Core i7-4960X "Ivy Bridge-E" Roughly 10% Faster than i7-3970X: Early Tests

[15th Warlock]

IB-E doesn't excite me the least, I hate how the "extreme" processors are almost a year behind mainstream processors in terms of architecture releases...

Haswell on the other hand would've been a better upgrade option, but I read Haswell-E will support DDR4, so there's simply no chance it'll work on X79 boards, by the looks of it, we won't be seeing it until mid 2014 anyways...

 

[Octavean]

If Intel wants us X79 hexa users to upgrade they need to give us either an unlocked 8 core cpu or an unlocked qpi 6 core to play with on enthusiast 2p mobos.

Not worthy draining my loop for a 5% increase in IPC.

Might aswell bump my 3930k to 5,3Ghz and call it a day until Haswell-E comes.

If you look here:

http://www.techpowerup.com/182238/Intel-Core-i7-quot-Ivy-Bridge-E-quot-HEDT-Lineup-Detailed.html

You'll see that the entry level Ivy Bridge-E processor is listed as a Core i7 4820K which if correct could mean all three processors in the lineup are fully unlocked.


That might mean something,....or not,...

 

[qwerty_lesh]

Id be quite happy still on my old 920 if I didn't have the opportunity to shift up to a 3930k free of cost.

The way I see it, IVY-E was never going to be appealing enough to upgrade to.

Its just like Gulftown was, on the tylersburg platform.. Nice to know its out there but not to actually go out and spend on.

The great thing now is if you have practically anything of the last several generations which has good processing power, you can sit on it for many years and be happy.

Conroe/Wolfdale Nehalem and Sandybridge were all big steps forward, and all quite sufficient for almost everybody.

To me, this is not really bad news at all, it doesn't phase me that its not a leap forward.
The same goes with Haswell, I can see why many will want to see a big performance jump when its out but if it doesn't happen, personally I'm not really concerned.

Oh i need to update my specs on here. lawl

 

[Hayder_Master]

seems no different at all, thanks for the 100mhz more which is make the different

 

[Fourstaff]

Might aswell bump my 3930k to 5,3Ghz and call it a day until Haswell-E comes.

Best decision for almost all users with a chip more powerful than 920. After all, Intel's focus is no longer brute power at all cost. They are taking a breather to balance other factors like iGPU and power consumption etc.

 

[radrok]

Best decision for almost all users with a chip more powerful than 920. After all, Intel's focus is no longer brute power at all cost. They are taking a breather to balance other factors like iGPU and power consumption etc.

If my RE2 didn't crap out I'd still be using my Gulftown chip.

I literally kept my 3930k inside its box for 6 months before upgrading.

 

[Octavean]

Anyone who had an LGA2011 based Sandy Bridge-E processor probably didn't expect much of a performance bump from Ivy Bridge-E.

Due to the Nomenclature as well it makes sense considering Sandy Bridge to Ivy Bridge on the same LGA1155 socket wasn't earth shattering either but at least there was an upgrade for the socket before moving on,....

Thats not small point either because think about the bitching and complaining that occurs when there are no upgrades before moving on to a new socket,....

At least the option is there.

Also note that not everyone is upgrading. Sometimes people and businesses find a need for an additional system or systems. Thats additive, so what would you buy if your in need of another system, the same old Sandy Bridge-E or the new Ivy Bridge-E. In that case I would probably buy the new Ivy Bridge-E,...but I still have to see retail product / reviews first.

Keep in mind that we have heard that Haswell will ship with the USB bug / older stepping chipset initially,...

 

[Dent1]

I cannot understand that everyone thinks there is a 10% bump in IPC! It isn't! Max gain is almost 10%. In the majority of the tests the gain is 5-7%. And that is sad since the clocks are the same as IB. Total gain is none to move someone to upgrade.

Ivy Bridge-E isnt mean to be an upgrade from Sandy Bridge or Ivy Bridge builds. It's for new builds from the Core 2 Duo or Phenom II generation or older wanting to jump on the fastest available. You don't need to jump on the latest architecture every round!

I hope AMD's SR is what expected Only then Intel is going to bring faster CPUs or lower the prices.

It doesn't work like that. Intel could release the fastest CPU on the planet, but the yield will always be disappointing without the proper software optimisation. We saw this with Bulldozer, on paper it should have outperformed Sandy Bridge but without software support the results didn't materialise.

