Esko Logo Back to Esko Support
Choose your language for a machine translation:
Skip to end of metadata
Go to start of metadata

 

Question

I would like to get some information on the graphic card usage for real-time rendering in Store Visualizer.

Answer

System requirements for Store Visualizer 14: https://wiki/display/SystemRequirements/Store+Visualizer+14+System+Requirements

KB on Stereoscopic rendering and preferred graphic cards: https://wiki.esko.com/pages/viewpage.action?pageId=79692911

Benchmark page on Graphic card performances: http://www.videocardbenchmark.net/high_end_gpus.html

Store Visualizer using only one core for rendering is by design. In fact, it doesn't make sense to use more cores for rendering as it would actually hurt the performance. All the rendering is exclusively done by the graphic card(s). The CPU supplies the graphic cards with the information that it needs to render the scene. The CPU doesn't do any rendering by itself. Since there is only one single 'lane' to send rendering commands to the graphic card(s), having more than one core to push data into it would create collisions and would reduce performance.

It is normal that one core goes to 100% usage when entering 'freelook' mode. We do this on purpose in order to have that core in an 'always ready' state to send data to the GPU when needed. In some very complex projects, with many thousands of products, the CPU core may struggle to send data fast enough to the GPU. However even here multiple cores wouldn't help, they would basically get into each others way. LOD-2 ('Level of Detail' in Visualizer 14.0.2) heavily decreases the amount of data that needs to be sent to the graphics card in the first place.

We currently use the remaining cores for physics simulation and various smaller tasks. They are also used on model import (Windows only). We plan on using multiple cores for project loading in the near future (for reducing the time it takes to open a project).

Adding a second graphics card will not double the available video memory, as all cards need to hold an identical copy of the data. The only way to increase video memory is to use a video card with more onboard memory. Multiple cards will improve rendering performance, but only if they are interconnected in an SLI or Crossfire configuration. Simply adding new cards without interconnection will neither improve memory nor performance, as Store Visualizer won't be able to use it at all (unfortunately this is the case on the new MacPros with 2 GPUs).

Store Visualizer uses OpenGL.

We do not recommend the Quadro K5000 (or K4200 which is slightly better). Check the benchmark page on top for a proper comparison.
It is rather overpriced and has bad performance. If the customer needs a Quadro card (really only needed for non-3D vision stereo), then we recommend the K6000 or M6000. In all other cases, the customer should prefer a GTX 780/980/Titan based system with one or multiple cards (Windows only).

Article information
Applies to

Store Visualizer 14

Created07-May-15
Last revised13-Jan-16
AuthorSDME
CW Number 
Contents