Elastic High Performance Computing (E-HPC) is a computing service provided by Alibaba Cloud. E-HPC delivers excellent and stable computing performance that is achieved by auto scaling. E-HPC aggregates computing capabilities and uses parallel computing to address major challenges in science, engineering, and commerce. E-HPC is used in a variety of fields, including scientific research, petroleum exploration, finance, weather forecast, pharmaceuticals, gene sequencing, and graphics processing.
Architecture
The following table describes the relationships between E-HPC and other services.
Alibaba Cloud service | Description | References |
Elastic Compute Service (ECS) | Each E-HPC cluster consists of multiple ECS instances. These instances serve as logon nodes, management nodes, and compute nodes to provide different features.
A security group is a virtual firewall that provides Stateful Packet Inspection (SPI) and packet filtering capabilities. It can be used to isolate security domains on the cloud. | |
Virtual Private Cloud (VPC) | A VPC is a private network on the cloud. VPCs are logically isolated from each other. You can create and manage clusters in a VPC. | |
Elastic IP address (EIP) | By default, the logon node in a cluster is associated with an EIP, and is allocated a specific amount of public bandwidth to access the Internet. | |
File Storage NAS (NAS) | NAS provides a path to shared files for all nodes in a cluster. | |
Object Storage Service (OSS) | You can use an OSS bucket to store your job data file and executable file. You can also import these files from an OSS bucket to a cluster. |
Features
The following table describes the features of E-HPC.
Feature | Description | References |
Computing resource creation | Before you use E-HPC, you must create a cluster. Then, you can submit jobs, run jobs, and view job results in the cluster. | |
Computing resource configuration | E-HPC allows you to configure multiple computing resources, including users, nodes, scheduler queues and auto scaling settings. | |
Job submission | You can submit jobs to a cluster by using the E-HPC console. If you need to submit a small job, you can modify and submit the job script in the console. If you need to submit a complicated job, you can upload its job script to an OSS bucket and import the script to the console. You can also use an SSH client to log on to the cluster and submit jobs. | |
Performance analysis | The E-HPC optimizer is a performance monitoring and profiling platform. You can monitor cluster performance metrics, view cluster performance history, and profile performance hotspots. | |
Virtual Network Computing (VNC) | You can use the VNC service to manage jobs in the E-HPC console. |