Steamroller won't change much, Intel can have a slower CPU priced higher and it'll still generate just as much sales. We've seen this in history with the Athlon/Athlon XP/Duron/Sempon vs P3/Celeron/P4/Pentium D. Lower performance doesn't always mean less sales or lower prices for Intel.

 

[Aquinus]

You don't need to jump on the latest architecture every round!

Just to prove your point a bit more, even this statement is incorrect. IVB-E is not a new architecture. It is a die shrink. SB-E and IVB-E for the most part will work exactly the same (like SB and IVB,) but there are a couple different features and smaller circuitry inside the CPU. That's it. Nothing earth shattering, nothing special, just simply a die shrink.

 

[Nordic]

here's no way that Intel can make any inroads into the phone, tablet or embedded devices market

I don't know about that. Amd has some really promising low power x86 chips coming that look reallllly good on paper. It has yet to be shown if it can compete with arm, but if amd can make it so can intel.

 

[Octavean]

I don't know about that. Amd has some really promising low power x86 chips coming that look reallllly good on paper. It has yet to be shown if it can compete with arm, but if amd can make it so can intel.

But AMD is also just as likely to slap some really good graphics on an ARM SoC since they have much less to lose from the success of ARM then Intel.

 

[Jstn7477]

Hey guys, have you considered that there is probably a nice power consumption decrease with these chips vs. SB-E? Apparently nobody seems to understand that what Ivy Bridge does with 80w is comparable to what Sandy Bridge does at around 130w (my 3770K @ 4.3GHz/1.18v vs. my 2600K @ 4.3GHz 1.32v). Just saying...

 

[EarthDog]

SB = 95W while IB =77W. Perhaps IB-e will fit in to sub 100W package. On the surface a ~25% decrease is nice, however unless you are running the thing 24/7... doesnt translate to much dollar wise.

 

[radrok]

Still not worth to upgrade based on power consumption.

I bet these chips will draw a ton of power too if kept clocked at what I run my 3930k (5 Ghz), we are talking upwards of 300w

 

[EarthDog]

I bet it will draw less than your 3930K...

 

[radrok]

I agree on that but will the difference be worth it?

 

[buggalugs]

Just to prove your point a bit more, even this statement is incorrect. IVB-E is not a new architecture. It is a die shrink. SB-E and IVB-E for the most part will work exactly the same (like SB and IVB,) but there are a couple different features and smaller circuitry inside the CPU. That's it. Nothing earth shattering, nothing special, just simply a die shrink.

Well its not just a die shrink, ivy bridge uses tri-gate transistors, that's a pretty major design change.

The biggest improvement is likely memory performance/latency, power consumption, and overclocking.

The biggest letdown for me is Intel still relying on X79 boards for this new chip. an unfinished platform with no Intel USB 3.0 only 2 Intel sata 6Gb/s etc. Its weird when the mainstream platform has better motherboard features than the highend that costs twice as much.

 

[Aquinus]

Well its not just a die shrink, ivy bridge uses tri-gate transistors, that's a pretty major design change.

Which has no impact on performance. This is a heat/power optimization more than anything else, also I think Intel would be hard pressed to not use a multi-gate transistor at 22nm considering the physical limitations with circuits that small.

The biggest improvement is likely memory performance/latency, power consumption, and overclocking.

Power consumption is due to the multi-gate transistors and the die shrink.
Overclocking isn't any better than SB. IVB just has a better IPC so each Mhz goes a bit further (like 10% further ,) so even though you might not get clocks as high as a SB chip, you're getting more work done because it's doing 10% more in the same amount of time with the same frequency.
You can thank the die shrink for the better memory latencies too.

So yeah, most of the performance benefits came from the die shrink. The power consumption improvements come from both the shrink and the multi-gate transistors.

The biggest letdown for me is Intel still relying on X79 boards for this new chip. an unfinished platform with no Intel USB 3.0 only 2 Intel sata 6Gb/s etc. Its weird when the mainstream platform has better motherboard features than the highend that costs twice as much.

