Page 2:Z270, Optane, Overclocking Tools, And HD Graphics 630
Page 3:Test Setup
Page 4:Results: Rendering And Office Applications
Page 5:Results: Workstation Applications
Page 6:Results: DTP And Multimedia
Page 7:Results: Gaming And Integrated Graphics (iGP)
Page 8:Intel Core i7-7700K: Power Consumption And Temperatures
Page 9:Intel Core i7-7700: Power Consumption And Temperatures
Page 10:Intel Core i5-7600K: Power Consumption And Temperatures
Page 11:Intel Core i5-7600: Power Consumption And Temperatures
Page 12:Inconsistent CPU Quality And Its Consequences
Intel Core i7-7700K: Power Consumption And Temperatures
It’s a well-known fact that processors vary widely in quality due to their manufacturing process. This time around, we didn’t get all of our test samples from Intel directly, but they’re the same retail models that anyone can buy at the store. This means that the manufacturer didn't provide pre-sorted CPUs. Unfortunately, the luck of the draw worked against our German lab, and we received a sample on the lower end of the quality range.
This doesn’t affect the benchmark results at all. However, it does have an impact on the power consumption and cooling results, as well as how far this particular CPU sample can be overclocked. Consequently, we have a page dedicated to the quality variations of the four CPUs that we tested. We were also able to use a test sample that Intel’s PR department gave us at the last minute.
The Intel Core i7-7700K
The “K” CPUs have an unlocked multiplier and a significantly higher base clock rate compared to non-K models. There’s also room for manual overclocking.
The Intel Core i7-7700K’s stock base frequency is 4.2 GHz. Even under extreme loads, it runs all four of its cores at a Turbo Boost frequency of 4.5 GHz. Let’s take a look at how it acts under different loads. Do keep in mind that this particular sample’s an extreme case and represents what happens if you have bad luck when buying this processor.
Core Voltage (Vcore)
We start with the actual core voltage (Vcore), which shouldn’t be confused with the voltage identification (VID), or what GPU-Z or CoreTemp show the Vcore to be. Our sensor readings come straight from the motherboard and represent the voltage that actually runs through its voltage converters.
The loads change both dynamically and quickly during gaming (grey curve). Voltage regulation works well with this kind of load pattern, since the CPU is only regulated down if there’s a strong load increase. This is also demonstrated by the blue curve, which represents the results from a constant computing load. The motherboard regulates the voltage down in order to keep the CPU safe from leakage currents. On the plus side, it does manage to keep the voltage constant for the entire duration, though.
There are a lot of observable fluctuations during the extreme stress test. This is because the CPU runs into its thermal limit in spite of our compact water cooling solution, and it throttles accordingly for safety reasons.
Normal Load: Gaming
We’re using a Watch Dogs 2 sequence to test the gaming load. The player stands in a well-populated spot in the city center, which creates a high CPU load due to the many NPCs and constant traffic. This makes the test very reliable. Across 30 minutes, the CPU load is very similar to the average CPU load that we measured during actual gaming. Of course, the average load might change based on the specific title you're looking at.
Now let's look at overall power consumption and the components that go into it, such as the IA cores, cache, and memory controller. After the warm-up phase, we measure a relatively high 77W for the entire CPU, 67W of which comes from the execution cores. The difference is covered by the needs of the Uncore (cache, memory controller, etc.) and leakage currents.
As the CPU heats up, power consumption goes up as well. This is a sign that leakage currents increase, pushing power up another 3W during our gaming workload.
The temperatures increase at a different pace, depending on the position of the sensor. They do stabilize after a maximum of 22 minutes at up to 71°C, though.
Heavy Load: Stress Test (Floating-Point Unit)
For our next test run, we tax the CPU's floating-point unit using the stability test included in FinalWire's AIDA64. This results in a power consumption measurement of 98W. Leakage currents increase right along with the temperatures, and this time the increase is significant.
With the Tcore reaching up to 85°C, we’re still in the clear. But that's not a good sign for the thermal headroom available when we want to start overclocking.
Maximum Load: Intel Power Thermal Utility (100%)
To prove the old adage that things can always get worse, we drop the hammer with Intel’s Power Thermal Utility, which isn't publicly available for Kaby Lake. The Core i7-7700K in our German lab consumed a hefty 137W. The one in our U.S. lab consumed 18W less than that!
Unsurprisingly, temperatures explode right along with the power consumption. The CPU hit 101°C very quickly, which means that it even exceeded its maximum temperature. Consequently, it throttled up to 25 percent. At least the processor survived our full 30-minute test.
Dissipating almost 140W isn't really a challenge for our water cooling solution. It’s able to deal with an Intel Core i7-6950X overclocked to 4 GHz, which generates a lot more heat than this. The problems lie elsewhere. First, there’s the CPU package's smaller surface area. Second, there’s the thermal paste, which might have made Intel's accounting department happy, but causes overclockers nothing but worry.
Intel Core i7-7700K vs. Core i7-6700K @ 4.5 GHz
For our performance benchmarks, we had the Core i7-7700K compete against a Core i7-6700K at the same frequency. The latter fared pretty well, but was also close to its maximum overclocking potential. We have an above-average chip, but it's far from a golden sample.
Consequently, we end up comparing a solid Skylake CPU to a bad Kaby Lake sample at the same clock rate. The result’s somewhat of a shocker: our Skylake CPU comes in just below Kaby Lake in our power consumption test, in spite of its manual overclock and a slightly higher voltage set in the BIOS.
This means that a decent retail Core i7-6700K beats a below-average retail Core i7-7700K. We already feel sorry for the online retailers that’ll have to deal with the returns and exchanges generated by enthusiasts trying to get the best possible sample.
The Core i7-6700K’s voltages are always higher than those of the Core i7-7700K with its new manufacturing process. This doesn’t guarantee better efficiency than the preceding processor generation, though.
We have to stress that the variability of the processors’ quality can have a major impact on the results. We go over this variability later on, on its own page.
We learn two things from our experiences with Intel's Core i7-7700K. First, Intel’s Power Thermal Utility is no joke. Second, the Core i7-7700K does have some thermal and performance reserves, so long as you don't hit it with such a worst-case workload. There should be some room for overclocking, though it won't be much.
- Z270, Optane, Overclocking Tools, And HD Graphics 630
- Test Setup
- Results: Rendering And Office Applications
- Results: Workstation Applications
- Results: DTP And Multimedia
- Results: Gaming And Integrated Graphics (iGP)
- Intel Core i7-7700K: Power Consumption And Temperatures
- Intel Core i7-7700: Power Consumption And Temperatures
- Intel Core i5-7600K: Power Consumption And Temperatures
- Intel Core i5-7600: Power Consumption And Temperatures
- Inconsistent CPU Quality And Its Consequences