There is a variety of customer groups in need of as much processing power as possible. Especially in the field of content creation, which includes editing pictures and videos or even rendering a complex scenery with beautiful effects, performance improvements quickly translate into higher income by the end of the month. With this series of articles we're going to compare two processors. The CPUs tested are being run at stock clocks as well as overclocked and apart from that we will also vary the memory frequency. It is our goal to show which CPU is how much faster or slower than another one.
|
Motherboard |
- ASUS Rampage V Extreme
- MSI Z170I Gaming Pro AC (BIOS 1.5)
|
| CPU |
- Intel Core i7-6950X @ Default
- Intel Core i7-6950X @ 4.2GHz
- Intel Core i7-6950X @ 4.2 GHz DDR4-3000
- Intel Core i7-6700K @ Default
- Intel Core i7-6700K @ 4.2 GHz
- Intel Core i7-6700K @ 4.2 GHz DDR4-3000
|
|
Memory |
- G.Skill Ripjaws F4-3000C15Q Quad Channel 4x4GB
|
|
Graphics Cards |
- nVidia GeForce GTX 980
Ti
|
|
OS |
|
|
System Drive |
|
|
PSU |
- Seasonic Platinum SS-1000XP / 1000 Watts
|
[pagebreak]
PCMark08
PCMark 8 includes tests based on popular Adobe and Microsoft applications. In order to test the max perfomance of our system, we run the Creative benchmark suite in “Conventional” mode without OpenCL acceleration. PCMark 8 Creative benchmark includes workloads typical for enthusiasts and professionals who work in the field of content creation. The score in PCMark 8 Creative is measured in points where higher is better.
Creative
| |
Points |
% |
| Core
i7-6950X @ Default |
5'869 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
6'287 |
107.12 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
6'344 |
108.09 |
| |
|
|
| Core
i7-6700K @ Default |
5'672 |
96.64 |
| Core
i7-6700K @ 4.2GHz |
5'722 |
97.50 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
5'753 |
98.02 |
[pagebreak]
3DMark
3DMark is a benchmarking tool created and developed by Futuremark Corporation to determine the 3D rendering performance of GPU and CPU. In this article we only use the Physix workload testing routing to find out how capable a processor is. The score is measured in points where higher is better.
Fire Strike Physix
| |
Points |
% |
| Core
i7-6950X @ Default |
20'915 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
25'807 |
123.39 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
26'016 |
124.39 |
| |
|
|
| Core
i7-6700K @ Default |
11'964 |
57.20 |
| Core
i7-6700K @ 4.2GHz |
12'648 |
60.47 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
13'614 |
65.09 |
3DMark 11 Physix
| |
Points |
% |
| Core
i7-6950X @ Default |
16'098 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
17'324 |
107.62 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
19'867 |
123.41 |
| |
|
|
| Core
i7-6700K @ Default |
9'776 |
60.73 |
| Core
i7-6700K @ 4.2GHz |
9'882 |
61.39 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
12'098 |
75.15 |
[pagebreak]
Microsoft Excel
Microsoft's Excel doesn't need much introduction, since we believe almost all of our readers have worked at least once with this application. In order to check performance we're using a file which conducts extensive calculations. In order to determine performance we measure the time it takes to finish the calculations.
| |
Points |
% |
| Core
i7-6950X @ Default |
390 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
386 |
101.04 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
380 |
102.63 |
| |
|
|
| Core
i7-6700K @ Default |
350 |
111.43 |
| Core
i7-6700K @ 4.2GHz |
346 |
112.72 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
342 |
114.04 |
[pagebreak]
Cinebench R15
Cinebench is a real-world cross platform test suite that evaluates your computer's 3D rendering capabilities. The test scenario uses all of your system's processing power to render a photorealistic 3D scene (from the viral "No Keyframes" animation by AixSponza). This scene makes use of various algorithms to stress all available processor cores.
| |
Points |
% |
| Core
i7-6950X @ Default |
1'856 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
2'156 |
116.16 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
2'201 |
118.59 |
| |
|
|
| Core
i7-6700K @ Default |
825 |
44.45 |
| Core
i7-6700K @ 4.2GHz |
907 |
48.87 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
914 |
49.25 |
[pagebreak]
SiSoft Sandra
SiSoft Sandra is a theoretical multi purpose benchmark suite which is able to show differences in CPU architectures. Therefore on this page you'll find Arithmetic-, Multimedia- and Cryptography-benchmarks.
