• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Leaked review of Sandy Bridge....

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Any word on the socket 1366 replacement? Specifically 6-core cpus. Want to know if I should upgrade to the 980X or wait a bit. (Currently running 920 at 3.8GHz stable).
 
Anything extra running or even idle will generate heat. That's the main reason I'd prefer any embedded graphics function it to be turned off or non-existent.

A dual socket 1366 hex core system looks pretty good to me right now. By the time I can afford the build there should be some decent benching/use data on Sandy Bridge systems to see if there is any cost/performance advantage over a dual 1366 build.

You can always wait until the next new thing comes out, but then there will always be a still newer thing on the horizon to wait even longer for. :shrug:
 
Since they come already at 3.5ghz area and they are all 32nm, it seems that doing 4.5ghz on air should be easy if those specs hold true. And apart from the IGP, what is so revolutionary about Sandy Bridge? When X58 came out that was a complete overhaul in memory, graphics and Hyper Threading. Seems like AMD's Bull Dozer will be more impressive.
 
On XS, I asked if the 2500K would be better than a i7 860 (considering the 2500K will be right around $200, and right now you can buy i7 860 on forums for $200), and here's one of the responses I got.

http://www.xtremesystems.org/forums/showpost.php?p=4666844&postcount=109

And one of the guys in the thread already bought a retail chip.

http://www.xtremesystems.org/forums/showpost.php?p=4665104&postcount=40

Interesting. One guy saying that the limit is different on every CPU. I know I saw a screenshot of a SB CPU using a 57x multiplier, but his is only 47x max. If the max multipliers are that low, what's the point in benching them?

EDIT: It was a 2600K using the 57x multiplier, whereas in that XS link it's a 2500K.
 
Audioaficionado said:
I don't want any video processors included in any new chip I get. Hopefully that feature is only going to be on the mobile processors.

As far as I know all Sandy Bridge chips will have IGP... so you'll be out of luck on that one. Maybe they'll release some down the road without IGP :shrug:

Might be a while. My uderstanding is Sandy Bridge IGP is DX10.x and 12 EU's, while Ivy Bridge plans are to go to DX11 and 24 EU's.

I've can't find anything pointing to non-IGP chips, desktop or mobile, before Socket R (LGA2011).

BigTerminator said:
... what is so revolutionary about Sandy Bridge?

Might be something in here you haven't seen or read?

*
 
Last edited:
Yes, the difference being the 2600K has HT, and the 2500K is a straight up quad-core. Also the final retail stepping is 7, while most of the ES I saw were 6 (D1). So they may have changed it yet again from the ES to the retail chips. I doubt Intel would change the max multi for each batch.
 
Might be something in here you haven't seen or read?
*
Excellent read :thup:

"A substantially new new microarchitecture
That mystery, it turns out, is pretty juicy, because Sandy Bridge is part of the unprecedented wave of brand-new x86 microprocessor architectures hitting the market. Just weeks after AMD disclosed the outlines of its Bulldozer and Bobcat cores, Intel has offered us an answer in the form of its own substantially new microarchitecture...

...At IDF, we got a better sense of how complete an AVX implementation Sandy Bridge really has, right down to a physical register file to store those 256-bit vectors. This chip should be in a class of its own on this front, at least until AMD's Bulldozer arrives later in 2011. Even then, Bulldozer will have half the peak AVX throughput of Sandy Bridge and may only catch up when programs make use of AMD's fused multiply-add (FMA) instruction—which only Bulldozer will support...

...Furthermore, Intel claims Sandy Bridge should have substantially more headroom for peak Turbo Boost frequencies, although it remains coy about the exact numbers there. One indication of how expansive that headroom may be is a new twist on Turbo Boost aimed at improving system responsiveness during periods of high demand. The concept is that the CPU will recognize when an intensive workload begins and ramp up the clock speed so the user gets "a lot more performance" for a relatively long period—we heard the time frame of 20 seconds thrown around. With this feature, the workload doesn't have to use just one or two threads to qualify for the speed boost; the processor will actually operate above its maximum thermal rating, or TDP, for the duration of the period, so long as its on-die thermal sensors don't indicate a problem..."

SB might be worth my wait after all. :comp:
 
Dont know if you guys have seen it, but someone leaked it:

http://www.hardwareonline.dk/nyheder.aspx?nid=10600

I know most of you dont get the text, but the graphs speak for them selfs.

Apparently Intel is also locking the clock generator, as it is build into the chip, leaving no way of overclokcing the chip, as the multiplier is locked as well. No BCLK clocking.
So with sandy we HAVE to buy the "OC" version of the chip, that is marked with a "k" at the end, to be able to overclock at all. :sly:

Try this: http://translate.google.com/transla...heder.aspx?nid=10600&sl=da&tl=en&hl=&ie=UTF-8

Google Translate isn't exactly perfect, but it is understandable.

x509 :chair:
 
So let me get this straight, these "Sandy Bridge" CPU chips will not fit onto any of the current Intel mobo's?
 
Screw that then, they'd have to be pretty ******* exemplar to warrant a complete motherboard overhaul.
 
Screw that then, they'd have to be pretty ******* exemplar to warrant a complete motherboard overhaul.

Yea, it would seem AMD has an edge in that respect since they've designed there AM3+ boards to accept AM3 chips. But, it's just a matter of time before socket changes are needed and it's always a tough pill to swallow. Personally, I'm not so bothered by the socket change as much as how there are two lines running concurrently (1156 & 1366), which are both being replaced (1155 & 2011(unconfirmed)). Really makes upgrading a hassle.
 
I'd only be interested in the socket 2011 systems. Otherwise I'll go with a dual socket 1366 w/hexcore CPUs.
 
Screw that then, they'd have to be pretty ******* exemplar to warrant a complete motherboard overhaul.

Yea, it would seem AMD has an edge in that respect since they've designed there AM3+ boards to accept AM3 chips. But, it's just a matter of time before socket changes are needed and it's always a tough pill to swallow. Personally, I'm not so bothered by the socket change as much as how there are two lines running concurrently (1156 & 1366), which are both being replaced (1155 & 2011(unconfirmed)). Really makes upgrading a hassle.

... i seem to remember not long ago.... Amd, socket A, 754, then 939 in a VERY short time... then AM2, AM2+ and AM3....

Intel had 775 for what... 5-6 yrs? 1336 and 1156 for 1.5-2 yrs now?

i dont see it as all that bad that intel needs to change sockets... although 1156 to 1155 coulda probably been forgon, i see their reasoning.

Regardless... those graphs make me feel good about my cheap i5 760 build... i had thought bout waiting till sandy bridge came out but just couldnt pass up the i5 760 and mobo i got... ad similar performance clock to clock, at MUCH less in price im fine with that haha.
 
... i seem to remember not long ago.... Amd, socket A, 754, then 939 in a VERY short time... then AM2, AM2+ and AM3....

Intel had 775 for what... 5-6 yrs? 1336 and 1156 for 1.5-2 yrs now?

Let's try again. First off, you've missed 423 and 478.

Pentium IV - 423/478/775 - three sockets, one CPU
Duron/Athlon/Athlon (X/M)P - A - one socket, five CPUs

Core/Core 2 - 775 - one socket, two CPUs
Sempron/Athlon 64 - 754/939 - two sockets, two CPUs

i3/i5/i7 - 1156/1366 - two sockets, three CPUs
Athlon 64/Phenom - AM2(+)/AM3 - technically two sockets, but maybe only counts as 1.5, because AM3 CPU will run in AM2(+) socket, 2 CPUs

Intel total - 6 sockets, 6 CPUs
AMD total - 4 sockets, 8 CPUs

Feel free to correct me, I'm sure I missed something. :cool:

Back on topic, what task other than gaming needs a strong FPU? We have GPUs for those annoying non-integer numbers, and they're not often needed for anything but gaming. Also, even outside of gaming, GPU makers have been doing it better for longer and likely can produce better performance for it than Intel will be able to any time soon (Tesla, FireStream).
 
Let's try again. First off, you've missed 423 and 478.

Pentium IV - 423/478/775 - three sockets, one CPU
Duron/Athlon/Athlon (X/M)P - A - one socket, five CPUs

but the durons, athlons and axp's were pretty much all the same architecture, weren't they?

with the p4's, there was willamets on 423 and 478, northwoods on 478, gallatins on 478, presscots on 478/775, and cedarmills on 775. also, if you are counting durons, you had better count all the celeron revisions as well. ;)

Core/Core 2 - 775 - one socket, two CPUs
Sempron/Athlon 64 - 754/939 - two sockets, two CPUs
the 'core' you speak of doesn't exist. there was just core2duo. unless you mean the pentium m -> core solo...those are mobile cpus.

anyways, both companies have had lots and lots of sockets. both have had long lived sockets and short lived sockets. that doesn't make anyone better than anyone else. :burn:
 
Back