• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Sandy E benches - disappointing?

Meh...at least they aren't worse than their previous high-end chips like Bulldozer was. I'm waiting on Ivy anyway.
 
I'm underwhelmed, but I'm not ready to write SB-E off yet based on one set of tests. The NDA lifts in a few hours so we'll have lots of independent verifiable benchies soon.
 
This isn't all that surprising, I can only see SB-E being good for gaming if you need more pci-e lanes for more graphics cards. Most games are gpu limited.
 
I wonder if its more that the 2600 is already capable enough of feeding the GPU fast enough to let it stretch its legs fully. Theres gotta be a point where youre not gonna get any more frames out of a video card no matter how fast the CPU is right?

I have a feeling that even though SB-E isnt any better than SB for games, that it will likely shred SB in productivity and stuff like video encoding and all that.
 
Why does everyone think E means new architecture or something? Its not. Its the same exact cores, with 2 more + more cache and some other enhancements. Other than that the gaming performance should be same exact IPC etc... as socket 1155. This system is meant for 3 or more GPUs and multithreaded applications.

I wouldnt get my panties in a wad over gaming performance. It has been way overkill on hype on bulldozer and look what happened. Lets not make the same mistake with this processor.
 
yeah, nothing surprising about gaming performance. The mobos will support more pci-e lanes or some bologna right? Better performance for quad sli?
 
I'm pretty excited about it. Looking to put my 1366 out to pasture and get this newest bad boy. The 3930K looks to be the one to get price/performance wise.
 
These results look great. A base Cinebench score of 10.4 makes for a really good jump when OC'd. It would take a 2600K running 5ghz to get a 10. Hopefully I can get my hands on a 3930 sometime and give it a moderate OC (~4.4) for a 24/7 setup.
 
I wonder if its more that the 2600 is already capable enough of feeding the GPU fast enough to let it stretch its legs fully. Theres gotta be a point where youre not gonna get any more frames out of a video card no matter how fast the CPU is right?

I have a feeling that even though SB-E isnt any better than SB for games, that it will likely shred SB in productivity and stuff like video encoding and all that.

Pretty much this. SB-E is shaping up to be a big "meh" - which is funny considering how many people were decrying the initial Sandy Bridges, saying everyone should wait for SB-E instead.
 
Pretty much this. SB-E is shaping up to be a big "meh" - which is funny considering how many people were decrying the initial Sandy Bridges, saying everyone should wait for SB-E instead.
3960x is 3.3 GHz, has 6 cores. 2600k is 3.4 GHz, has 4 cores. Why is anyone the least bit surprised that the 3960x is narrowly beaten out by the 2600k in single-threaded gaming tasks?

The people saying "wait for SB-E" probably already had quad-cores with hyperthreading and didn't see the benefit in a side-grade to a new mainstreak quad-core chip. Now that SB-E is here, those that waited are ready to move and those that moved are ready to wait - again, why should any of this surprise anyone?

The only disappointing thing for me are the power consumption numbers. Those two extra cores appear to add 30W of idle power and a whole lot of full load consumption. I had been hoping that SB-E would be the best of both worlds, power consumption and performance.
 
I'm now using my PC without my 5970 installed, running caseless until my replacement case arrives.
Using the HDMI hook up on the board and this feature alone is so convenient, I'll pass on SB-E.
As for gaming, those results are expected. Most games still run on two cores, very few use 4+ cores efficiently for gaming.
 
People could read the writing on the gaming wall for months now:

1. Socket 1155 + NF200 = the perfect triple-card gaming platform. The extra PCIe lanes of 1366 are redundant...and the same can be said of 2011.

2. Sandy Bridge quad-core is already awash in memory bandwidth at DDR3 1600 dual-channel. In fact, even a heavily-overclocked 4.7 GHz Core i7 2500K sees less than %5 performance improvement between 1600 CL7 and 2133 CL9. Hell, Gulftown 6-core processors don't see more than %10 benefit in desktop usage moving from 2 to 3 channels of cheap DDR3 1333 9-9-9 memory. You can just imagine how little impact that third and fourth memory stick is having on SB-E, even with 6 cores.

3. In every single game four Sandy cores are the same speed or faster than 6 Gulftown cores.

SB-E serves no purpose except for people who really use their computers as massive workstations. Only the people locked in a state of denial were expecting something more impressive.
 
