
Tuning Core i7-4770K up to 4.2 GHz (using the four-core Turbo Boost setting) has a nearly 10 percent impact in our FineReader OCR workload. From there, the stock Core i7-4790K shaves off a couple of seconds more, while overclocking further whittles away at the completion time.

Same story in Visual Studio. This time around, the overclocked 4770K and stock 4790K nearly tie, which is the outcome we'd expect from two processors able to push four cores at a consistent 4.2 GHz. Nudging the clock rate up to 4.4 GHz across four cores gives Devil's Canyon an additional edge.

This is the only page with single-threaded tests, but TotalCode Studio isn't one of them. The bump you get from overclocking Core i7-4770K is easily replicated (and then some) by a stock Core i7-4790K.

The same goes for HandBrake, though the performance improvement is even more pronounced. From stock 4770K to stock 4790K, we report an almost 11 percent boost.

From the factory, a Core i7-4770K jumps as high as 3.9 GHz with one core active while the 4790K gets up to 4.4 GHz. That 500 MHz difference translates to an improvement of almost 12 percent. Pushing Devil's Canyon's clock rate up to 4.7 GHz knocks another couple of seconds off of the completion time.

Our iTunes test is similarly set up to utilize one core. The percentages work out to be fairly similar to LAME: expect double-digit gains from the frequency headroom freed up by more effective cooling.
- Intel Core i7-4790K: Devil's Canyon Is For Enthusiasts
- Overclocking Core i7-4790K And TIM Performance
- How We Tested Core i7-4790K
- Results: Synthetics
- Results: Content Creation
- Results: Adobe CC
- Results: Productivity And Media Encoding
- Results: Compression Apps
- Results: Power Consumption
- Core i7-4790K Adds Enthusiast Appeal To Haswell
bring back solder intel.
I built my PC at the end of last year, beginning of this one and went with a i7-4930k. I really wanted a six core processor and have not been disappointed. I have been itching to build another PC because it was really fun to put the plan of components together and although my hands were to big and my medical conditions prevented me from getting to do a lot of the building, my wife helped a lot with that part and it was nice to see the finished product in action. With that being said, I don't have a lot of money for anything right now and hope that my disability pay finally comes through so I can start picking together parts for a computer for my wife. She won't need anything as powerful as I have, and the i7-4790k sounds pretty sweet.
On a side note, this website annoys me. I click to add comment and the default fields are for signing up, no logging in, and when I do, I am back to the homepage. Great.
bring back solder intel.
Second problem: Why is Tom's using 1.275 V for 4200MHz on both units? Is that actually 1.275v with LLC disabled or is LLC on a setting resulting in the lowest load voltage? If one of those units need 1.275v to be stable at 4200 you have a real donkey sample on your hands. Even the worst i7-4770k are stable at 1.20v @ 4200. Or was the over voltage designed to test an unrealistic incompetent situation to either emphasize or DE-emphasize the TIM difference?
Voltage wall is still at the approximate same place. Heat is still the limiting factor. I expect some of the better binned 4670K will hit equal or better than the 4690K.
I'll stay with my 3570K @ 4.3ghz - this clearly isn't much of a step up. Looks like I'm waiting for a DDR4 system in a couple years.
Not impressed.
My interest lies in, will the devils canyon i5 4690k be able to hit the same clocks and temperatures as the i7 4790k at the same voltages, or for the i7 is there just a more extensive binning process or something of the sort? Your comment at the beginning of the article when discussing these realistic and sustainable overclocks really hit home in relation to this.
Second problem: Why is Tom's using 1.275 V for 4200MHz on both units? Is that actually 1.275v with LLC disabled or is LLC on a setting resulting in the lowest load voltage? If one of those units need 1.275v to be stable at 4200 you have a real donkey sample on your hands. Even the worst i7-4770k are stable at 1.20v @ 4200. Or was the over voltage designed to test an unrealistic incompetent situation to either emphasize or DE-emphasize the TIM difference?
The point of running both CPUs at the same voltage and the same clock rate is measuring the difference of the TIM. For the rest of the tests, each chip is pushed as fast as it'll go, stably.
The FX-8350 at stock was roughly the same time as the i7-3770K in your 3DSM fly-by render ...
So why would you go out of your way to compare Devil's Canyon to the A10-7850K ??
Still, even though the performance is another step faster than AMD, for my purposes the cost is still a bit high. Still, I'd love to see what it can do at 5.0 GHz.