导航|地址|刷新
Welcome to RDHPCS’s Documentation — NOAA…










Accounts, Projects, and Access Tokens
Applying for a user account

Request access to RDHPCS projects

Requesting access to MSU-HPC systems
General Access Requirements

Foreign Nationals


Account Activity Requirements: Suspension, Deactivation, Reactivation
Deactivated Accounts


Request a New Allocation or Project

Common Access Card (CAC)

CAC/PIV SSH Key Enrollment

Yubikey
Configuring Yubikey for the NOAA RDHPCS

Yubikey Registration

Yubico Authenticator Setup Instructions for Windows/Mac

Yubico Authenticator Setup Instructions for Linux

Testing your Yubikey Registration for NOAA RDHPCS

Troubleshooting RDHPCS Yubikey Authentication


Role Accounts
Requesting a new role account

Requesting changes to a role account

Accessing a role account from your user account

Accessing a role account for Data Transfers or Bastion SSH Login

X Applications With role accounts

Using CRON or SCRON



Connecting
Connecting for the first time

Secure Shell (SSH) Access

Bastion Hostnames

Common Access Card (CAC) SSH Login
Linux

Mac OS

Windows


Yubikey SSH Login

Selecting a Node

Tectia
Tectia Initial Setup procedure

Install the Tectia Client

Configure the Tectia Client


Port tunnels

X11 forwarding

Reference

X11 Graphics

SSH Port Tunnels

Web based ParallelWorks Access


Getting Help
Submitting a Good Help Request
Use a Good Subject

Provide Detailed Description of the Problem

Provide Job Information

Describe How to Reproduce the Problem

Only Report One Problem Per Help Ticket

Follow up With Additional Information or Questions


Required Information for Specific Types of Help
Basic Ticket Information

File System Problems

Compilation Problems

Job Submission Problems

Job Completion Problems

Providing a Reproducer

Reporting Data Transfer Issues


Recording a Terminal Session
Using script to Record a Terminal Session

Submitting a Recording to the Help Desk


Managing Help Tickets
Reply to a Ticket

Search for a Ticket

Create a New Ticket



Systems
Cloud User Guide
Cloud Processing
Getting Help


Parallel Works
Using ACTIVATE

Using Parallel Works to access on-prem HPC Systems

Access the head node from within Parallel Works [PW]

Running a Jupyter workflow on a Slurm compute node

Accessing nodes in a cluster


Cloud Projects
Create/request a new project


Best Practices

Storage Types and Storage Costs
Lustre file system

Bucket/Block blob storage

Contrib file system


Costing
Cost Calculator


Errors
Warning messages from the on-prem system about exceeding quota


Usage Reports

Cloud Presentations

Knowledge Base from Asked Questions
General Issues

Storage functionalities

Clusters and snapshots

Data Transfer

Configuration Questions

AWS Lustre explained

Azure Lustre explained

Best practice in resource configuration

Slurm

Errors

Miscellaneous



Gaea User Guide
System Overview
Node types

Compute nodes

Login nodes

Data transfer nodes

System interconnect

File systems

Operating system


Connecting

Data and storage
NFS file systems

GPFS file systems

Move data to and from Gaea


Programming environment
Environment Modules

Compilers

MPI


Compiling
Compilers

Dynamic linking


Running jobs
Login vs Compute Nodes

Slurm


Software

Debugging
Linaro DDT

GDB

GDB4HPC

Valgrind4hpc


Profiling
HPE Performance Analysis Tools


Tips and tricks
GPFS (F5) Performance


Known issues
Open issues



Hera User Guide
Action Required

System Overview
System Configuration


Hera Partitions

Lustre File System Usage
Lustre Volume and File Count


Lustre
Hera Lustre Configuration

File Operations

Types of file I/O

File Striping

Userspace Commands


Applications and Libraries
Using Miniforge3 and Python on Hera

MATLAB

Using IDL on Hera

Using ImageMagick on Hera

Using R on Hera

Libraries


Using Modules

Using MPI
Loading the MPI module

Using PGI and mvapich2

Tuning MPI (TBD)

