[Contents] [Index] [Top] [Bottom] [Prev] [Next]


Contents

Preface

Audience
LSF Suite 3.2
LSF Enterprise Edition
LSF Standard Edition
Related Documents
Online Documentation
Technical Assistance

1 - Introduction

What is LSF?
LSF Features
Host Resources
Batch Processing
Interactive Processing
Clusters
Fault Tolerance
Structure of LSF Base
Load Information Manager
Remote Execution Server
LSF API
LSF Utility Programs
Applications
Structure of LSF Batch

2 - Getting Started

Getting Cluster Information
Displaying the Cluster and Master Names
Displaying Available Resources
Getting Host Information
Displaying Static Host Information
Displaying Load Information
Running Jobs
Running Jobs on Remote Hosts
Load Sharing Commands With lstcsh
Parallel Processing With LSF Make
Listing Hosts
Submitting a Job
Selecting a Job Queue
Tracking Batch Jobs
xbsub and xlsbatch GUI Applications

3 - Cluster Information

Finding the Master
Listing Resources
Listing Hosts
Displaying the Load
Graphical Load Display

4 - Resources

Introduction to Resources
Load Indices
Static Resources
Shared Resources
Boolean Resources
Listing Resources
Resource Requirement Strings
Selection String
Order String
Resource Usage String
Job Spanning String
Specifying Shared Resources
Configuring Resource Requirements
Remote Task List File
Managing Your Task List
Using Resource Requirements

5 - Using LSF Batch

Batch Jobs
Fairshare Scheduling Policy
Host Partition Fairshare Scheduling
Queue-Level Fairshare Scheduling
Hierarchical Fairshare
Other Scheduling Policies
Preemptive Scheduling
Exclusive Scheduling
Processor Reservation
Backfill Scheduling
Scheduling Parameters
Load Thresholds
Scheduling Conditions
Time Windows for Queues and Hosts
Run Windows
Dispatch Windows
Batch Queues
Finding Out What Queues Are Available
Detailed Queue Information
Automatic Queue Selection
Specifying Default Queues
Queue Selection Mechanism
Choosing a Queue
Batch Users
Batch Hosts
User and Host Groups
Viewing Hierarchical Share Information
Queue-Level Job Starters
Configuration Parameters
User Controlled Account Mapping

6 - Submitting Batch Jobs

Input and Output
Resource Requirements
Resource Reservation
Host Selection
Host Preference
Resource Limits
Pre-Execution Commands
Job Dependencies
Job Dependency Examples
Remote File Access
Start and Termination Time
Parallel Jobs
Minimum and Maximum Number of Processors
Specifying Locality
Job Arrays
Creating a Job Array
LSB_JOBINDEX Environment Variable
Array Job Dependencies
Handling Input/Output/Error Files for Job Arrays
Specifying a Share Account
Re-initializing Job Environment on the Execution Host
Other bsub Options
Job Scripts
Embedded Submission Options
Running a Job Under a Particular Shell
Submitting Jobs Using the Job Submission GUI

7 - Tracking Batch Jobs

Displaying Job Status
Finding Pending or Suspension Reasons
Monitoring Resource Consumption of Jobs
Displaying Job History
Viewing Chronological History
Checking Partial Job Output
Tracking Job Arrays
Displaying Queue and Host Status
Job Controls
Killing Jobs
Suspending and Resuming Jobs
Controlling Job Arrays
Sending an Arbitrary Signal to a Job
Moving Jobs Within and Between Queues
Job Modification
Submitted Job Modification
Dispatched Job Modification
Job Array Modification
Job Tracking and Manipulation Using the GUI

8 - Running Interactive Jobs

Shared Files and User IDs
Running Remote Jobs with lsrun
Running Parallel Jobs with lsgrun
Load Sharing Interactive Sessions
Load Sharing Login
Load Sharing X Sessions
Command-Level Job Starters
Interactive Batch Job Support
Shell Mode for Remote Execution

9 - Using lstcsh

Starting lstcsh
Using lstcsh as Your Login Shell
Automatic Remote Execution
Host Redirection
Job Control
Built-in Commands
The lsmode Command
The connect Command
Modes of Operation
Differences from Other Shells
Writing Shell Scripts in lstcsh
Limitations

10 - Using LSF Make

Parallel Execution
Invoking LSF Make
Specifying the Number of Processors
File Server Load
Tuning Your Makefile
Building in Subdirectories
Running lsmake as a Batch Job
Differences from Other Versions of make

11 - Checkpointing and Migration

Approaches to Checkpointing
Kernel-level Checkpointing
User-level Checkpointing
Application-level Checkpointing
Checkpoint Directory
Uniform Checkpointing Interface
The echkpnt Command
The erestart Command
Submitting Checkpointable Jobs
Checkpointing a Job
Restarting a Checkpointed Job
Job Migration
Queues and Hosts for Automatic Job Migration
Automatically Rerunning and Restarting Jobs
Submitting a Job for Automatic Migration
Building Checkpointable Jobs
The Checkpoint Library
The Checkpoint Startup Routine
Linking
Limitations

12 - Customizing Batch Jobs for LSF

Environment Variables
Parallel Jobs
Getting the Host List
Starting Parallel Tasks With lstools
Using LSF Make to Run Parallel Batch Jobs
Submitting PVM Jobs to LSF Batch
Submitting MPI Jobs to LSF Batch
Submitting POE Jobs to LSF Batch
Using a Job Starter for Parallel Jobs

13 - Using LSF MultiCluster

What is LSF MultiCluster?
Getting Remote Cluster Information
Running Batch Jobs across Clusters
Running Interactive Jobs on Remote Clusters
User-Level Account Mapping Between Clusters

14 - Interoperation with NQS

Choosing an LSF Batch Queue
Submitting a Job from LSF to NQS
Controlling Jobs Running on NQS
Forwarding of Output Files

A - Customizing xlsbatch Menu Items

B - Frame Arrays

Overview
Distribution
Frame Array Concepts
Submitting Frame Arrays
Tracking Frame Arrays
Controlling Frame Arrays

C - Using LSF with Alias Renderer

Overview
Distribution
Installing the queue-level job starter
Submitting Checkpointed Frame Arrays
Tracking Checkpointed Frame Arrays

D - Using LSF with FLUENT

Overview
Distribution
Configuring the Checkpointing Executable Files
Submitting the FLUENT Job
Checkpointing the FLUENT job
Restarting the FLUENT job

Index


[Contents] [Index] [Top] [Bottom] [Prev] [Next]


doc@platform.com

Copyright © 1994-1998 Platform Computing Corporation.
All rights reserved.