Last edited:
http://www.techpowerup.com/155041/S...Leaked-Disappointing-Gaming-Performance-.html

Gaming is right in line if not similar to current sandy, but I'm sure it has it's advantages as they say in encoding.

Seeing as though a Core i3 keeps up with a Core i7 in Battlefield 3, its not surprising at all. SB-E in my eye is more of a workstation board (quad channel memory), I would like to see what kind of ppd you'll get out of a 3930k while folding.

On another note I don't see the 40 lanes total meaning much to be honest. I know [H]ardOCP did a write up while using a GTX480 in SLI 8x/8x vs 16x/16x.

defaultluser already nailed it.
 
People could read the writing on the gaming wall for months now:

1. Socket 1155 + NF200 = the perfect triple-card gaming platform. The extra PCIe lanes of 1366 are redundant...and the same can be said of 2011.

LOL You are just like the others in stating facts before the chip was even released or tested. Just sayin'

And who is getting 2011 that doesnt run quad cards....? A lot and they will be disappointed however I am running quad gpus. So I will be happy. And its future proofing.
 
LOL You are just like the others in stating facts before the chip was even released or tested. Just sayin'

It's true I'm jumping the gun on "real" performance ests of the new platform, but through analysis you can gauge expected performance.

If more PCIe lanes makes no performance difference on a set of gaming tests, then adding more lanes is not going to make anything faster.

The test I linked compares a 4 GHz Core i7 920 to a 4 GHz Core i7 2600K, so the tests are not CPU-limited. Even at those extreme speeds bandwidth was not a problem for the LGA 1155 triple-GPU platform with NF200: it slightly outperformed the X58 with NF200 overall, which was what you would expect from the slightly higher IPC of Sandy Bridge.

And who is getting 2011 that doesnt run quad cards....? A lot and they will be disappointed however I am running quad gpus. So I will be happy. And its future proofing.

I never said anything about 2011 being bad for people who run quad-GPU setups. I said it was a bad choice for triple-card *or less* gaming, which covers roughly %99 of our forum members.

Congratulations on being the %1.
 
Last edited:
I never said anything about 2011 being bad for people who run quad-GPU setups. I said it was a bad choice for triple-card *or less* gaming, which covers roughly %99 of our forum members.

Congratulations on being the %1.

More than 1% bud. And I gotcha now.
 
It's true I'm jumping the gun on "real" performance ests of the new platform, but through analysis you can gauge expected performance.

If more PCIe lanes makes no performance difference on a set of gaming tests, then adding more lanes is not going to make anything faster.

The test I linked compares a 4 GHz Core i7 920 to a 4 GHz Core i7 2600K, so the tests are not CPU-limited. Even at those extreme speeds bandwidth was not a problem for the LGA 1155 triple-GPU platform with NF200: it slightly outperformed the X58 with NF200 overall, which was what you would expect from the slightly higher IPC of Sandy Bridge.



I never said anything about 2011 being bad for people who run quad-GPU setups. I said it was a bad choice for triple-card *or less* gaming, which covers roughly %99 of our forum members.

Congratulations on being the %1.
The problem is your thinking of benefits right now. Were all thinking of benefits later on.
 
The problem is your thinking of benefits right now. Were all thinking of benefits later on.



Yeah, I think most people on the [H] that are considering it (and are knowledgeable about it, not buying it cause they have lots of money only) are doing so for using multiple cards. How will it work (vs. 16 lanes) when the GTX 680 is out and you are using three cards?

and I know little about this but I don't think we will see very similar performance between the NF200 and native 40 lanes. We'll see soon though.
 
Yeah, I think most people on the [H] that are considering it (and are knowledgeable about it, not buying it cause they have lots of money only) are doing so for using multiple cards. How will it work (vs. 16 lanes) when the GTX 680 is out and you are using three cards?

and I know little about this but I don't think we will see very similar performance between the NF200 and native 40 lanes. We'll see soon though.

Exactly. Ive also seen lots of people saying that PCI-E v3 is worthless. Going by their logic there is no reason to put PCI-E v3 on board and should just stick to v2. Guess what. The more boards that have v3 the sooner the cards will become v3. People running tri-sli and quad sli will benifit. I plan to runa tri-sli or quad sli setup along with a dedicated physx and a PCI-E sound card.
 
What time is the NDA suppose to be lifting? Are we going to see something tonight around 12:00 PST or not till in the morning?
 