Profiling an MPI application with Intel MPI


Debugging Codes
Debugging Intel MPI Applications

Application Debuggers

Invoking DDT on Hera with Intel IMPI


Profiling Codes
Linaro Forge

TAU


Managing Contrib Projects


Jet User Guide
Jet Data Disposition


Mercury User Guide
System Overview

Interactive Logins

Data Transfer

Per User Data Management on Mercury

Software Stack

CRON Services

HPSS Access

Getting Help


MSU-HPC User Guide
Introduction
MSU’s Official HPC Documentation


General Information
Logging In


Running Jobs on MSU-HPC Systems
Submitting a Job

Specifying a Partition

Monitoring Jobs

Getting Information about your Projects


MSU-HPC System Configuration
File Systems

Orion Compute System

Hercules Compute System


Account Management
Overview

MSU Account Management Policies

Managing Project and Role Account Members

NOAA Portfolio, Project, and User Management on MSU-HPC

Getting An Account

Account Renewal

Managing Portfolios, Projects and Allocation

Role Accounts


Help, Policies, Best Practices, Issues
MSU-HPC Help Requests

Policies and Best Practices

Protecting Restricted Data


MSU FAQ


PPAN User Guide
Login to Analysis
ssh setup for GFDL Workstations

C-shell Setup

C-shell keybinds for analysis

xterm and gnome-terminal

Logging into the PP nodes


File Systems
/home and /nbhome

/archive

/vftmp & $TMPDIR

/work

/ptmp

/net, /net2, /net3


Batch Software
Access

Local Commands

Gotchas


Podman

Analysis Software
Using Modules

netcdf Library

MATLAB Licenses


About Archrpt
Report Option [-r]

Show Archive Report By Specified Group [view]

Show Archive Report By Specified User [view]

Show Archive Report By Specified Group and Sort By Files [view] [sort]

Show Archive Report By Specified Group and Sort By Bytes [view] [sort]

Show Archive Report By Specified Date [date]

Summary Option [-s]

Show Archive Summary

Show Archive Summary and Sort By Files [sort]

Show Archive Summary and Sort By Bytes [sort]

Show Archive Summary By Date [date]


Group Quotas

User Quotas
Info

Configuration


Enforcing Quotas

Analysis Hosts

Post-Processing Hosts


Ursa User Guide
System Overview

System Configuration
Partitions

Node Sharing

Front Ends and Service Partition


Using GPU Resources on Ursa
Using H100 GPU Resources With a GPU allocation

Using H100 GPU Resources Without a GPU allocation


Using the Exploratory GPU Resources

Software Stack

HPFS File Systems
Usage/Quota information for /scratch[3,4] file systems


Cron and Scrontab Services

Getting Started with Ursa

Getting Help



Data Storage and Transfers
Data Storage
Summary of Storage Areas
Notes on User-Centric Data Storage

User Home Quotas

User Home Permissions

User Home Backups

User Archive Access

User Archive Accounting

Project Home Path, Quota, and Permissions

Project Home Backups

Project Work Areas to Facilitate Collaboration (Gaea)

Permissions

Backups

Permissions

Project Archive Access


NESCC HPSS
Gaining Access to use HPSS

New HPSS User Requests

Adding New Projects to HPSS

Portfolios Using HPSS

Expired Data Deletion Process

User Notification

Expired Data - Deletions

HTAR

HTAR Cookbook

HSI

HSI Basic Usage

HSI Cookbook

File Expiration Commands

Sample HPSS Batch Job


GFDL Archive
Gaining Access to use the GFDL Archive

Allocation and Quota

Finding Files



Data Transfers
Overview
Data Transfer Methods

Requests for Firewall Exceptions

Firewall Exception Terms


Transferring Data
Globus Connect

Trusted Data Transfer Nodes (DTN)

Untrusted Data Transfer Nodes (UDTN)

Transfer and Syntax Examples

Firewall Modification Requests for DTNs

Unattended Data Transfers or Password-less Transfers to/from RDHPCs Systems

Using a Pre-Established SSH Port Tunnel


Globus Online Data Transfer
Overview

RDHPCS Globus Collection Summary

NOAA RDHPCS Globus Endpoint Types

Globus Command Line Interface (CLI)

Transferring Data to and from Your Computer

Data Sharing with External Collaborators

Globus Example

Using Globus Online Data Transfer

NOAA RDHPCS Globus Endpoint Types

NOAA RDHPCS UDTN’s (Globus Untrusted Endpoint)

