Task Scheduling
Task scheduling is one of the core capabilities of the ZopToken platform.
The platform assigns suitable AI workloads to available devices based on task requirements, device status, and resource conditions.
Tasks Suitable for Distributed Execution
Distributed devices are better suited for tasks that are decomposable, verifiable, and independently executable, such as:
- Batch inference
- Image or video-related computation
- Rendering tasks
- Embedding generation
- Data processing
- OCR
- Speech processing
- Small model inference
- Automated computation tasks
Actual available task types are subject to the platform's current stage.
Cautious Notes
Not all AI tasks are suitable for distributed device execution.
Tasks requiring low latency, high bandwidth, or strong synchronization — such as large model training or real-time inference — typically demand stricter hardware, network, and scheduling conditions.
Therefore, the current ZopToken documentation uniformly describes tasks as AI workloads, without making excessive promises about specific training, inference, or rendering capabilities.
Factors That Scheduling May Consider
The platform may assign tasks based on the following factors:
- Whether the device is online
- Device hardware resources
- Current device load
- Client version
- Task type
- Task priority
- Network stability
- Historical completion rate
- Provider or device group policies
Task Status
Tasks may exhibit the following statuses:
| Status | Meaning |
|---|---|
| Pending | Awaiting execution |
| Assigned | Assigned to a device |
| Running | Currently executing |
| Completed | Completed |
| Failed | Execution failed |
| Cancelled | Cancelled |
| Retrying | Retrying |
Status naming and display are subject to the current platform version.