Hardware resources are needed for many Yonohub actions, such as building an environment, running a job, running a YonoArc pipeline, etc. In order to meet the requirements of different workloads, Yonohub offers four resource models with different CPU, memory, SSD, and GPU resources as follows:
Resource Model | CPU Cores | Memory (GB) | SSD (GiB) | GPU |
C1 | 1 | 4 | 30 | – |
C2 | 2 | 8 | 50 | – |
C8 | 8 | 32 | 200 | – |
GK1 | 6 | 56 | 340 | Tesla K80 12 GB Memory |
Whenever Yonohub needs to allocate hardware resources for an action, Yonohub allows you to select the resource model you wish. It is important to note that the resource model you select for a particular action does not limit your options for future actions. For example, you can build an environment using C1, but later use the same environment to start a custom app using C2.
Pricing information is detailed here. With Yonohub, you only pay as long as you are using the resource. Yonohub uses Per-Second Billing with a minimum of two minutes for each action.
You can track your current balance and your transactions as follows:
|
You can always use Express Build/Launch to allocate your resource model faster, but this incurs an additional 30% cost compared to the corresponding non-express resource model. If Yonohub wasn’t able to allocate your resource model fast enough, no additional cost will be incurred. Non-express build/launch are hidden by default, but you can show them by checking Enable normal launch from your Account Settings.