Intel Devil's Canyon: Core i7-4790K CPU Review

[yRaz]

That is true but it will happen since it will be better to market more cores. Some people just look at cores and think, "wow, this has twice as many cores. It must be 2x as fast!"

if that were true AMD would be out selling Intel. They tried that marketing strategy with 4x4 when they released socket AM2. Even tried marketing a new measurement, "coreghz" multiplying cores by frequency. Let me just say there is a reason it isn't around anymore.

 

[Movatheaiur]

My 8 threads clocked @4.4GHZ is still not left behind.

 

[Steve]

Great review @Steve. I have a comment about the Photoshop benchmark though: it might be best to use Photoshop CC since it has numerous engine improvements over CS6. It may lead to more accurate/realistic/up-to-date benchmarks.

Overall it looks like a nice chip, I think Intel has squeezed a lot out of 4 cores. It is time for more cores .

I am a bit annoyed with myself. In my haste to get the review done I copy and pasted the wrong heading into Excel. Those are Photoshop CC results, I haven't used CS6 for a while. I will fix the graph now thanks.

 

[Jad Chaar]

I am a bit annoyed with myself. In my haste to get the review done I copy and pasted the wrong heading into Excel. Those are Photoshop CC results, I haven't used CS6 for a while. I will fix the graph now thanks.

No problem! Great work .

 

[Alpha Gamer]

"Intel's Australian PR company demanded to have the processor back within 7 days"
you don't get to keep the stuff you test?
I was pretty sure you did, so that you could test them again on newer games, gpus or whatever comes to mind.

 

[Steve]

"Intel's Australian PR company demanded to have the processor back within 7 days"
you don't get to keep the stuff you test?
I was pretty sure you did, so that you could test them again on newer games, gpus or whatever comes to mind.

You would think that would be the case since it makes sense. Thankfully for 90% of the companies that we do deal with this is the case.

However Intel PR is managed by Spectrum in Australia and they are massively tight with samples. Another company that is a pain to deal with is Asus and that is why you almost never see an Asus product reviewed at TechSpot. Both Asus and Intel are managed by Spectrum in Australia. Still I can’t just blame Spectrum, I know for a fact that the people working for Asus in Australia have no idea about PR.

http://www.channelnews.com.au/sales.../W8D2H5J5-asus-pr-goes-pear-shaped-again.aspx

 

[misor]

I think Intel has squeezed a lot out of 4 cores.
It is time...
Four more cores!!!

jc713, fixed that for you.

I hope amd will release Angel's Breath processor series to counter the Devil's Canyon.

 

[theeviljohn]

At the same clock speeds what're the temperature differences between the 4770K and the 4790K? I'm wondering if they really went back to the Sandybridge TIM (Solder,) or are using something else inferior. At this point it seems Intel is content with forgetting about OC on their budget platforms, and focusing on forcing people to the more expensive X79, soon to be X99, platform.

Its actually on the last page of the article. It runs a degree or 2 lower. *edit: whoops mis-read your comment. But considering it runs cooler at a higher clock/wattage it would imply that have improved TIM as they claim

 

[Guest]

Meh still isn't worth upgrading from a OC'd 2500k, think I'll wait for Skylake and DDR4

 

[Guest]

I have the i7-4790K - its actually a great deal for those of us who simply want a fast cpu at good price and have zero interest in overclocking.

I have to be also thinking this release for Intel is not so much about getting those who have the 4770k to jump and buy again as those who have been on the fence and didn't upgrade to the 4770K.

I just don't see the 4790K getting beat this time by the extreme models (5960x) for gaming due to is slow core speeds (even though of course there are more cores) - Hence my opinion is once the x99 cpu's release another wave will upgrade and many of them will settle on the 4790K in the end.

Thank you Steven for the review, TS does a great job with all the reviews (and yes, its pretty hard to keep everyone happy all the time).

