Sandy Bridge-E Details Revealed

[LOL_Wut_Axel]

1-2%

That's pretty much what the 100MHz difference will do.

 

[bridito]

Aside for the ground breaking overclocking crowd it seems that the i7-3930 is going to be the "sweet spot" for really serious powerusers. If Intel can bring it in at under $600 it seems that I might have to abandon my plans for the i7-3820... of course unless BD can pull a rabbit the size of Godzilla out of its hat before then.

 

[AdamK47]

And when was I arguing over the extra features on the SB-E platform? I already mentioned it in the past: the real reason why you should buy SB-E is if you need the more feature-rich platform and/or the extra cores. Anything apart from that and there's simply no reason.

No, memory bandwidth is not an issue. When comparing Lynnfield to Bloomfield it was an academic difference in programs (under 1% performance increase), and given the fact that Sandy Bridge has even more memory bandwidth than Lynnfield I expect this fact to be accentuated.

Extra cache size, again, will make little to no difference. This is the case going from the 2500K's 6MB L3 cache to the 2600K's 8MB. Just because Intel loves to show off huge numbers in SB-E doesn't mean anything in the real-world. They put them in for marketing purposes, nothing else. It's 15% faster overall than Bloomfield in IPC.

So what you're saying is that if someone where to build a 1155 system and a 2011 system and clocked them at the same speeds (CPU, memory, etc) that there would be little to no difference at all, correct?

 

[toyota]

So what you're saying is that if someone where to build a 1155 system and a 2011 system and clocked them at the same speeds (CPU, memory, etc) that there would be little to no difference at all, correct?

thats probably about right if you are referring to gaming. what does the 2600k provide over the 2500k in games? its HT and extra 2mb cache make basically zero difference.

 

[AdamK47]

Gaming with what card and how many? You could probably max out many graphics cards and get the same framerate with different CPUs and speeds.

 

[toyota]

Gaming with what card and how many? You could probably max out many graphics cards and get the same framerate with different CPUs and speeds.

well I have no idea at what point the cpu will start being the limitation comparing those 2 cpus but I am sure it takes a pretty massive gpu setup. even for high end gpu users we will have Ivy Bridge cpus out long before anyone sees more than a 1 fps difference between Sandy Bridge 1155 and 2011 clock for clock at 1920 or higher. the limitation for 1155 will likely be the boards themselves with their reduced pci-e lanes if running high end multi gpus. I know you like to justify your level of cpu but a $1000 cpu does nothing more than a 2500K even with high end gpus at settings any normal person would use. its like adding more horsepower to a car that can barely get any traction with the power it has.

 

[LOL_Wut_Axel]

So what you're saying is that if someone where to build a 1155 system and a 2011 system and clocked them at the same speeds (CPU, memory, etc) that there would be little to no difference at all, correct?

Less than 1-2% difference overall.

It's the same architecture. Look back at Bloomfield vs Lynnfield and the difference between the 2500K and 2600K with HT disabled.

 

[bridito]

Less than 1-2% difference overall.

It's the same architecture. Look back at Bloomfield vs Lynnfield and the difference between the 2500K and 2600K with HT disabled.

Not that I believe anything any longer unless it's independently verified, but you're saying that these slides:

http://www.donanimhaber.com/islemci...lemcisi-Core-i73960Xin-ilk-test-sonuclari.htm

are all wrong and LGA2011 brings a minor tweak that's next to nothing, right? I'm of the opposite opinion as I found the i7-3960 performance lower than I expected!

 

[AdamK47]

Here's what I'll do once I get an i7 3960X and an X79 board. I'll clock the CPU to the same 4200MHz I'm running my 980X at, the memory at the same 2000 DDR, and keep the three GTX 580s at the same 850/1100/1700. I'll show 980X/X58 scores, 3960X/X79 scores, and the max overclock scores of the 3960X/X79 (whatever I can get it up to 100% stable). Nobody on this forum can truly say they know what the results would be at this point. Anyone who says they do is pulling it from opinion, speculation, and favoritism to their platform of choice.

 

[RussianSensation]

Here's what I'll do once I get an i7 3960X and an X79 board. I'll clock the CPU to the same 4200MHz I'm running my 980X at, the memory at the same 2000 DDR, and keep the three GTX 580s at the same 850/1100/1700. I'll show 980X/X58 scores, 3960X/X79 scores, and the max overclock scores of the 3960X/X79 (whatever I can get it up to 100% stable). Nobody on this forum can truly say they know what the results would be at this point. Anyone who says they do is pulling it from opinion, speculation, and favoritism to their platform of choice.

