Request: Network Rendering How To For Beginners

[thickly]

A little background first....I am working for a large clothing retailer as a store designer and before I came along, all design and design viz was being done by consulting architects. I started getting renderings in front of the boss and now we have moved all design work in-house and I'm pretty much the dedicated viz guy.

 

That being said, I now have a great machine to work on but every day I walk past our server room and see racks and racks and racks and racks of servers (you get the point) that I just dream of being able to use all of that processing power on my renders - after all I am using Maxwell and we know how long those renders can take. I have approached IT about the idea of network rendering and they didn't say no, but if I was able to present them with some direction on how to set it up -- it might be what I need to get them to invest the time in setting it up. Does anyone know of a good resource that explains the theory/process, etc... The more leg work I do the easier it will be to get them to help me out.

 

I know I'll need additional licenses, but I can get my boss to approve that cost -- the tough part is getting another department on board to do the actual work of setting everything up. God bless huge beauracratic corporations.

 

I'm also planning on adding Vray to my skill set and with that doing some animations, so even though I know of the issues with Maxwell's network rendering, I still think I could benefit from have network capability set up now.

 

Thanks in advance for any advice.

SD

[Devin Johnston]

Sean,

Setting up network rendering for Maxwell isn't that difficult, what you need for your initial setup is this.

 

- a dedicated manager machine that isn't running anything else, the reason is stability and the more programs this machine is running in conjunction with the Maxwell manager the more unstable it becomes.

 

-You need to have an accessible server that all your render nodes can see where all of your maps are located. Unlike 3D Studio's network rendering where you can include maps, Maxwell must have them accessible to all nodes at all times. You will also need a place to store files that is also accessible by all nodes on the network. All of these machines need to have full read and write access to the server.

 

-distributed rendering won't work with more than 10 machines at a time so you will have to keep that in mind.

 

The problem you are going to have from what you described is that all of the servers in the server room are busy doing other tasks. Maxwell is a resource hog even in low priority mode it must have enough memory to open your file and the minimum amount of memory recommended is 2GB. Once you start rendering you are going to put a serious dent into the processing power of those servers and if they are critical to your company’s daily operations it's going to be very noticeable how slow they have become to user requests. The best solution is of course to purchase dedicated render nodes, memory is very important and so is the processor, you don't need a video card at all and a decent hard drive will due fine.