 

[Guest]

Are you sure about that? I've had a chance to try them both on custom builds, the 4820k is impressive but the 4770k definitely outperforms it.

 

[Guest]

Why didn't you use a Z97 motherboard for your testing? That is the chipset made for Devil's Canyon. Not a lot of difference between Z97 and Z87 I know, but there could have been some. Enough to give it an even more little edge over i7 4770k. Should have went with Z97.

 

[Guest]

Indeed you are correct, the i7 4970k is faster than the i7 4820k, and runs cooler as well all things being equal.

The only advantage the 4820k has is the extra PCIe lanes.

 

[Guest]

Why didn't you use a Z97 motherboard for your testing? That is the chipset made for Devil's Canyon. Not a lot of difference between Z97 and Z87 I know, but there could have been some. Enough to give it an even more little edge over i7 4770k. Should have went with Z97.

Agreed, I enjoyed Steve's review, but a Z97 board should have been used. Very possible this is a variable that would have made a small difference in performance. Yes they are "almost" the same, but not exactly so.

 

[Steve]

Why didn't you use a Z97 motherboard for your testing? That is the chipset made for Devil's Canyon. Not a lot of difference between Z97 and Z87 I know, but there could have been some. Enough to give it an even more little edge over i7 4770k. Should have went with Z97.

A Z97 board was used, I have fixed the system specs but we did talk about using a few Z97 boards to try and improve overclocking performance.

Finally the performance difference between the Z87 and Z97 chipsets for either the 4770K or 4790K processors should not be described as "Not a lot of difference" but rather NO difference. Why would the Z97 chipset make the 4790K faster but not the 4770K faster? Simply put it wouldn't.

 

[Guest]

Maybe, but it should be benched/tested and proven so. If there was no difference at all in Z87- Z97 a i7 4790k would work on any Z87 board right out of the box - right now. As of right now it won't, so there are obviously some differences in both the board and the CPUs, minute as they may be. Only certain Z87 boards after BIOS updates will work with DC according to Intel (and they may still be glitchy), and since its still so early in the game with DC, who's to say that an i7 4970k will perform exactly the same on a ASUS Z87 and an ASUS Z97 board with all other things being equal. We don't know yet. I'm betting they'll be some difference. Maybe tiny, maybe big. But some difference. But hey, this is just MHO.

Apologies if you used a Z97 board for your review. I didn't catch that.

 

[Guest]

***sorry obviously I meant i7 4790k.

 

[Steve]

Maybe, but it should be benched/tested and proven so.

It was tested, sorry for the typo. Also it was proven many years ago that chipsets have nothing to do with performance once AMD moved their memory controller on-die.

If there was no difference at all in Z87- Z97 a i7 4790k would work on any Z87 board right out of the box - right now. As of right now it won't, so there are obviously some differences in both the board and the CPUs, minute as they may be.

The boards BIOS cannot detect the processor and therefore won’t post. That is the only issue here and it has nothing to do with performance. Flashing the BIOS with a version that recognizes the CPU ID fixes that.

Only certain Z87 boards after BIOS updates will work with DC according to Intel (and they may still be glitchy), and since its still so early in the game with DC, who's to say that an i7 4970k will perform exactly the same on a ASUS Z87 and an ASUS Z97 board with all other things being equal.

The 4790K works on all Z87 boards as long as a BIOS update to identify the CPU is available. I am "to say" the 4790K will perform exactly the same on an Asus Z879 as a Z97, you can take that to the bank. My opinion might mean little but I am willing to put it out there.

We don't know yet. I'm betting they'll be some difference. Maybe tiny, maybe big. But some difference. But hey, this is just MHO.

There won’t be a difference, at least nothing that can be accurately measured using benchmark software. Hell we don’t even see a difference in performance between brands anymore. If you take a budget MSI board for example and a high-end Asus board the performance is very much the same as long as everything else remains equal. This has been the case for a long time now making motherboard reviews quit boring.