Arithmetik Dhrystone
| |
GIPS |
% |
| Core
i7-6950X @ Default |
295.16 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
352.17 |
119.31 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
353.66 |
119.82 |
| |
|
|
| Core
i7-6700K @ Default |
162.97 |
55.21 |
| Core
i7-6700K @ 4.2GHz |
199.84 |
67.71 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
202.41 |
68.58 |
Arithmetik Whetstone
| |
GFLOPS |
% |
| Core
i7-6950X @ Default |
247.50 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
298.00 |
120.40 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
299.12 |
120.86 |
| |
|
|
| Core
i7-6700K @ Default |
107.65 |
43.49 |
| Core
i7-6700K @ 4.2GHz |
113.91 |
46.02 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
114.25 |
46.16 |
Multimedia Integer
| |
MPixels/sec |
% |
| Core
i7-6950X @ Default |
895.54 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
1'007.25 |
112.47 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
1'054.84 |
118.12 |
| |
|
|
| Core
i7-6700K @ Default |
498.75 |
55.69 |
| Core
i7-6700K @ 4.2GHz |
521.65 |
58.25 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
525.73 |
58.71 |
Multimedia Floating Point
| |
MPixels/sec |
% |
| Core
i7-6950X @ Default |
921.00 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
1'001.86 |
108.78 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
1'087.28 |
118.05 |
| |
|
|
| Core
i7-6700K @ Default |
486.51 |
52.82 |
| Core
i7-6700K @ 4.2GHz |
506.65 |
55.01 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
515.13 |
55.93 |
Cryptography AES
| |
GB/sec |
% |
| Core
i7-6950X @ Default |
19.32 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
18.18 |
94.10 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
18.31 |
94.77 |
| |
|
|
| Core
i7-6700K @ Default |
7.84 |
40.58 |
| Core
i7-6700K @ 4.2GHz |
10.46 |
54.14 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
10.53 |
54.50 |
Cryptography SHA
| |
GB/sec |
% |
| Core
i7-6950X @ Default |
11.58 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
13.72 |
118.48 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
13.98 |
120.73 |
| |
|
|
| Core
i7-6700K @ Default |
1.56 |
13.47 |
| Core
i7-6700K @ 4.2GHz |
1.62 |
13.99 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
1.62 |
13.99 |
[pagebreak]
Blackhole Benchmark Multithreaded
Black Hole evaluates the raw double precision performance with floating point as well as integer operations. There two different routines: multithreaded and singlethreaded. Multithreaded will load all available threads, while singlethreaded will use only one thread.
Single Thread
| |
Points |
% |
| Core
i7-6950X @ Default |
2'852 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
3'508 |
121.30 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
3'587 |
124.03 |
| |
|
|
| Core
i7-6700K @ Default |
3'852 |
133.20 |
| Core
i7-6700K @ 4.2GHz |
4'004 |
138.45 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
4'075 |
140.91 |
4 Threads
| |
Points |
% |
| Core
i7-6950X @ Default |
3'680 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
4'428 |
120.33 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
4'436 |
120.54 |
| |
|
|
| Core
i7-6700K @ Default |
5'236 |
142.28 |
| Core
i7-6700K @ 4.2GHz |
5'444 |
147.93 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
5'484 |
149.02 |
Multithreaded
| |
Points |
% |
| Core
i7-6950X @ Default |
11'700 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
14'080 |
120.34 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
14'121 |
120.69 |
| |
|
|
| Core
i7-6700K @ Default |
8'760 |
74.84 |
| Core
i7-6700K @ 4.2GHz |
10'432 |
89.16 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
10'920 |
93.33 |
Overall
| |
Points |
% |
| Core
i7-6950X @ Default |
18'272 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
22'016 |
120.49 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
22'163 |
121.29 |
| |
|
|
| Core
i7-6700K @ Default |
17'848 |
97.68 |
| Core
i7-6700K @ 4.2GHz |
19'880 |
108.80 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
20'479 |
112.08 |
[pagebreak]
WinRAR
WinRAR is one of the most common file archiving programs. In our test we use the application’s in-built benchmark utility to determine CPU performance. Performance is measured in KB/s (kilobytes per second), where higher is better.