:thumbsup: Looking forward to those benches.

 

[bridito]

:thumbsup: Looking forward to those benches.

Especially if they could be published, say, next week!

 

[AtenRa]

Here's what I'll do once I get an i7 3960X and an X79 board. I'll clock the CPU to the same 4200MHz I'm running my 980X at, the memory at the same 2000 DDR, and keep the three GTX 580s at the same 850/1100/1700. I'll show 980X/X58 scores, 3960X/X79 scores, and the max overclock scores of the 3960X/X79 (whatever I can get it up to 100% stable). Nobody on this forum can truly say they know what the results would be at this point. Anyone who says they do is pulling it from opinion, speculation, and favoritism to their platform of choice.

Nice to have those data but your 980X is not an SB and we clearly know that SB-E 3960X will be faster Clock to Clock.

What i really would like to see is a comparison that include Core i7 9xx (1366), SB Core i7 2600K (1155) and SB-E (2011) quads 4C8T.

 

[RussianSensation]

Nice to have those data but your 980X is not an SB and we clearly know that SB-E 3960X will be faster Clock to Clock.

It would still be nice to see the overall performance increase from a 6-core 12 threaded 980X to SB-E for reference purposes.

 

[AdamK47]

Nice to have those data but your 980X is not an SB and we clearly know that SB-E 3960X will be faster Clock to Clock.

What i really would like to see is a comparison that include Core i7 9xx (1366), SB Core i7 2600K (1155) and SB-E (2011) quads 4C8T.

I would if I had the desire to get an 1155 board and a 2600K CPU. It would make for a very interesting comparison.

 

[bridito]

Nice to have those data but your 980X is not an SB and we clearly know that SB-E 3960X will be faster Clock to Clock.

I'm not believing anything that comes out of Donair Haver until I see it independently verified. I still am not convinced those slides are accurate.

 

[LOL_Wut_Axel]

Not that I believe anything any longer unless it's independently verified, but you're saying that these slides:

http://www.donanimhaber.com/islemci...lemcisi-Core-i73960Xin-ilk-test-sonuclari.htm

are all wrong and LGA2011 brings a minor tweak that's next to nothing, right? I'm of the opposite opinion as I found the i7-3960 performance lower than I expected!

When comparing a 990X and a 3960X at the same clock speeds expect a 15-17% difference overall. When comparing a 2600K and a 3820 at the same clock speeds expect a 0-2% difference overall.

Intel's claims of 30-40% higher performance than the 990X by 30-40% are due to the new AVX instructions, nothing else. Their claims of this much performance enhancement due to higher memory bandwidth are absolutely false and should be dismissed. We already know this from Bloomfield and Lynnfield. The other ones, they're synthetics, so they should be dismissed as well. As for SB vs SB-E:

Let's be realistic here. It's the exact Sandy Bridge architecture we know with more theoretical bandwidth and more cache which will do next to nothing as we can see comparing the 2600K w/ HT disabled vs the 2500K.

Again, you buy into Sandy Bridge-E if you need those additional cores and/or more feature rich platform.

I don't know why people are expecting Intel to pull out more performance out of thin air. It's the same architecture.

Not really surprising at all. Sandy Bridge has 15% higher IPC; therefore, it's 15% faster overall than Nehalem.


Intel claiming much higher performance when due to AVX instructions is true. I find their claims of higher performance coming from higher memory bandwidth due to the additional memory channel to be absolute BS. It makes less than a 1% difference when comparing Bloomfield to Lynnfield. Bandwidth is definitely not an issue.

 

[Aikouka]

It seems everyone is focusing on clock speeds and such, but what about the fact that socket 2011 comes with twice as many PCI-E lanes for the GPUs? Right now with Socket 1155, I'm always looking at the fine print on motherboards to see what crazy limitations they put on me. For example, on one board, if I use the front-panel USB 3.0 connector on the motherboard, it limits my second PCI-E 16x slot to 1x or maybe it was 4x... but whatever the case, it really slows it down.

Hopefully with SB-E, we'll see something like a 16x + 8x configuration that leaves another 8 lanes for the "extras" such as USB 3.0 and bonus SATA 6Gbps ports, or maybe just a 2x 16x configuration without the extra fluff for those that don't need it.

