The new Cycles CPU 2.79 Benchmark

Hi all, here is the new Cycles CPU benchmark for Blender 2.79

Please use official release of Blender 2.79 only (NOT Buildbot, NOT Graphicall)

Open file, hit F12,write down:

OS
CPU model @stock frequency
Render time

Please do not use your machine until rendering is complete.
Please use orginal Blender stamp format xx:xx.xx
Please don´t upload images, render time and specs only, as stamped.

CPU_benchmark_279_denoise_disney.blend (3.05 MB)

It is better you check the frequency your CPU is working. But If you do not know the frequency your CPU is running or if it is running overclocked with some profile on the motherboard, you simply indicate the CPU model. This will be taken as running at stock frequency.
If you want you can overclock and publish render time and stating that you are overclocking and at what freq it is running (e.g, @4.2GHz OC). You should keep in mind that this benchmark should be useful for other users to decide the hardware to buy. So please do not overclock too much, use feasible overclocking to production render under long hours of hard work.

Many thanks to mib2berlin for review and improvements to the file, and for the support.

====
Results (This may not be frequently updated)
Click to enlarge (it may be necessary to click again when the image is opened)

If you think that I have made some mistake in the spreadsheet, please let me know.

*Personal Note: If you are thinking of a machine intended only for CPU rendering with Cycles, you should surely pay attention to these results. But you should keep in mind that overall performance of many tasks in Blender depend on single thread performance. So please, you also keep in mind this if you are planning to work with Blender on that computer:
https://www.cpubenchmark.net/singleThread.html

Kubuntu 14.04
intel i7-3770 @stock freq
05:47.75

Opensuse Tumbleweed
Intel i5 3570K
08:48.38

Thanks YAFU!

Windows 10 PRO x64
i7 980x @stock
07:03.97

Antergos
e5-2683V3
02:10.87

03:57:89

i7 5820k @ 4GHz

win 10 64bit

Sabayon linux, 4.12.0
Intel, [email protected] GHz, stock.
08:25:92

Win7 Pro
Intel i7 [email protected]
10:50:34 :wink:

Ubuntu 16.04
Intel i7 4930K OC @ 4.4 GHz
04:59.99

Windows 10 pro x64
AMD ryzen 7 1700 @stock frequency
03:14:30

Win 7 Pro 64bit
Dual Intel Xeon E5-2690 v4 @ 2.60 Ghz (stock)
0:56.28

Windows 10 Pro 64bit
Core-i7 5960X @ 4Ghz
03:03:26

Ok, two dumb questions. What does this info do? Is this the same .blend file as the “GPU” thread.

It is the same configuration except for the tile size configured in 16x16 for CPU.
Just split into 2 threads for clearer results and make the task a little simpler for mib2berlin and me.

So you can also compare results between spreadsheets of the GPU in the other thread and CPU of this thread.

Edit:
Tiles sizes can be somewhat trivial for experienced users, but in many cases users may confuse settings. So better to just share different files with tile sizes correctly configured for each device so that the user only has to press F12 to render.

Windows 10 Home 64bit
Intel Xeon E3-1240 @3,4GHz
05:47.92

Windows 10 Pro x 64
i7-6700HQ
08:22.23

windows7
2x intel X5650 / 24 threads @ 2.67 GHz

05:14.46 @ 16x16
05:14.91 @ 32x32
05:29.71 @ 64x64

Windows 10 64 Family
Intel [email protected] (stock)
05:47.91

Have several machines knocking about so tried the following, best to worst:

W10 Pro
2 x Intel Xeon X5670 @2.93 GHz (stock)
04:09.29

W7 Pro
i7 3770 @ 3.40 GHz (stock)
07:08.85

W10 Pro
i7 3740QM @ 2.70 GHz (stock)
07:49.98

W10 Home
i7 920 @ 4GHz (overclocked)
09:11.47

W10 Home
i7 920 @ 2.67 GHz (same machine but stock)
13:08.14

Ubuntu 17.04
Dual Xeon E5 V4 ES 18 cores @ 2.2 GHz
00:53.65

Using optimal tile settings 16x18:
00:53.51