| |
KB/s |
% |
| Core
i7-6950X @ Default |
30'073 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
33'282 |
110.67 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
34'981 |
116.32 |
| |
|
|
| Core
i7-6700K @ Default |
10'760 |
35.78 |
| Core
i7-6700K @ 4.2GHz |
10'856 |
36.10 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
12'751 |
42.40 |
7-Zip
7-Zip is an open-source file archiving program. We use the application’s built-in benchmark utility to measure CPU performance regarding compression and decompression of files. We test with a dictionary size of 32MB and use all availables threads. The results is displayed in MIPS (millions of instructions per second), where higher is better.
| |
MIPS |
% |
| Core
i7-6950X @ Default |
49'807 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
58'387 |
117.23 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
58'498 |
117.45 |
| |
|
|
| Core
i7-6700K @ Default |
24'989 |
50.17 |
| Core
i7-6700K @ 4.2GHz |
26'013 |
52.23 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
26'883 |
53.97 |
[pagebreak]
Frybench
Frybench is based on fryrender, which is one of the most demanding engines in its category. Fryrender uses every CPU core available without wasting as little CPU cycles as possible. The benchmark has been written to take as much advantage as possible of the new multi-threading processors. The final score is disaplyed in minutes. To keep things easier for comparison we're displaying the results in seconds and lower is better.
| |
Sec |
% |
| Core
i7-6950X @ Default |
109 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
96 |
113.54 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
92 |
118.48 |
| |
|
|
| Core
i7-6700K @ Default |
258 |
42.25 |
| Core
i7-6700K @ 4.2GHz |
234 |
46.58 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
218 |
50.00 |
[pagebreak]
HandBrake
HandBrake is an open-source, GPL-licensed, multiplatform, multithreaded video transcoder, available for MacOS X, Linux and Windows. This software puts multithread CPUs to good use and it allows to transcode videos using different codec, like for instance h.264. For our test we have used a short 4K video file and transcode it with the h.264 and the h.265 codec using the so called high profile and the file is being saved in a MP4 container. The result you see below is average frames per second, where more is better.
h.264
| |
Avg FPS |
% |
| Core
i7-6950X @ Default |
30.13 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
34.83 |
115.60 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
35.14 |
116.63 |
| |
|
|
| Core
i7-6700K @ Default |
16.78 |
55.69 |
| Core
i7-6700K @ 4.2GHz |
17.01 |
56.46 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
17.09 |
56.72 |
h.265
| |
Avg FPS |
% |
| Core
i7-6950X @ Default |
19.13 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
21.81 |
114.01 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
22.02 |
115.11 |
| |
|
|
| Core
i7-6700K @ Default |
11.37 |
59.44 |
| Core
i7-6700K @ 4.2GHz |
11.41 |
59.64 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
11.71 |
61.21 |
[pagebreak]
Photoshop
We guess almost anybody knows Photoshop. In the case of our performance comparison we're using an action script from Keith Simonian, which applies "radial blur" to a 4.8MB test picture. Apart from that Photoshop can measure the time it takes until the operation is complete.