 

[Compddd]

I'll wait for Ivy Bridge since it's cheaper.

 

[AtenRa]

It would still be nice to see the overall performance increase from a 6-core 12 threaded 980X to SB-E for reference purposes.

Sure thing, that's why i said its nice to have the data

 

[bryanW1995]

Here's what I'll do once I get an i7 3960X and an X79 board. I'll clock the CPU to the same 4200MHz I'm running my 980X at, the memory at the same 2000 DDR, and keep the three GTX 580s at the same 850/1100/1700. I'll show 980X/X58 scores, 3960X/X79 scores, and the max overclock scores of the 3960X/X79 (whatever I can get it up to 100% stable). Nobody on this forum can truly say they know what the results would be at this point. Anyone who says they do is pulling it from opinion, speculation, and favoritism to their platform of choice.

That will be very interesting to see. If scaling is the same for the 6 cores as the 4 cores, then I would estimate a 15% improvement clock/clock from the cpu and probably another 10-15% from higher clocks. so something ~ 25-30% total improvement. Not bad at all, though this sort of iterative improvement can be attributed imho more to a lack of competition from AMD rather than any sort of incompetence on intel's part. It's still probably worth it for type of high end user who bought a gulftown 18 mos ago.

 

[bryanW1995]

I would if I had the desire to get an 1155 board and a 2600K CPU. It would make for a very interesting comparison.

If I get one I'll send it to you for the comparo, though I admit the prospects of my buying a skt 1155 mobo are dim. I'll probably either get a BD when they get the kinks work out or hold off until IB-E.

 

[bridito]

When comparing a 990X and a 3960X at the same clock speeds expect a 15-17% difference overall. When comparing a 2600K and a 3820 at the same clock speeds expect a 0-2% difference overall.

Intel's claims of 30-40% higher performance than the 990X by 30-40% are due to the new AVX instructions, nothing else. Their claims of this much performance enhancement due to higher memory bandwidth are absolutely false and should be dismissed. We already know this from Bloomfield and Lynnfield. The other ones, they're synthetics, so they should be dismissed as well. As for SB vs SB-E:

Let's be realistic here. It's the exact Sandy Bridge architecture we know with more theoretical bandwidth and more cache which will do next to nothing as we can see comparing the 2600K w/ HT disabled vs the 2500K.

Again, you buy into Sandy Bridge-E if you need those additional cores and/or more feature rich platform.

I don't know why people are expecting Intel to pull out more performance out of thin air. It's the same architecture.

With respect to your opinion, I'd like to express mine. Intel is not stupid enough to try and sucker users to shelling out for an LGA2011 X79 just so they can have an extra couple of SATAs or other gimcracks. I am not going to argue the architecture angle, but it is my profound (albeit unfounded by evidence) belief that SB-E will not perform at a par with 2600K. No way. No how. Again, that's my opinion and I'm entitled to it.

 

[RussianSensation]

With respect to your opinion, I'd like to express mine. Intel is not stupid enough to try and sucker users to shelling out for an LGA2011 X79 just so they can have an extra couple of SATAs or other gimcracks. I am not going to argue the architecture angle, but it is my profound (albeit unfounded by evidence) belief that SB-E will not perform at a par with 2600K. No way. No how. Again, that's my opinion and I'm entitled to it.

Well it already happened with X58 vs. P55. After P55 came out, there was little benefit in owning the X58 platform:

- The motherboards were more expensive, but practically offered no difference (i.e., PCIe 8x/8x = 16x/16x, you could have purchased P55 with SATA3 off PCIe lanes too)
- It was a lot more expensive to get 3x2GB of Ram vs. 2x2GB of Ram, since back then Ram prices were much higher
- There was little benefit in the extra memory bandwidth for non-professional market space
- Core i7 CPUs had far inferior idle and load power consumption compared to the i7 860 / 870 on S1156, and didn't really have that much better overclocking either. Sure some hit 4.2-4.4ghz, but at that point power consumption on the CPU alone approached 300 Watts!

Of course at the time, there were power users who needed hexa-cores and a lot of RAM slots for expandability/video work/photoshop, etc.. S1366 also proved valuable because it launched 12 months prior to P55. So if you wanted the fastest platform and could afford it in 2008, you didn't really need to wait for LGA1155. You paid a premium but you owned the fastest platform for a full year before LGA1155 that you knew wasn't going to be faster.