Last edited:
What time is the NDA suppose to be lifting? Are we going to see something tonight around 12:00 PST or not till in the morning?

Nothing for sure yet.

20111112122345.jpg


20111112122420.jpg
 
Kyle did a good write up on the utterly useless 60x. Good one too.

I'm tired of staying awake to see these go on sale. I'll get it later this week sometime when Microcenter drops price on k.

Edit** Reviews are starting to be released left and right using google to search for them. So many sites are saying to hold off on this one for a bit longer. I am pausing for now. To read more reviews and product availability etc....
 
Last edited:
Being disappointing at SB-E gaming performance...

That's pretty much like being disappointing with your new 1500W PSU not giving you additional FPS in Crysis 2 - you have every right to be, but it just shows how clueless you are.
 
http://www.techpowerup.com/155041/S...Leaked-Disappointing-Gaming-Performance-.html

Gaming is right in line if not similar to current sandy, but I'm sure it has it's advantages as they say in encoding.

What's suprising in it?
The same logic and reasons why Bulldozer sucks in gaming apply here - lack of above 4 threaded soft.

Difference is that you can have a machine which rocks multithreaded tasks without compromising your single threaded performance (apart from a bit smaller clocks when overclocking)
 
Why aren't these on newegg yet? I got a buddy that is going to buy one because he needs the 8 ram slots for a pc running vms.
 
Just read a whack of different reviews on various sites. A little bit faster here and there and so what? Pricetag is not justified by 99.99% of all computer users. I'm happily hanging onto my 2600K until IB-E. I really do want those 8 RAM slots!!!!
 
The only disappointing thing for me are the power consumption numbers. Those two extra cores appear to add 30W of idle power and a whole lot of full load consumption. I had been hoping that SB-E would be the best of both worlds, power consumption and performance.

The board has a 30W or so IOH chip. That adds a hell lot of power full-time, even when idle. Furthermore you have four more DIMMs, another memory controller and two extra cores. The power usage is entirely reasonable compared to 1155 - it's like 1366 vs 1156.


Exactly. Ive also seen lots of people saying that PCI-E v3 is worthless. Going by their logic there is no reason to put PCI-E v3 on board and should just stick to v2. Guess what. The more boards that have v3 the sooner the cards will become v3. People running tri-sli and quad sli will benifit. I plan to runa tri-sli or quad sli setup along with a dedicated physx and a PCI-E sound card.

SLI will benefit exactly zero. GPUs have no remote chance of saturating the 4GB/s a x8 PCIe 2.0 port provides; they barely saturate 2GB/s on x4, with high end GPUs. PCIe 3.0 doubles that, so it's even more ridiculous to waste 16 lanes on a card that needs worst-ever-case 8 (2.0, high end or dual GPU), 4 (2.0, mainstream GPU or 3.0, high end or dual GPU), or hell, two lanes (PCIe 3.0 mainstream cards).

Socker 2011 is meant for high-end workstations and mainstream servers - it replaces 1366. There's a few desktop processors just like 1366 because Intel wants to milk some money out of rich enthusiasts. The point of the platform is the Xeon chips, capable of quad-processor operation.


Just read a whack of different reviews on various sites. A little bit faster here and there and so what?

Cause those reviews are GPU limited. Even if they weren't, games barely scale above two-four threads, let alone TWELVE. This thing is meant for productivity, not games. Maybe game servers - several fully loaded ones per processor that is.
 
(snip) Socker 2011 is meant for high-end workstations and mainstream servers - it replaces 1366. There's a few desktop processors just like 1366 because Intel wants to milk some money out of rich enthusiasts. The point of the platform is the Xeon chips, capable of quad-processor operation.

Cause those reviews are GPU limited. Even if they weren't, games barely scale above two-four threads, let alone TWELVE. This thing is meant for productivity, not games. Maybe game servers - several fully loaded ones per processor that is.

I am not a gamer at all and what I use my PC for is to make money doing a variety of online work including some superheavy Photoshop. I don't encode vids, or run folding, etc. Strictly business. I want to throw 8 fully loaded RAM sticks at my Adobe CS5 really badly, but I don't want to toss away my 2 month old 2600K rig. I have to admit that I am tempted by a dual socket 8-core SB-E Xeon system with 16 DIMMs but I haven't seen any pricing and I'm afraid that it's going to cost as much as a small car. So I think that the best policy for me (and me alone) is to hang on for IB-E. What I won't do for Photoshop is to throw a big GPU at it as I am thrilled with my fanless 6770 due to the fact that virtual silence is just as important to me as getting through a filter and resize 0.7 seconds faster.
 