| |
Points |
% |
| Core
i7-6950X @ Default |
7.8 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
6.9 |
113.04 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
6.6 |
118.18 |
| |
|
|
| Core
i7-6700K @ Default |
12.9 |
60.47 |
| Core
i7-6700K @ 4.2GHz |
12.7 |
61.42 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
12.4 |
62.90 |
[pagebreak]
Performance Rating
| |
% |
| Core
i7-6950X @ Default |
100.00 |
| Core
i7-6950X @ 4.2GHz |
114.07 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
117.06 |
| |
|
| Core
i7-6700K @ Default |
65.88 |
| Core
i7-6700K @ 4.2GHz |
70.13 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
72.47 |
[pagebreak]
Power Consumption
Idle
| |
Watt |
% |
| Core
i7-6950X @ Default |
65 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
106 |
163.08 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
123 |
189.23 |
| |
|
|
| Core
i7-6700K @ Default |
55 |
84.62 |
| Core
i7-6700K @ 4.2GHz |
59 |
90.77 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
63 |
96.22 |
Load
| |
Watt |
% |
| Core
i7-6950X @ Default |
180 |
100.00 |
| Core
i7-6950X @ 4.2GHz |
290 |
161.11 |
| Core
i7-6950X @ 4.2 GHz DDR4-3000 |
335 |
186.11 |
| |
|
|
| Core
i7-6700K @ Default |
141 |
78.33 |
| Core
i7-6700K @ 4.2GHz |
147 |
81.67 |
| Core
i7-6700K @ 4.2 GHz DDR4-3000 |
163 |
90.56 |
[pagebreak]
Conclusion
Default
First of all we’re going to have a closer look at performance differences with standard clock speeds. From our overall performance rating we see that the Core i7-6950X is on average
51.79 % quicker than the Core i7-6700K.
If we start searching for the benchmarks with best possible scaling on the
i7-6950X, then we find that WinRAR runs 279.49% faster on this CPU and Frybench benefits by 236.69%. Apart from that it's interesting to see what happens when we have a look at a single threaded benchmark like Blackhole Single-Threaded. In this case the new Core i7-6950X is 33.20% slower because of a lower max Turbo clock compared to the i7-6700K.
We also had a closer look at power consumption and in idle we notice that our test setup with Core i7-6950X was burning
18.18% more power than the system equipped with Core i7-6700K. In case of load power consumption the difference the new Core i7-6950X is 27.67% less efficient.
4.2 GHz
Overclocking the Core i7-6950X increases the average performance by 14.84%, while the i7-6700K benefits by 6.45%.
A closer look at power consumption reveals that overclocking these CPUs make for
quite a difference. With the Core i7-6950X we see system power consumption rise
by no less than 63.07% in idle and 61.11% under load, when comparing with the default power
consumption values. In the case of the Core i7-6700K system power consumption
goes up by 7.27% in idle and 4.26% under full load. This clearly shows that from an efficiency
point of view, overclocking these processors is not attractive.
4.2 GHz DDR4-3000
Having arrived at this point we’re not going to go through every single benchmark anymore since the overall results are still very similar. What’s much more interesting now is checking what performance differences there are when running DDR4 memory at 3000MHz with both processors, while the stock clocks are DDR4-2400.
Overclocking the system with the Core i7-6950X inside we notice that the performance goes up by
2.74%. In the case of the Core i7-6700K the difference is 3.34% on average.
Looking at idle power consumption, the system with Intel Core i7-6950X turns
16.03% more energy into heat, whereas with the i7-6700K consumption goes up by
6.00%. Under full load the differences are 10.89% in the case of the i7-6700K and
15.52% for the i7-6950X.
Recommendation
A quick look at the performance differences between these two CPUs shows, that only select applications benefit from the additional cores/threads the Intel Core i7-6950X offers compared to the i7-6700K. Nevertheless it is interesting to see that Intel's new flagship performaned 65% faster in Photoshop, when we compared the two CPUs at default clock speeds. The difference in rendering applications is even more impressive. Therefore investing in such a costly CPU can make sense if you're using it in a professional environment. If you're saving precious time, which in translates into increased revenue, then deciding for an i7-6950X over an i7-6700K is going to pay off rather sooner than later. In all other cases justifying the i7-6950X's price tag, which is way north of 1600 Euro, is really difficult.
