All in one cluster management software
ClusterEngine is the cluster management software, developed by Inspur, provides the comprehensive management service for HPC cluster, with its functions including hardware monitoring, job scheduling, HPC application management, and user account maintenance. ClusterEngine delivers the visual mode for users, to maximise system resource utilization.
Four functional models management
Hardware monitoring, job scheduling, HPC application management, and user account management.
Hardware monitoring management
- Smart workload: dynamically retrieve and release computing resources, based on the job queue and workload status.
- Cluster monitoring: cluster utilization status of CPU, MEM and network.
- Alarm system: SMS, Email, voice and web alarm, with alarm threshold setting.
- Heterogeneous nodes: status of GPU and MIC accelerators, including temperature, power, utilization of GPU & MEM.
Job scheduling management
- Job submission: upload, download, create, open, copy, rename, compress, delete.
- Job scheduling: priority policy, sharing policy, resource constraints policy, resource backfill policy.
- Breakpoint recovery: breakpoint auto setting, breakpoint saving, recovery and migration. Minimizing side effect of accident.
HPC application management
- Application integration: Fully integrated with VASP, FLUENT,CFX,ANSYS,GAUSSION and more than 20 scientific applications
- Application templates: workflow of job submission, job queue, job results and job analysis, after simple parameters setting up.
User account management
- User Management: creating, deleting users, user authority setting.
- Authority permission: different users with different permissions, including job number, maximum number of nodes and processors, and accessible queues and nodes.
- Expense statistics: different pricing for different resources, with balance check and payment alarm.
- Resources statistics: resource utilization record, job statistics, and .XLS and PDF reports.
- Friendly interface: the real layout view of cluster, shell & remote login, cluster status monitoring.
- Workload visualization: job running monitoring, pre and post job status checking, including convergence graphics, energy-band graphics, and density graphics.
- Customization: application templates customized, interface customized, and Logo customized.
Converging with big data
With the trend of HPC and big data converging for data computing and platform sharing, ClusterEngine enable to manage HPC and Hadoop resource accordingly, dynamically switch between hpc node and Hadoop node, scheduling both HPC and Hadoop jobs. Inspur HPC is ready for BigData era.
ClusterEngine provides the comprehensive and efficient management of HPC cluster with full visual mode, to maximise system resource utilization.