More and more 3D artists are approaching render farm rental services to serve their workflow. Renting a render farm is very simple and saves them a lot of time. So is renting a render farm expensive, and which render farm is the cheapest?
Why should you use a render farm?
Why do we need to hire a rendering farm service? The question seems very simple, but many people want to find the answer. And below, we give three reasons:
- Render Farm will help you complete difficult and heavy projects that your personal computer cannot handle.
For instance, Weta Digital utilized a 10,000-square-foot server farm to render Avatar. There are 4,000 servers and 35,000 CPU cores altogether. A single frame typically takes two hours to generate, which is equivalent to 160 minutes of video. 2880000 hours are required for the entire rendering process, which is the same as 328 years of work for one server. It will take a long time to complete a movie if you don’t employ render farm support because this is a really large project. And no home computer has the power to accomplish this.
-
- Reduce costs and render time
The same project can take you days or even months to handle. But when using the render farm service, it only takes a few hours or a few days. Because a “render farm” is a high-performance computer system of interconnected computers that will allocate and process your work many times faster than using a personal computer.
-
- Large storage memory
Each renders farm has a large data storage memory ranging from 64 GB to 256 GB. This means you can load multiple projects with multiple frames without worrying about running out of space.
How does calculator price rent render farm service?
Each render farm has a different pricing method. Some common price units are GHzH, Obh, core hour and node hour.
Explain the meaning of each unit
GHz Or GHzH
GHzH hay GHz hour stands for gigahertz per hour, which is how CPU rendering services are charged.GHz hour is the price for 1 hour of rendering at 1 GHz.
Render cost = Number of cores x Clock speed (GHz) x Price per GHz hour x Render time
You use the service specifically with CPU information. The 22-core Intel Xeon E5-2699 v4 processor has a clock speed of 2.20 GHz. An Intel Xeon E5-2699 v4 server in a render farm costs $0.015 per GHz per hour. Therefore, the price for a project to render for an hour on this particular machine is 22 x 2.2 x 0.015 x 1 = $0.726. An Intel Xeon E5-2699 v4 node costs $0.726, and we’ll examine this more in the “node hour” section.
This is also the common pricing method for Ranch Computing, Garage Farm, and Rebus Farm.
Core hour
Core Hour is also calculated for the CPU rendering service, like GHz, but instead of us using calculations on the clock of each core, the Core Hour calculation is simpler:
Render cost = Number of cores x Price per core hour x Render time
For example, if you use a render farm service at $0.06 per core hour with 16 core CPU nodes, then your 1-hour usage price will be 0.06*16*1 = $0.96.
Similarly, the price will differ if you use a package of 24 or 36-core CPU nodes.
OBh
OBh is the way to the time when you use GPU rendering services. This is an acronym for OTOY’s OctaneBench. OctaneBench is able to define the total graphics processing power of a computer. Once it is launched, it produces a “score” for the system being benchmarked, called Octanebench (OB) points. OBh is a simple unit of GPU cost. Here’s how to calculate OBH hours:
Render cost = Octanebench score x Price per Obh x Render time
For example, suppose you have an RTX 3090 with an Octanebench score of around 652 and a render farm that charges around $0.009 OBh. The bill for 1 hour of rendering on this RTX 3090 node will be 652 x 0.009 x 1 = $5,868.
The OBh benchmark is very popular, so many render farms use this calculation to calculate their GPU rendering services.
Node hour
Node hour calculates rendering costs based on the entire node. The number of nodes is calculated based on OctaneBench for GPU, and for CPU, we use the CPU Rendering Benchmark (Cinebench).
Render cost will be calculated according to the following formula:
Render cost = Price per node hour x Render time.
A node with an Intel Xeon E5-2699 v4 processor costs $0.726 per node hour.
An RTX 3090 node costs $5,868/node hour.
This node-hour calculation is used in render farms that can recognize the specific type of machine you’re using, such as clear information about the CPU and GPU card, and you can thoroughly examine the machine information.
Which is the cheapest render farm?
If you are looking for a render farm that is suitable for your project and has a low cost, you can check out some of these render farm prices. This is the cheapest render farm available today.
iRender’s IaaS service supports all 3D software and render engines with a massive number of nodes ranging from the RTX 4090 to the AMD RyzenTM ThreadripperTM PRO 5955WX @ 4.0-4.5GHz. With this configuration, you can handle all projects, including high-volume projects, in a short time.
Give a 20% discount to a newly registered customer.
Garage Farm is a render farm that operates on the SaaS model. They bill OBh for GPU rendering and Per GHz Hour for CPU rendering. Both GPU and CPU rendering are divided into three basic packages: low priority, medium priority, and high priority. With prices like these, Garage is also one of the cheapest render farm.
Give a 25% discount to a newly registered customer.
Fox uses Per node per hour and core hour as the cost units for their rendering services. They also have four different fees for the four deposit levels.
Give a 20% discount to a newly registered customer.
You can refer to their price list below:
Conclusion
Through the article, you can also understand the ways to calculate the cost when renting a render farm and the top 3 render farms with strong configurations at the best cost for you to choose from. With IaaS render farms, you can comfortably use the service without worrying about software and support versions. But with SaaS render farms, you should pay attention to the software, render engines, and plugins they support on the system.
Leave feedback about this