Setting Up Your AWS Account
Nimbus allows you to render your Blender projects on your own render farm that lives in Amazon’s computing cloud. In order to use Nimbus, you must set up an Amazon AWS account. Note that this is not your usual Amazon account that you use to buy books and gadgets.
Nimbus will help you set up your account, but first we want to make sure you’re aware of a few things. Once you start rendering in the cloud, Amazon will charge you for computing time and data storage and transfer. The price is typically measured in cents per hour, cents per GB, or cents per GB-month, but you should consult the Amazon pricing pages if you have any questions.
You can view your bill at any time by logging into your AWS account and going to My Billing Dashboard.
When you first start Nimbus, you’ll be greeted by a wizard that will help you set up your AWS account so that Nimbus can use it. If you close out of the Wizard and would like to start it again, you can always get back to it by selecting File → Setup Account…
Subscribing to a Nimbus Rendering Machine
In order to render your projects, you’ll need to choose a rendering machine. To do this, you’ll need to subscribe via the AWS Marketplace. Go to the AWS Marketplace and search for “Nimbus”. Select “Nimbus Render” and subscribe. Take note of the pricing and accept the software terms. There is no need to “Launch with the EC2 Console”.
Once you’ve subscribed and accepted the terms, take note of the AMI ID (AMI-XXXXXXXX).
Preparing Your Blender File
Before submitting your .blend file to Nimbus (or any render farm, really), you must make sure it’s as self-contained as possible. If your project references image files that reside at some random location on your hard drive, then render farm won’t be able to access them. So first you need to tell Blender to pack all files into the .blend. In Blender, go to File → External Data → Pack All Into .blend. There, that’s it!
Submitting Your Project
Now add all your project files to Nimbus either by dragging and dropping them, or by selecting Submit → Add Files…. For a simple project, you’ll just have your .blend file. But if you’ve baked particles into your .blend, for instance, then you’ll have a directory with a name like blendcache_MyProject. Be sure to add that directory as well.
Now fill in the fields:
|Project Name||Give your Nimbus project a name.|
|Frames to Render||Tell Nimbus the starting and ending frames to render.|
|Step||Usually set to 1. If you’d like Blender to render every other frame, change this to 2 (and change “Per Task” to 2 as well).|
|Per Task||Usually set to 1. If you’d like each task to handle more than one frame, change this value. If you’re skipping frames, also change this value.|
|Image||This is the Amazon Machine Image that your render farm will use. Each image has a different version of Blender installed. Currently, we only have one image available.|
|Scene||Set this to the scene in your .blend file you would like to render. If you leave it blank, Blender will render whichever scene your .blend file was set to when saved.|
|Frame Name||This controls the names of the output files that Blender generates. “######” will be replaced with the frame number, so if this field is set to “frame_######”, then the first frame will be called “frame_000001.png”.|
|Instance Type||Amazon has many different types of machines available for rendering. Machines with low memory, high memory, slow CPUs, fast CPUs. Amazon will charge you based on which instance type you decide to use. See here for a list of prices.|
|Num.||This controls how many machines you have in your farm running your tasks in parallel. In general, 10 machines will run 10 times as fast as one machine. Remember that Amazon charges for each fraction of an hour, so if you start 100 instances, and your render project finishes in one minute, you’ll still be charged for 100 hours of computing time.
If you’re not sure your project has been set up properly, you can set this to 1 and wait for the output of a single frame to make sure you’re getting what you expect. Then later we’ll show you how to add more instances to a project that’s already running.
Once you’ve completed the fields, just click Submit and your job will be submitted to your render farm. Depending on how large your file is, it might take a few minutes to finish submitting.
Monitoring Your Render Tasks
Once your render tasks are submitted, you can monitor their progress on the Monitor tab. In the lower right corner, click refresh to see their current status.
In the Completed Frames box, you can see the frames that have finished rendering so far. Right click on one to download and view it.
In the Instances box, you can see the status of all the instances you’ve launched. You can right click on one to view its log file and confirm that Blender is running properly on it. You can also stop the instance if you desire. Note that once stopped, you cannot start it rendering again.
Near the bottom, you can launch new instances by entering the number to launch and then clicking on Add. The new instances will be of the same type as your original instances.
Once all your jobs have finished, you can go to the Finish tab. Click Download to download all your rendered frames at once.
Finally, click the Clean Up button. This will clean up all the AWS cloud resources you’ve used, including removing all rendered frames from the cloud. You don’t want to clean up until after you’ve downloaded all your frames. But it is important to clean up when you’re done, or Amazon will continue to charge you for cloud storage.