Altair Control™ User Guide

2021.1.0

Home
Configure the Workload Manager
Configure WLM server and scheduler settings, queues, resources, defaults and limits.
Manage Scheduling Parameters
Make changes to settings to modify the behavior of the Workload Manager Scheduler.

What's New
Get Started
Tutorials
Establish Access Controls
Manage Clusters
Configure the Workload Manager
Monitor and Manage an HPC Cluster
Configure and Manage Allocations
Run a Simulation
Analyze HPC Resource Usage
Configure and Manage Cloud Bursting
Node Management
Real-time Dashboard
Log out of Control

2021.1.0

Altair Control™ User Guide

What's New
New features available with Altair Control™.
Get Started
Use Control to monitor and manage your HPC clusters.
Tutorials
Learn more about Control by watching a product video.
Establish Access Controls
Add users and assign them roles to grant and restrict access to the various features of Control.
Manage Clusters
Add, update, or delete HPC clusters.
Configure the Workload Manager
Configure WLM server and scheduler settings, queues, resources, defaults and limits.
- View the Current Workload Manager Configuration
  View the current Workload Manager server and scheduler settings, queues, resources, defaults and limits.
- Manage Workload Manager Server Settings
  Make changes to settings to modify the behavior of the Workload Manager Server.
- Manage Queues
  Add, change, and delete execution and routing queues.
- Manage Custom Resources
  Add, change or delete custom resources.
- Manage Server Defaults
  Specify the default execution queue and default arguments that are added to any job submission or deletion request.
- Manage Resource Defaults
  Add, edit or delete resource defaults.
- Manage Server Limits
  Set Workload Manager Server limits for maximum subjobs per job, maximum size of a job script, and maximum number of vnodes that can be provisioned.
- Manage Resource Usage Limits
  Add, edit or delete resource usage limits.
- Manage Server Security Settings
  Grant PBS Manager and Operator privileges, specify whether users can query the status of jobs owned by another user, and specify how usernames are treated across all machines in a HPC cluster.
- Manage Scheduling Parameters
  Make changes to settings to modify the behavior of the Workload Manager Scheduler.
  - Implement Backfilling
    Implement backfilling to keep resources from becoming idle when the top job cannot run.
  - Implement Starving Jobs
    Track the amount of time a job has been waiting to run and then mark the job as starving if this time has passed a specified limit.
  - Implement Placement Sets
    Use placement sets to improve task placement by exposing information on system configuration and topology.
  - Adjust the Scheduling Cycle Iteration
    Adjust the time between scheduling iterations.
  - Use a Formula for Computing Job Execution Priority
    Provide a formula that is used to calculate job execution priority.
  - Enable Scheduling
    Enable or disable scheduling.
  - Implement Fairshare
    Implement fairshare to ensure that jobs are run in the order of how deserving they are.
  - Configure Fuzzy Backfilling
    Trade-off between scheduling cycle speed and the fineness of the granularity with which estimated start times are calculated.
  - Improve Throughput of Jobs
    Improve job throughput by allowing the scheduler to run asynchronously.
  - Specify the Prime and Non-Prime Time Queue Prefix
    Specify the prefix that indicates that a queue is a prime time or non-prime time queue.
  - Specify Resources to Schedule On
    Specify those resources which are not to be over-allocated, or if Boolean are to be honored, when scheduling jobs.
  - Specify Scheduler Log Events
    Specify Scheduler events which should be excluded from the scheduler's log.
  - Specify the Dedicated Time Queue Prefix
    Specify the prefix that indicates that a queue is a dedicated time queue.
  - Specify the Express Queue Minimum Priority
    Specify the minimum queue priority required for a queue to be classified as an express queue.
  - Configure Whether Jobs Can Cross Boundaries
    Control Whether Jobs Cross Primetime Boundaries
  - Set Minimum Job Priority Value for Job Execution
    Specify a minimum job priority value for jobs to run.
  - Configure Host-level Custom Resources
    Configure a dynamic host-level custom resource such as scratch space.
  - Select the Provisioning Policy
    Choose the policy to determine how vnodes are selected for provisioning.
  - Set Resources to Treat as Infinite if Unset
    Specify a list of host-level resources that will behave as if they are infinite when they are unset.
  - Set the Duration of the Scheduler Cycle
    Configure the maximum duration of the scheduler cycle.
  - Implement Preemption
    Preempt currently running jobs in order to run higher-priority work.
  - Examine Jobs Queue by Queue
    Specify whether all jobs that can be run from the highest-priority queue are run.
  - Implement Round-Robin Scheduling
    Select jobs from queues by examining the queues in round-robin fashion.
  - Implement Strict Ordering
    Use strict ordering to tell PBS that it must not skip a job when choosing which job to run based on execution priority.
  - Sort Jobs
    Sort jobs on a key that you specify.
  - Sort Vnodes
    Sort nodes on a key that you specify.
- Manage Hooks
  Configure when the Python Interpreter should be restarted for Hooks.
Monitor and Manage an HPC Cluster
Monitor jobs and nodes to understand what your HPC system is doing and improve performance.
Configure and Manage Allocations
Configure allocation periods, clusters and service units and manage allocation budgets, accounts, and transactions.
Run a Simulation
Run a simulation to perform what-if analysis to determine the most productive way to scale and configure an HPC.
Analyze HPC Resource Usage
View HPC resource usage charts and simulation results.
Configure and Manage Cloud Bursting
Burst HPC cluster nodes in the cloud when the demand for computing capacity spikes.
Node Management
The Node Management page enables you to monitor, filter, add and configure nodes quickly through an easy to use interface.
Real-time Dashboard
The real-time dashboard in Control (currently in preview release stage), provides live data in a single page to view several aspects about the overall operation of a cluster.
Log out of Control
Logout once your work is complete.

