Jaa


Overview of HPC Job Manager

Microsoft HPC Pack provides an integrated application platform for running, managing, and developing parallel computing applications. HPC Job Manager is your primary interface for submitting and monitoring jobs on a cluster.

Using HPC Job Manager, you can submit, monitor, and manage all of your compute jobs from a single location.

Note

You can also use a command prompt window or HPC PowerShell to submit, manage, and view jobs on a cluster.

In this topic:

Basic HPC job terms

Term Definition
Job A resource request that is submitted to the HPC Job Scheduler Service that contains, or will contain one or more tasks.
Task The execution of a user application. The task Type property helps define how many times to run the command. For more information, see Understanding Task Types - Job Manager.

A task cannot be run independently of a job. Each job has one or more tasks that must run before the job is complete. To have tasks run in a specific order, you can define dependencies between the tasks. For more information about defining dependencies, see Define Task Dependencies - Job Manager.
Job Template A job template is required for submitting jobs. Job templates are created by the cluster administrator for different types of jobs. Job templates define default values and constraints for job properties and assist users in creating jobs that meet their needs and conform to resource allocation policies. For more information, see Understanding Job Templates - Job Manager.
Job or Task Description File A user-named XML file that contains job or task specifications. This file allows you to preserve a job or task as a pattern for future submissions. When you create a new job or task from a description file, you can modify any of the properties before submission. For more information, see Save a Job or Task to a File - Job Manager.
Queue A list of jobs that have been submitted to the Job Scheduler service to be run by the cluster. A set of scheduling policies determines the order in which to run jobs from the queue. Cluster administrators configure the scheduling policies.
HPC Job Scheduler Service A service that runs on the head node and is responsible for validating your credentials and your job parameters, queuing jobs and tasks, allocating resources, dispatching tasks to compute nodes, and monitoring the status of jobs, tasks, and nodes. In this guide, this service is also referred to as the “job scheduler”.

Basic HPC cluster resource terms

Term Definition
Cluster The top-level organizational unit of an HPC cluster, comprising a set of nodes, a queue, and jobs. The cluster is referred to by the name of the head node.
Head Node A server that coordinates and monitors the cluster. The head node coordinates job scheduling through the HPC Job Scheduler Service.

You submit jobs to the head node. The head node validates your credentials and your job parameters and adds your job to the queue.
Node A single, named host in the cluster.
Node Group A collection of nodes. There are default node groups that you can use to run jobs, including compute nodes and workstation nodes. Depending on your version of HPC Pack, there might be additional default node groups that can run jobs such as Windows Azure nodes and unmanaged server nodes.

Cluster administrators can create additional custom groups and assign nodes to one or more groups. For example, nodes are often grouped according to network proximity, memory, specific hardware, or installed applications.

Job templates can determine what node groups you can use for specific types of job.
Core An individual CPU on a node. For example, a dual-core processor is considered two cores.
Socket In terms of the HPC Job Scheduler Service, a socket is a set of processors with a dedicated memory bus. This is also known as a non-uniform memory access (NUMA) node.

Opening HPC Job Manager and connecting to a cluster

To submit jobs to a Windows HPC cluster:

  • You must have the client utilities in HPC Pack installed on your computer. The client utilities include HPC Job Manager.

  • To connect to a cluster, your computer must be joined to the same domain or to a domain that has a trust relationship with the cluster domain.

  • You must have been added as a user on the cluster, or you must be a member of a domain group that has been added as a user on the cluster. For more information, contact your cluster administrator.

To open HPC Job Manager in Windows HPC Server 2008 R2 and earlier versions

  1. Click Start, click All Programs, click Microsoft HPC Pack, and then click HPC Job Manager.

    The first time you open HPC Job Manager, the Select Head Node dialog box appears. Click the head node of the HPC cluster that you want to connect to, and then click OK.

  2. If you want to connect to a different cluster, in the File menu, click Select Head Node, then click the head node of the HPC cluster that you want to connect to.

To open HPC Job Manager in Windows Server 2012

  1. Open the Start screen using one of these methods:

    • Press the Windows logo key. In a virtual machine, you can press Ctrl+Esc.

    • Hover the mouse cursor in the upper right corner of the screen, and then click Start.

    • On the desktop, hover the mouse cursor in the lower left corner of the screen, and click when the thumbnail of the Start screen appears.

  2. Right-click the Start screen and click All apps in the bottom right corner.

  3. Navigate to HPC Job Manager.

    The first time you open HPC Job Manager, the Select Head Node dialog box appears. Click the head node of the HPC cluster that you want to connect to, and then click OK.

  4. If you want to connect to a different cluster, in the File menu, click Select Head Node, then click the head node of the HPC cluster that you want to connect to.

Additional references