Pretty much what I expected, 50% more cores and performance than a 2600k. Though I wasn't expecting the idle and load TDP to jump quite so much.

It'd be a lot more interesting if it had 8 cores with the same price and TDP.

I want 8+ DIMMs, so it's still very much on my radar. I'll probably end up waiting for the dual socket LGA 2011 boards. I'll likely start with 2, 4 core e5 Xeons, and later upgrade to 2, 6 or 8 core e5 Xeons, depending on price.
 
just a minor letdown.

Not really a let down. For $599.00 where else you gonna get 4 cores of exact same performance as 2600K + 2 more cores when you need them with 8 RAM slots and a ton more expandability with GPU's?

No where for now.

People are making me LOL for real @ how disgusted they are with nothing at all. There is no substance to be disgusted with.

Kyle was harping on the fact that Intel has sincerely overpriced the 60X. Kyle made no mention of the K series being overkill or unwanted. He just said that the 60X was utterly useless and unwanted by almost anyone due to the prohibitive price tag and lack of features to justify the price.,

Gaming wise its still a freaking Sandy Bridge processor. Dont fret over something you are not going to get anways people. Let us that are going to spend the dough for something we need the 2 cores and 8 slots for get it and be happy.

Thanks for the Review of the 60X OCP well be waiting on the K series when you can do it.
 
SB-E is about those two extra cores and two extra memory channels. I think it might have had a chance at enthusiast popularity with two extra cores, but no extra memory channels (reduction of motherboard size) and not $1000.

For instance, a 2500K with 6 cores and 9MB cache for $330 (which is in line with 4 core 6 MB for $220). Now that would probably sell like hotcakes!

The SB-E power numbers seem too high for the additional hw it brings to the table.
 
Pretty much what I expected, 50% more cores and performance than a 2600k. Though I wasn't expecting the idle and load TDP to jump quite so much.

I was.

There was a roughly 20w power consumption difference between X58 and P55 at idle and load. For X79, the increase in I/O is to blame for the higher power consumption at idle. It really should not be the extra two cores: Sandy Bridge is fully clock-gated, so the extra cores can be completely deactivated when idle.

I/O is a power hog. Know your needs before you pull the trigger on X79.
 
The board has a 30W or so IOH chip. That adds a hell lot of power full-time, even when idle. Furthermore you have four more DIMMs, another memory controller and two extra cores. The power usage is entirely reasonable compared to 1155 - it's like 1366 vs 1156.




SLI will benefit exactly zero. GPUs have no remote chance of saturating the 4GB/s a x8 PCIe 2.0 port provides; they barely saturate 2GB/s on x4, with high end GPUs. PCIe 3.0 doubles that, so it's even more ridiculous to waste 16 lanes on a card that needs worst-ever-case 8 (2.0, high end or dual GPU), 4 (2.0, mainstream GPU or 3.0, high end or dual GPU), or hell, two lanes (PCIe 3.0 mainstream cards).

Socker 2011 is meant for high-end workstations and mainstream servers - it replaces 1366. There's a few desktop processors just like 1366 because Intel wants to milk some money out of rich enthusiasts. The point of the platform is the Xeon chips, capable of quad-processor operation.




Cause those reviews are GPU limited. Even if they weren't, games barely scale above two-four threads, let alone TWELVE. This thing is meant for productivity, not games. Maybe game servers - several fully loaded ones per processor that is.
So SLI will benefit exactly Zero from X79?
 
I still feel like the achilles heel of the platform is the lack of a truly unlocked quad-core part.
 
So SLI will benefit exactly Zero from X79?

Current SLI/crossfire will benefit extremely close to zero from X79, if there are no CPU bottlenecks. Next gen will also probably benefit close to zero. It'll take at least another generation (speculation here) before graphics cards are able to consistently max out 8x PCI-E 2.0. And by then we'll have a new set of platforms. Although, it is a good chance to mature PCI-E 3.0 before the hardware really needs it.

Edit: I really don't get what people were expecting with SB-E. It's the exact same architecture as used in the 2600k, so single-threaded and 4-threaded tasks would perform similarly at the same clock speeds.
 
Back
Top