Altair Control™ User Guide

2021.1.0

Home
Configure the Workload Manager
Configure WLM server and scheduler settings, queues, resources, defaults and limits.
Manage Scheduling Parameters
Make changes to settings to modify the behavior of the Workload Manager Scheduler.

What's New
Get Started
Tutorials
Establish Access Controls
Manage Clusters
Configure the Workload Manager
Monitor and Manage an HPC Cluster
Configure and Manage Allocations
Run a Simulation
Analyze HPC Resource Usage
Configure and Manage Cloud Bursting
Node Management
Real-time Dashboard
Log out of Control

Manage Scheduling Parameters

Make changes to settings to modify the behavior of the Workload Manager Scheduler.

Implement Backfilling
Implement backfilling to keep resources from becoming idle when the top job cannot run.
Implement Starving Jobs
Track the amount of time a job has been waiting to run and then mark the job as starving if this time has passed a specified limit.
Implement Placement Sets
Use placement sets to improve task placement by exposing information on system configuration and topology.
Adjust the Scheduling Cycle Iteration
Adjust the time between scheduling iterations.
Use a Formula for Computing Job Execution Priority
Provide a formula that is used to calculate job execution priority.
Enable Scheduling
Enable or disable scheduling.
Implement Fairshare
Implement fairshare to ensure that jobs are run in the order of how deserving they are.
Configure Fuzzy Backfilling
Trade-off between scheduling cycle speed and the fineness of the granularity with which estimated start times are calculated.
Improve Throughput of Jobs
Improve job throughput by allowing the scheduler to run asynchronously.
Specify the Prime and Non-Prime Time Queue Prefix
Specify the prefix that indicates that a queue is a prime time or non-prime time queue.
Specify Resources to Schedule On
Specify those resources which are not to be over-allocated, or if Boolean are to be honored, when scheduling jobs.
Specify Scheduler Log Events
Specify Scheduler events which should be excluded from the scheduler's log.
Specify the Dedicated Time Queue Prefix
Specify the prefix that indicates that a queue is a dedicated time queue.
Specify the Express Queue Minimum Priority
Specify the minimum queue priority required for a queue to be classified as an express queue.
Configure Whether Jobs Can Cross Boundaries
Control Whether Jobs Cross Primetime Boundaries
Set Minimum Job Priority Value for Job Execution
Specify a minimum job priority value for jobs to run.
Configure Host-level Custom Resources
Configure a dynamic host-level custom resource such as scratch space.
Select the Provisioning Policy
Choose the policy to determine how vnodes are selected for provisioning.
Set Resources to Treat as Infinite if Unset
Specify a list of host-level resources that will behave as if they are infinite when they are unset.
Set the Duration of the Scheduler Cycle
Configure the maximum duration of the scheduler cycle.
Implement Preemption
Preempt currently running jobs in order to run higher-priority work.
Examine Jobs Queue by Queue
Specify whether all jobs that can be run from the highest-priority queue are run.
Implement Round-Robin Scheduling
Select jobs from queues by examining the queues in round-robin fashion.
Implement Strict Ordering
Use strict ordering to tell PBS that it must not skip a job when choosing which job to run based on execution priority.
Sort Jobs
Sort jobs on a key that you specify.
Sort Vnodes
Sort nodes on a key that you specify.

Copyright (c) 1994-2021. Altair Engineering, Inc. All Rights Reserved. Contains trade secrets of Altair Engineering, Inc. Copyright notice does not imply publication.

Intellectual Property Rights Notice | Technical Support