the heck the casual user will need all that speed for. But for us gamers, editors, etc...:D
Click to expand...

not unless it's an XP 17000+ ;)

yeah, I have an 1800+, and already it's less than 1/2 speed of the top speed...
 
L

LordOfAll

Senior member
Nov 24, 1999
838
0
0
I couls see prescott at 5.2, but if tejas really is a 90nm core it won't come close to the 9.2 they predict. Maybe if tejas is a 65nm core it will, but there are other issues besides just shrinking the die when they start to get that small.
 
J

jeffrey

Golden Member
Jun 7, 2000
1,790
0
0
It's not the speed increases that have been impressing me the most from Intel. It's their support of the overall processing enviornment that has been awesome since launch.

FSB - Quad-pumped busses are awesome and Intel has not been shy increasing the speeds
400mhz to 533mhz to 800mhz is what I call progress!

Cache - 256k of level two cache turned into 512k of level two cache and we will see 1mb of level two cache on Prescott!

Hyperthreading - This technology was a good idea and will grow in the future on the desktop, but as far as right NOW the increases in a database/server platform are freaking awesome, Anand showed that in his article.

Memory - Launching with the best bandwith available through Rambus, but then abandonning the standard after market rejects it. Intel is adopting dual-channel PC3200, something that might never have been approved by JDEC if Intel did not want to push fsb and memory speeds as far as they could reasonably go.
 
BFG10K

BFG10K

Lifer
Aug 14, 2000
22,709
3,000
126
10.2 GHz? Just what we need for the likes of UT2003. Bring it on, I can't wait! :)
 
A

Atlantean

Diamond Member
May 2, 2001
5,296
1
0
Thats pretty sweet, that won't be till late 2005 though right? Who cares I will still want one, or whatever equivalent amd processor is out, unless the intel performs better.
 
D

Dug

Diamond Member
Jun 6, 2000
3,469
6
81
I wish Intel spent that much r&d into storage solutions that could keep up with processor speed.

Even if we had 5-10Ghz right now we would still be waiting for that damn hd to spin around.

 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom