Sandy Bridge architecture overview

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org
Am I the only one that is more interested on SB for laptops than desktops? Laptops certainly can use more CPU power, and finally a decent integrated GPU not be to avoided like the plague...whereas desktop CPUs once you OC are pretty much fast enough for most things anyway.

The Zacato preview looked very promising in this arena too.

I am kind of excited that the graphics "floor" is coming up quite a bit with these new CPUs.

I'm hoping Intel also has a lot of room to gain in drivers for graphics. Judging from the stability in the non-blockbuster games they have been obviously prioritizing for driver stability, there are gains to be made.
 
Last edited:

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
The Zacato preview looked very promising in this arena too.

I am kind of excited that the graphics "floor" is coming up quite a bit with these new CPUs.

I'm hoping Intel also has a lot of room to gain in drivers for graphics. Judging from the stability in the non-blockbuster games they have been obviously prioritizing for driver stability, there are gains to be made.

The zaPata scores are fudged . Reader and pys. are proven false and the game demo will also prove to be fudged
 

ilkhan

Golden Member
Jul 21, 2006
1,117
1
0
More graphics info on Sandy Bridge.

-Supports up to 4 displays independent or concurrently
Thats actually a decent upgrade, especially for laptops (where 2 external screens is the max for docked business laptops, can't use the built in with 2 externals. Which sucks, I could have a third screen up right now.)
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Heres what happened Ananda doing the mini review of SB . When the graphics was 2x faster than HD . Which is exactly what 1 eu is suppose to be intels own words . The cork came out of the bottle over at AMD central . Wait until they see performance of Oak Trail in the zapata class of cpus. The ATOM gamer won't run and hide from zapata. Intel has zapped the zapata and AMD knows it.
 
Last edited:

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
LOL. I know two people that will push their limits for forum privileges and you are one of them. :)

Did I make a false statement if so I retract it. But I have everthing right here in front of me or I should say next to me. Just as I said SB was way more than what forum members were saying . Bob has many friends in the tech sector and Bob is my best friend. Did I not clearly state last night , IDF day 1 is over day 2 tomorrow and this years IDF is exciting stuff for us types.

I may not get a chance to tell you this so I tell u right now up front Bob says he has arranged it so I have an 8 core cpu next to me within a month.
 
Last edited:

Riek

Senior member
Dec 16, 2008
409
15
76
so because intel drivers suck the gaming bench was useless? one can keep in mind that all the ULVC from intel have a much lower frequency then the tested parts, or even the desktop parts. (500MHz -> 133MHz).

the only way intel can compete with zacate is either have Atoms with better graphics and lower power, or bring their huge ULVC i series to the bottom with higher graphic performance (probably with lower cpu frequencies as a result).
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,787
136
http://news.mydrivers.com/1/174/174822.htm

The site is saying Sandy Bridge's GPU will unofficially support DX11. So I guess they might be taking it conservatively now. When they first claimed DX10 support with G965 and GM965, they were having problems delivering on the promise. Oh well, so Nemesis might not be a false prophet after all. :)
 

Riek

Senior member
Dec 16, 2008
409
15
76
http://news.mydrivers.com/1/174/174822.htm

The site is saying Sandy Bridge's GPU will unofficially support DX11. So I guess they might be taking it conservatively now. When they first claimed DX10 support with G965 and GM965, they were having problems delivering on the promise. Oh well, so Nemesis might not be a false prophet after all. :)

That would be great. Would be a huge boost for most OEM systems and for battery life of higher end laptops. Maybe 2011 will be a time where we won't need additional gpus in laptop to actually do something with it.
 

piesquared

Golden Member
Oct 16, 2006
1,651
473
136
I don't know what unofficialy support means, but the DX11 spec has to be strictly adhered to. Not partially supported. If the unofficial support is true, I suspect it means that sandy bridge is not fully compliant.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,787
136
No they haven't talked about Socket 2011.

I take it unofficial mean the hardware is there but software isn't. But at the moment its all a rumor, I don't see a tessellator in there.
 
Last edited:

StrangerGuy

Diamond Member
May 9, 2004
8,443
124
106
http://news.mydrivers.com/1/174/174822.htm

The site is saying Sandy Bridge's GPU will unofficially support DX11. So I guess they might be taking it conservatively now. When they first claimed DX10 support with G965 and GM965, they were having problems delivering on the promise. Oh well, so Nemesis might not be a false prophet after all. :)

Lol, DX11 support so it can play the "next DX11 blockbuster with tesslation" for 3 fps.
 