The situation today is the complete opposite:

- We have had P67 platform since January, which was the fastest platform. So in fact, there was almost no reason to wait for SB-E, except for those very same power users who still require 6- core processors
- The fact that P67/Z68 fully support IB drop in upgrades means that the fastest quad-core CPUs may actually end up first on the 1155 platform as well
- It was pretty much rumored all this time that SB-E was unlikely to bring any benefits per clock since it wasn't a new architecture or a refresh like IB was
- X79 platform now has 4 Dimms instead of 6, which is actually a step back.
- The increased memory bandwidth of SB on 1155 has almost ensured that no memory bandwidth bottlenecks exist for consumer applications. The Quad-Channel memory support is therefore obviously aimed at non-consumer apps (such as server/workstation apps).

Practically, the X79 platform makes no sense whatsoever unless you are getting at least a 6-core processor with it imo. I don't even understand why Intel is bothering releasing a quad core on X79 unless it has a some magical overclocking headroom. The fact that they are throwing 14 SATA ports with it also sends a signal that the main purpose of X79 is to be a workstation/server platform.

 

[bridito]

Well it already happened with X58 vs. P55. After P55 came out, there was little benefit in owning the X58 platform:

- The motherboards were more expensive
- It was a lot more expensive to get 3x2GB of Ram vs. 2x2GB of Ram, since back then Ram prices were much higher
- There was little benefit in the extra memory bandwidth in the real world
- Core i7 CPUs had far inferior idle and load power consumption compared to the i7 860 / 870 on S1156, despite hardly overclocking that much better on average

Of course at the time, there were power users who needed hexa-cores and a lot of RAM slots for expandability/video work/photoshop, etc.. S1366 also proved valuable because it launched 12 months prior to P55. So if you wanted the fastest platform and could afford it in 2008, you didn't really need to wait for LGA1155.

The situation today is the complete opposite:

- We have had P67 platform since January, which was the fastest platform. So in fact, there was almost no reason to wait for SB-E, except for those very same power users who still require 6- core processors
- It was pretty much rumored all this time that SB-E was unlikely to bring any benefits per clock since it wasn't a new architecture.
- X79 platform now has 4 Dimms instead of 6, which is actually a step back.
- The increased memory bandwidth of SB on 1155 has almost ensured that no memory bandwidth bottlenecks exist for consumer applications. The Quad-Channel memory support is therefore obviously aimed at non-consumer apps (such as server/workstation apps).

Practically, the X79 platform makes no sense whatsoever unless you are getting at least a 6-core processor with it imo. I don't even understand why Intel is bothering releasing a quad core on X79 unless it has a some magical overclocking headroom. The fact that they are throwing 14 SATA ports with it also sends a signal that the main purpose of X79 is to be a workstation/server platform.

Granted about the workstation bit, but if LGA2011 is as much of a no-news platform why couldn't the workstation users go with the standard LGA1155? I'm sticking hard to my guesstimate of significantly higher performance all-round... but the benchys will tell the tale... I just hope I live long enough to see them released!

BTW, X79 has 8 slots in the Intel Silar mobo. That's the one I am lusting after!!!!!

 

[LOL_Wut_Axel]

Granted about the workstation bit, but if LGA2011 is as much of a no-news platform why couldn't the workstation users go with the standard LGA1155? I'm sticking hard to my guesstimate of significantly higher performance all-round... but the benchys will tell the tale... I just hope I live long enough to see them released!

BTW, X79 has 8 slots in the Intel Silar mobo. That's the one I am lusting after!!!!!

Why are you expecting Intel to get more performance out of the exact same architecture, just with more theoretical memory bandwidth, more cache, and a more feature rich platform?

As RussianSensation said, Sandy Bridge is not bandwidth starved (bottlenecked) AT ALL. That's one down. As for the cache difference, Sandy Bridge isn't cache starved at all either. You can see this comparing a 2600K with HT off to a 2500K. As for the more feature rich platform, that along with support for six-core CPUs should be the reason to get Sandy Bridge-E.

Don't expect any significant performance improvements at the same clocks. They'll be extremely minimal, if there's any.

Do we need to pull out Lynnfield and Bloomfield again? Guess we do.

http://www.anandtech.com/bench/Product/100?vs=192

What leads you to believe that Sandy Bridge-E will have more performance per clock than Sandy Bridge? Intel isn't gonna get more performance out of thin air.