Supposed Intel i7-3770K vs. i7-4770K Benchmarks Leaked

[N.Broekhuijsen]

Some Benchmarks have been leaked where the i7-3770K has been pitted against the upcoming i7-4770K.

Supposed Intel i7-3770K vs. i7-4770K Benchmarks Leaked : Read more

 

[ssdpro]

"Coolaler also mentioned that, while Haswell will be very nice for overclocking, it still suffers from the same temperature problems that Ivy Bridge does, although perhaps not to as great an extent"
This was an over-baked Ivy issue anyway. As a 2600k and 3770k user I never found 3770k to be wildly hotter at all. Clock for clock, yes, it is a bit warmer - few degrees C. Something to worry about or a reason to avoid as some did, not at all. Is the chip "suffering"? That is just clown stuff.

 

[Derbixrace]

meh, still not a big jump from sandy bridge.

 

[Augray37]

so an 8-10% increase in performance? not good, not bad, about what i expected. but these are 4 benchmarks only, and i'm guessing the 4770k used was an engineering sample? what i really want to see is how well they OC.

 

[darkchazz]

I upgraded from Phenom II X4 955 to i7 3770K in June last year. One of the best purchases I have made, next to my Crucial M4 SSD.
The performance jump was unbelievable. It also draws much less power and runs cooler. The phenom bottlenecked my gtx670 even when overclocked to 4ghz.

 

[icrf]

I'm actually most interested in something that utilizes AVX2. I have high hopes that x264 can incorporate them and have gains well in excess of 10%.

 

[Augray37]

I upgraded from Phenom II X4 955 to i7 3770K in June last year. One of the best purchases I have made, next to my Crucial M4 SSD.
The performance jump was unbelievable. It also draws much less power and runs cooler. The phenom bottlenecked my gtx670 even when overclocked to 4ghz.

hmm...that's weird, the gtx 670 didn't bottleneck my phenom ii x4 955 much at all really, not noticeably so. i was playing bf3 ultra at 1080p with no real lag at all. don't remember the fps exactly, but it rarely dipped into the 30s if i remember correctly.

 

[icemunk]

yay... for 5-10% speed gains... zzzz

 

[redeemer]

So these gains are from architectural changes or updated optimization coding that Intell pumps into these benchmark softwares?

 

[crisan_tiberiu]

well, since SB the TIC-TOC strategy was forgot since SB we have TIC-TIC. The only thing that we forget is that intel cand aford this, they are far ahead...

 

[ismaeljrp]

Still no reason to upgrade from my 2500k, I think I'll just wait for DDR4 systems.

 

[loops]

A 5-10% bump that will cost you a new chip and mobo. That is some expensive 10% gains.

 

[yhikum]

I would be interested in what were clock speeds using this benchmarks. There is no mention of that in article.
I have hard time believing that CPUMark99 scores were achieved on stock speeds. A bit of searching across forum boards show scores below 600 and scores themselves correlate to CPU speeds. This is for both AMD and Intel processors. Here is one for reference: http://forums.anandtech.com/showthread.php?t=2057154
Jump to 10% is quite significant when it runs non-optimized code for specific processor. This is due to simple fact that instructions on CPU cannot be executed faster than they already execute, unless you apply speed factor of clock speed. And since CPUMark99 is single threaded benchmark, clock speed plays important role.

 

[dimar]

For me 6x SATA 6Gb native ports on 8 series motherboards would be good reason to upgrade.

 

[loops]

A 5-10% bump that will cost you a new chip and mobo. That is some expensive 10% gains.

 

[CaedenV]

At this rate it's gonna be like 5 years before OC Sandy Bridge users have a reason to upgrade.

Just keep in mind that Intel is not looking at improving raw performance right now. Their largest competitor is not AMD, but the hordes of ARM manufacturers who are looking to break into the desktop and laptop space over the next few years. While the chip wattage is up a little bit this generation, the platform wattage is down, way down. The wattage increase is due to better IGP and integrated voltage regulation. These are useless improvements for desktop users, but it means a lot to laptop/netbook/tablet/AIO/server manufacturers where power usage vs performance is one of the largest considerations when choosing a chip.
Broadwell will take this to the next level by integrating more and more stuff on the chip. Eventually Intel will move the chipsets themselves onto the CPU so that you essentially have an SOC solution to properly fight against ARM. Once they get to that point then we will see a move back to improved raw performance again.

Until then I will enjoy my SB i7 and just throw money at the GPU as needed.

 

[ilysaml]

If steamroller takes on Sandy Bridge performance, there's a top chance AMD catch up with Intel.

 

[milktea]

I'll wait for 9770k before my next upgrade.

 

[nhat11]

The point of Haswell is the power savings for mobile devices. At the moment I'm waiting for that instead of a simple performance boost for my next laptop.

 

[Hupiscratch]

This could be the time to upgrade. I think it would be a nice improvement over my Athlon 64 FX-60

 

[InvalidError]

gogogogogogogo AMD
I'm tired of Intel getting away with this lazy junk

If AMD had a rabbit they could pull out of their hat, they would/should have done so already.

The cold reality is that single-threaded performance has pretty much reached as high as it is going to go so improvements there will remain pretty slow. With very few mainstream applications making reasonable use of more than two cores, there is very little demand for more than that so do not expect a desktop core-count race any time soon either.

Expect the number of people upgrading their PCs only every 5-7 years to increase - provided they do not fail first or get prematurely replaced due to degraded performance from push-pin HSF no longer making adequate contact.

 

[chumly]

I'm going to wait until I see what these things overclock like.

 

[nhat11]

The point of Haswell is the power savings for mobile devices. At the moment I'm waiting for that instead of a simple performance boost for my next laptop.

 

[killerclick]

Wow, 10% improvement? I used to upgrade when I could get twice the performance for the same amount of money, so now it's gonna be like 2024?

 

[ericjohn004]

This is actually sort of impressive. I mean that Cinebench score jumped a pretty decent amount. I'd say, if the performance gains are 10% across the board then the 4770k is a winner.
I know I know, I have a 3570k and I don't want the 4670k to be much faster than mine just so I won't have to spend more money on an upgrade or have a CPU that performs below par. But you can't denie that 10% is a rather good improvement. If it was more like 5% then yeah, who cares, but 10% is rather impressive considering they are already some powerful CPU's.
And as far as Synthetic's like SiSoft Sandra, the 4770k dominates, but I hate synthetic's anyways. I'm far more concerned with real world benchmarks like Cinebench. Of course it'd be nice If my 3570k had great synthetic's too but I really don't think that matters at all. I'd really like to see some game testing because that's what I do and that's what I care about. High Cinebench scores are fun too but gaming is where it's at.
I just think some of you need to start realizing that 10% is a nice gain. Not something to laugh at. AMD got a 15% out of the 8350 but had to raise the clocks to 4.0Ghz, 4.2Ghz if you include Turbo Boost. Intel got 10% more real world and crazy Synthetic's just from IPC, pretty impressive.