computer for visualization
, ,

Recommendations for choosing a computer for 3d Rendering in 2017

CPU selection for 3D visualization

Multi-core processor: the more cores, the better. From Intel processors, it is a family of Core i7:

  I7 7700K I7 6800K I7 6850K I7 6900K I7 6950X
Technical processor 14 nm 14 nm 14 nm 14 nm 14 nm
Cores 4 6 6 8 10
Streams 8 12 12 16 20
Frequency 4.2 3.4 3.6 3.2 3.0
Turbo 4.5 3.6 3.8 3.7 3.5
Cache 8 MB 15 MB 15 MB 20 MB 25 MB
Powerfulness, in W 91 140 140 140 140
Cinebench R15 multi 988 1132 1154 1486 2041
           

From AMD processors, it is Ryzen 5 and 7:

  1500X 1600X 1700 1700X 1800X
Technical processor 14 nm 14 nm 14 nm 14 nm 14 nm
Cores 4 6 8 8 8
Streams 8 12 16 16 16
Frequency 3.5 3.6 3.0 3.4 3.6
Turbo 3.7 4.0 3.7 3.8 4.0
Cache 8 MB 15 MB 15 MB 20 MB 25 MB
Powerfulness, in W 65 95 65 95 95
Cinebench R15 multi  — 1136 1419 1540 1628
           

CPU for 3D visualization

The comparison shows that the AMD processor for the same money is faster in the rendering in 1.5 times! This does not mean that the Ryzen 7 processors are the best in everything. In single-streamed tasks, they give the same performance or worse at the same frequencies. But in 3D visualization, in the parameter of combining performance on one price unit, they go far ahead.

Video card for rendering

In some cases, rendering using the video card works much faster than the one done by the CPU. GPU rendering supports iRay, V-Ray RT, Cycles in Blender, and other visualizers.

Here, the two main competitors are AMD and NVIDIA again. The main difference is in the supported technologies as NVIDIA video cards use CUDA, while AMD video cards – OpenGL. So it always makes sense to find out in advance what exactly you will use. For example, VRay help section tells about the choice clearly: “It is highly recommended to use the CUDA engine on nVidia GPUs.

I have the impression that there are fewer problems with NVIDIA video cards, so I propose to choose among them. The main criterion, besides the price, of course, is the number of CUDA cores – the more the better. There is no sense to overpay for a professional card for rendering – you will not get a real gain in the speed of rendering but you will lose a lot of money.

GEFORCE GTX 1050TI / 1050 1060
6 GB / 3 GB
1070 1080 1080TI
CUDA of the core 768/640 1280/1152 1920 2560 3584
Frequency base 1290/1354 1506 1506 1607 1480
Frequency turbo 1392/1455 1708 1683 1733 1582
Memory 4 GB GDDR5 / 2 GB GDDR5 6 GB GDDR5 / 3 GB GDDR5 8 GB GDDR5 8 GB GDDR5X 11 GB GDDR5X
Memory base 128-Bit 192-Bit 256-Bit 256-Bit 352-Bit
The pass. capability of the memory (Gb/s) 112 192 256 320 484
Consumed power, W 75 W 120 W 150 W 180 W 250 W
           

Video card for rendering

RAM

The standard memory is DDR4, the bigger volume the better. 8 Gb is the absolute minimum, while 16 Gb is optimally for the beginning. Further quality is up to taste and purse. In severe cases, it makes sense to add up to 32 Gb or more.

Hard Drive and SSD

They do not directly affect the rendering speed. I propose the following options:

  • HDD – SATA-III, 7200, cash – the more the better; volume is 1 Tb.
  • SSD – volume is 256 Gb and more. The speed of work is clearly defined by the capabilities of your purse.

Motherboard

It is selected strictly under the processor taking into account the dimensions of the case, the necessary number of any connectors like USB, sATA, for memory modules and so on. It is difficult to give any specific recommendations in advance.

Vrender Company 2017.