Help Center/ High Performance Computing/ HPC-S² User Guide/ Overview/ What's HPC-S²?
Updated on 2024-05-20 GMT+08:00
View PDF
Share

What's HPC-S²?

High-performance computing (HPC) is the practice of aggregating a large number of compute nodes to work concurrently to solve large-scale computing problems. It has been widely used in scientific research, weather forecast, simulation experiment, biopharmaceuticals, gene sequencing, and image processing.

The requirements for high-performance computing have been increasingly growing. HUAWEI CLOUD launched the HPC Solution Service (HPC-S²) to provide highly scalable high-performance computing resources as a cloud service. With basic services and software pre-integrated, HPC-S² allows you to provision and deploy HPC resources with just a few clicks. HPC-S² can automatically deploy job scheduling software such as Simple Linux Utility for Resource Management (Slurm), Sun Grid Engine (SGE), and Huawei commercial HPC Donau Scheduler to facilitate task assignments on parallel computers. On the HPC-S² console, you can view and manage clusters, nodes, and shared storage in an intuitive way.

You can also use the HPC-S² console to manage multiple clusters in different VPCs, as shown in Figure 1.

Figure 1 HPC-S² architecture

HPC-S² Functions

The HPC-S² is mainly used by enterprises that require a large amount of compute resources. These enterprises can use the HPC-S² console to easily manage clusters, nodes, tasks, and shared storage, provision clusters, and deploy applications.

Function

Description

Cluster management

Create HPC-S² clusters; view cluster creation, deployment, scale-in/out progress; view details about clusters; add, start, stop, and remove clusters; and manage instance topology.

Node management

Query, start, stop, restart, and delete nodes in HPC-S² clusters.

Shared storage management

Provide storage for HPC-S² clusters by using Scalable File Service (SFS).

Task management

Manage HPC-S² cluster-related job queueing, dynamic scheduling (including GPU scheduling), bin packing and workload consolidation, and quotas for CPU, memory, and instances.

HPC-S² Advantages

Compared with traditional HPC, HPC-S² offers the following advantages:

  • HPC resources are provisioned on demand and are available instantly upon provisioning.
  • Multiple types of cloud servers are available for you to choose the most appropriate compute servers for your workloads.
  • HPC-S² can work with security services, such as Anti-DDoS, Web Application Firewall (WAF), and Vulnerability Scan Service (VSS), to provide multi-dimensional protection to improve data security and availability.
  • With HPC-S², you do not need to build your own equipment rooms or worry about hardware upgrade or replacement. You can obtain resources for exclusive use from virtual resource pools anytime and these resources can be elastically scaled based on service changes.
  • HPC-S² can automatically configure password-free login, network information services (NISs), and schedulers for users.
Parent topic: Overview