Don't call it a letdown unless you own one and have legitimately have been let down by it. I'm perfectly happy with my X79 board and I think that most people who insult skt2011 don't really know what they're talking about. I find it astonishing the people complain about really stupid things like X79 not having enough SATA 6 ports or not many USB 3.0 ports (mine has 6 on the back, plus headers for another 4 so that's a matter of opinion,) when the CPU has 40 PCI-E lanes. You need more ports? Get a RAID card. They didn't load the CPU full of PCI-E slots and lanes for nothing.

I would also like to see your 3770k use VT-d and run 64Gb of ram like my 3820 can. What about features again?

People complain about X79 when the real power house is SB-E. The PCH does so little now, it almost hardly matters if you really need more than what it offers. The PCH does enough and if you need more, you really should get dedicated hardware. Remember, the PCH is on DMI not QPI or PCI-E. It can only do so much.

 

[jihadjoe]

I don't see why anybody is too surprised or disappointed by this considering we've already seen what the transition from Sandy to Ivy did on LGA1155.

 

[LAN_deRf_HA]

SB = 95W while IB =77W. Perhaps IB-e will fit in to sub 100W package. On the surface a ~25% decrease is nice, however unless you are running the thing 24/7... doesnt translate to much dollar wise.

I think power consumption tests have shown that the power difference is much smaller, like 4 watts. The TDPs aren't that relevant these days.

 

[Patriot]

DDR3-1600 at 11-11-11-28 ?
Really pushing the envelope.

Nice catch.... Cinebench loves high frequency and low latency... cas 11 ddr3 1600 is no doubt hampering the performance.

 

[GamerGuy]

I hopped on LGA2011 platform at launch, and from what I had read then, IB-E could be a consideration for an upgrade even when LGA1155 is replaced by LGA1150. But, looking at what IB-E has to offer (still a hexacore, IF it had been octocore, I might be mentally masturbating over this), or lack thereof, I'd be perfectly happy with this setup I have for a while more. My board has 4x USB 3.0 ports at the back, and I have 2x USB 3.0 ports used on my case.....I use these ports mainly for USB 3.0 external HDDs. All other peripheral devices that use USB can be done on the USB2.0 ports. A couple of the reasons why I'd gone LGA2011 is the PCIe lanes on this chipset, native 40 PCIe 3.0 lanes (my cards doing PCIe 3.0 x16/x8/x16) as well as a host of features the chipset support.

 

[AsRock]

SB = 95W while IB =77W. Perhaps IB-e will fit in to sub 100W package. On the surface a ~25% decrease is nice, however unless you are running the thing 24/7... doesnt translate to much dollar wise.

Well i have known both of my i5 and i7 idle around 72w-77w. But no were near 100w for either of them idle.

You gotta be careful on how this is tested too as some mobo takes a load more power even at idle.. My Asus Maximus (x38) used to run 190w idle were with another board around 100w with the same chip.

I think power consumption tests have shown that the power difference is much smaller, like 4 watts. The TDPs aren't that relevant these days.

It is about 4w when idle and that's if it disable HT or not compering my 2 chips..

 

[jihadjoe]

I don't think any of the current CPU's would last as long as this one did.

My opinion is totally different. ANY modern cpu will probably last as long, if not even longer because applications aren't getting much more demanding.

I mean how much can a game demand before any extra CPU is totally irrelevant? If you can map the game world at 60fps, maybe throw in a few cycles for AI and stuff that's in-motion, that's about all it's ever going to need. Everything else goes to the GPU.

Office apps, even more so. I can only imagine how many trillions of clock cycles are wasted while Word waits for your next keystroke. The fact of the matter is CPUs now are more than good enough for what we need, and Intel's direction in optimizing toward greater integration and lower power (as opposed to more outright computing power) is totally justified.

 

[Aquinus]

Well i have known both of my i5 and i7 idle around 72w-77w. But no were near 100w for either of them idle.

You gotta be careful on how this is tested too as some mobo takes a load more power even at idle.. My Asus Maximus (x38) used to run 190w idle were with another board around 100w with the same chip.

Were you measuring the 8-pin EPS connector to get those numbers? My rig idles at 200-watts but that doesn't mean the CPU is idling at that. According to Cadaveca's review of my board, the VRMs use very little power when the CPU is idle, so the majority of that must be my video cards and hard drives.

This is a good read: http://www.linuxfordevices.com/c/a/...PU-power-consumption-a-challenge-authors-say/