Scali

Banned
Dec 3, 2004
2,495
0
0
When they first claimed DX10 support with G965 and GM965, they were having problems delivering on the promise.

True. I bought a laptop with GM965 about 3 years ago, with Vista installed.
But the drivers only supported DX9 at first, and things like hardware T&L/vertex shading were also disabled mostly.
I was surprised by that, as I thought that originally the part was introduced as a DX10 part.
Many months later, I happened to find a beta driver which finally enabled DX10. It sort of worked, although certain DX SDK samples for DX10 had rendering issues.
As even more months passed, Intel finally launched an official DX10 driver.
But even today, my GM965 still cannot run all the DX10 examples of the DX SDK. Some of the rendering issues have been fixed, but some stuff just crashes still.
It also does DX11 by the way (in DX10 downlevel mode). But neither DX10 nor DX11 perform anywhere near as well as the same code in DX9.

G965 isn't the same as the GM965, and has never received DX10 support. G965 shares some architectural features with GM965, but either some of it was bugged and fixed in GM965, or G965 was never meant as a full DX10 part.

So in short: Intel has had a lot of problems getting DX10 to work, and it still isn't quite 100%. However, it seems that it was pretty much all on the driver side.
It could be that they now have hardware that is capable of DX11, but they again need a lot of time to get the drivers to support it. Or perhaps they'll run into hardware bugs along the way, and have to stick to DX10.1 support. So they may not want to get ahead of themselves here.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
No they haven't talked about Socket 2011.

I take it unofficial mean the hardware is there but software isn't. But at the moment its all a rumor, I don't see a tesslator in there.

No discussion of socket 2011...hmmm

Any discussion of Tukwilla, or Itanium at all?

Everyone likes to jump on top of IBM's back over the whole 45nm HKMG powerpoint release...but what has Intel done with Itanium? Not much different...here's yer slideware, take THAT power7!
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,787
136
They were talking about Sandy Bridge for embedded(Jasper Forest successor?) but nothing else. No, there was no Itanium talk.

The Tukwila systems exist: http://www.supermicro.com/itanium/

But the noteworthy systems come from HP, and they are severely delayed, so I heard.

Bull/Hitachi/HP/Inspur/Supermicro
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
They were talking about Sandy Bridge for embedded(Jasper Forest successor?) but nothing else. No, there was no Itanium talk.

The Tukwila systems exist: http://www.supermicro.com/itanium/

But the noteworthy systems come from HP, and they are severely delayed, so I heard.

Bull/Hitachi/HP/Inspur/Supermicro

http://en.wikipedia.org/wiki/Tukwila_(processor)
It was released on 8 February 2010 as the Itanium 9300 Series

Well I'll be...totally missed this. WTH was I?

Die size is 21.5×32.5 mm or 698.75 mm²

Jesus Christ :eek:

And I see they incorporated Turbo Boost.

OK...nevermind, sorry for derailing your Sandy thread IntelUser, I will go hide my ignorant shame elsewhere for a while :$
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
No kidding. It's not quite full field (that would be 26×33mm) but holy cow that has got to be the largest mass-produced commercial chip on record.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
64
91
I'm perplexed how it is that they did not plan to support DX11 with a leading edge design.

I wonder if it is really a DX11 design but either the driver support is still lacking or the design itself turned out to be borked so they down-graded its functionality via driver limits to just use the DX10.1 function.

For example say they have Tesselators included but it turned out their design of them limited the clocks to 500MHz max or the output itself was not DX11 compliant?

They had 4yrs to get this right after all, and it is being released a full year and change from the first debut of DX11. Why ensure AMD gets the upper-hand in marketing Llano and eOntario with their DX11 support by making SB a mere DX10.1 part?
 

Scali

Banned
Dec 3, 2004
2,495
0
0
Heck, if it does GPGPU well enough to do OpenCL (and also DirectCompute then), they could implement tessellation through a compute shader.
The DXSDK contains some examples of tessellation using DirectCompute. Good fallback for DX10 cards.
 

ydnas7

Member
Jun 13, 2010
160
0
0
http://translate.googleusercontent....&twu=1&usg=ALkJrhhH_d4KLSRUdktdOjCMjCpmsQXjKA

coolaler compares SB to Nehalem (i875K), both at same frequency 3Ghz, both with turbo off. same graphics card GeForce CTX 460
cinebench 11.5 SB CPU score is 12% better than Nehalem
SB OpenGL score is 23% better than Nehalem
SB OpenGL score is 0.3% better than a Magny Cours (with ATI Firepro..)

nb this was with the 6mb cache SB, not the 8mb version
 
Last edited: