Help Center> Cloud Container Engine> User Guide> User Guide for Standard and Turbo Clusters> Scheduling> GPU Scheduling> GPU Virtualization> Overview
Updated on 2024-07-02 GMT+08:00
View PDF

Overview

CCE uses xGPU virtualization technologies to dynamically divide the GPU memory and computing power. A single GPU can be virtualized into up to 20 virtual GPU devices. Virtualization is more flexible than static allocation. You can specify the number of GPUs on the basis of stable service running to improve GPU utilization.

Advantages

The GPU virtualization function of CCE has the following advantages:

  • Flexible: The GPU computing ratio and memory size are finely tuned. The allocation granularity for computing is 5% of GPUs, while the memory allocation is in MiB.
  • Isolated: The memory of a single GPU can be isolated, and both GPU computing and memory can be isolated concurrently.
  • Compatible: There's no need to recompile services or replace the CUDA library.

Prerequisites

Item

Supported Version

Cluster version

v1.23.8-r0, v1.25.3-r0, or later

OS

Huawei Cloud EulerOS 2.0

GPU type

T4 and V100

Driver version

470.57.02, 510.47.03, and 535.54.03

Runtime

containerd

Add-on
The following add-ons must be installed in the cluster:

Notes and Constraints

  • A single GPU can be virtualized into a maximum of 20 xGPU devices.
  • After GPU virtualization is used, init containers are not supported.
  • GPU virtualization supports two isolation modes: GPU memory isolation and isolation between GPU memory and computing power. A single GPU can schedule only workloads in the same isolation mode.
  • Autoscaler cannot be used to automatically scale in or out GPU nodes.
  • xGPU isolation does not allow you to request for GPU memory by calling CUDA API cudaMallocManaged(), which is also known as using UVM. For more information, see NVIDIA official documents. Use other methods to request for GPU memory, for example, by calling cudaMalloc().
  • When a containerized application is initializing, the real-time compute monitored by the nvidia-smi may exceed the upper limit of the available compute of the container.
Parent topic: GPU Virtualization