This topic provides answers to frequently asked questions about Elastic GPU Service.

Why does Windows Graphics not support DirectX-based features?

Problem description: On GPU-accelerated Windows instances where the installed GPU drivers have taken effect, Windows Remote Desktop Protocol (RDP) do not support DirectX- or OpenGL-based applications.

Solution: Install the VNC service and client or use other protocols that support these applications, such as PC over IP (PCoIP) and XenDesktop HDX 3D.

Do GPU-accelerated instances support Android emulators?

No, GPU-accelerated instances do not support Android emulators.

Can I change the configurations of GPU-accelerated instances?

You cannot change the configurations of the following GPU-accelerated instances:
  • GPU-accelerated instances with local storage
  • vGPU-accelerated instances

For information about the GPU-accelerated instance families that support instance type changes, see Instance families that support instance type changes.

Do pay-as-you-go GPU-accelerated instances support the No Fees for Stopped Instances (VPC-Connected) feature?

GPU-accelerated instance families such as gn5 that are equipped with local storage do not support the No Fees for Stopped Instances (VPC-Connected) feature. For more information, see No Fees for Stopped Instances (VPC-Connected).

How do I view GPU monitoring data?

You can log on to the CloudMonitor console or call the DescribeMetricList operation to view GPU monitoring data. For more information, see GPU monitoring.

How do I transmit data between GPU-accelerated instances and common Elastic Compute Service (ECS) instances?

In addition to GPU acceleration, GPU-accelerated instances can be managed in the same manner as common ECS instances. GPU-accelerated instances can communicate with common ECS instances that belong to the same security groups as the GPU-accelerated instances over the internal network.

What are the differences between GPU and CPU?

The following table compares GPU and CPU.
Item GPU CPU
Arithmetic logic unit (ALU) A large number of ALUs good for large-scale parallel computing. A small number of powerful ALUs.
Logic control unit (LCU) Simple LCUs. Complex LCUs.
Cache Small amounts of cache that serves threads instead of storing accessed data. Large amounts of cache that can store data to speed up data access and reduce latency.
Response mode Integrates all tasks for batch processing. Responds in real time and delivers fast single-task responses.
Scenario Compute-intensive, high-throughput scenarios where multiple threads run in parallel to process highly similar tasks. Serial processing scenarios that involve complex logic and require high response speeds.

Can instances be upgraded or downgraded from non-GPU-accelerated instance families to GPU-accelerated instance families?

No, instances cannot be upgraded or downgraded from non-GPU-accelerated instance families to GPU-accelerated instance families.

For information about the GPU-accelerated instance families that support instance type changes, see Instance families that support instance type changes.