Some useful information regarding KeyShot Network rendering speeds. 


Prerequisites


  • Network Rendering



Procedure


How does it work?
KeyShot is 100% CPU-powered and it will use 100% of your CPU’s processing power. If you double the number of cores, you potentially cut your render time in half.* KeyShot Network Rendering is designed to divide each render job into sections or regions. Imagine that the Manager is putting your image through a paper shredder. Regions are horizontal strips W pixels wide where W is your image width in pixels. The number of regions depends on the size of the image/frame.


Here is the typical order of operations involved when rendering over the network:
1. The manager receives a job from the client##netw
2. The manager breaks up the image/frame into regions.
3. The manager assigns two regions to each available Worker (regardless of how many cores has the Worker or how large is the image).
4. Worker renders the two regions and sends them back to the Manager once these two regions are done.
5. The worker repeats steps 3 and 4 until the entire image is rendered and until all frames in an animation are done.


A couple of things to keep in mind:
1. The real-time render engine inside KeyShot (not Network Rendering) is designed to use 100% of your CPU power 100% of the time.
2. KeyShot Network rendering is designed to use 100% of your CPU power but Workers are not rendering 100% of the time. A small percentage of that time is dedicated to transferring data to and from the Manager. This “downtime” percentage is negligible.
3. Network Rendering was designed to connect multiple machines in an office network and combine their processing power.


Note: Improvements in render speed vary. For some jobs, render time cannot be lowered regardless of how many cores one has. A rendering will always take at least a few minutes/seconds to render. Hence it is better to test Network Rendering performance with renderings that take a few hours.