NOAA RDHPCS Object Stores in the Cloud

Accessing Cloud Endpoints in our environment

Globus Command Line Interface (CLI)

Transferring Data to and from Your Computer

GFDL Institutional Data Portal

Migrating Data Between Local File Systems

Known Issues




Policies and Best Practices
System Usage
Login Node Usage


Cron and scrontab usage
Cron/Scrontab Dos

Cron/Scrontab Don’ts


Allocations
Request an Increase in Allocations

Adding a Project to an Allocation

Cloud Computing Allocations


Quotas
Requesting Additional Storage for a Project


File System Usage Practices and Policies
High Performance File System (HPFS - Scratch)

General Parallel File System (GPFS)

/data_untrusted

HFS

Long Term Data Archive Storage


Filesystem Backup and Data Retention
Recover Recently Deleted Files from /home

HPSS (Data Retention)
Expired Data Deletion Process

Data Recovery Policy



Data Disposition
HPFS (Scratch) Data

Closed Projects

Home File System (HFS) Data

Protecting Restricted Data


Managing Packages in /contrib
Overview of contrib Packages

Responsibilities of a contrib Package Maintainer

contrib Packages Guidelines

contrib Package Maintainer Requests

Managing a contrib Package

Maintaining “Metadata” for contrib Packages

contrib Package Directory Naming Conventions


Queue Policy
Overview

Specifying a Quality of Service (QOS)

Changing QOS’s

Hera and Ursa QOS

Gaea

General Recommendations

Priorities Between QOS

Debug & Batch QOS



Software
Modules
View Active Modules

Finding Modules

Load Modules

Module Collections

Adding Additional Module Paths

Using Modules in Batch Jobs

Command Summary


Python
Overview

Python Guides
Conda Basics

Installing Miniforge

Jupyter on RDHPCS Systems


Module Usage
Python

Conda


Python and Conda Environments
Conda Environments


Running Python
RDHPCS compute nodes


Python best practices
Cloning the base environment using conda

Environment locations (storage)

Adding known conda environment locations

Keep your Python and Conda caches trimmed

Deactivate your environments before running batch jobs

Unbuffered input


Additional Resources


Workflow Software
Cron and Slurm crontab
Cron

Viewing currently running crontab processes

Slurm Crontab


Flexible Modeling System Runtime Environment

Rocoto
Rocoto on RDHPCS Systems

Rocoto Documentation

Rocoto Help

Rocoto Best Practices



Debugging with Forge DDT
DDT remote connection

Debugging an MPI process
First time configuration

Submit a debug job



Compilers
Using the Intel compiler with Intel MPI
Recommended Intel Compiler Options for Optimization


Compiling for reproducible results
Recommended Intel compiler options for debugging

Thread-safe compilation

Other potentially useful Intel compiler options


Using the NVIDIA compilers
Documentation on NVIDIA compiler options

NVIDIA compiler options for optimization

NVIDIA compiler options for debugging



Containers
Introduction
Background

Supported RDHPCS Container Solutions


Singularity
Differences between SingularityCE and Apptainer

Limitation, Exception and Liability

How to create images

Run a Single Node or Single Core Containers

Run an MPI-dependent container



VSCode
Setting up VSCode for use on RDHPCS systems
Set up the install directory

Set up Port Forwarding on your Local client

Create a session with the tunnel and keep this session open

Test to make sure the tunnel is working

Login with your VSCode client




Slurm
Overview
Running a Job
Batch Scripts

Interactive Jobs

Common sbatch Options

Specifying Partitions and QOS


Determine and specify a memory limit for your jobs
Specify a processor layout for your job (uniform layout example)

Using report-mem utility in batch jobs

Using report-mem utility on a job that is currently running

Determining the amount of memory used by a process

Slurm Environment Variables

State Codes

Job Reason Codes

Job Dependencies

Srun


Monitoring Jobs
Show Pending and Running Jobs

Show Completed Jobs

Getting Details About a Job


Priority and Fairshare
Understanding Slurm Fairshare

Fairshare Priority Factor

Fairshare Definitions

Fairshare Reporting

Priority Reporting


Getting Information About Your Projects
sfairshare

saccount_params


Generating Reports
Sreport

Shpcrpt


References


Gaea Batch Job Overview
Compiling

Running

Staging/Combining

Transferring Data to/from Gaea

Allocation

Running a Simple Job

Running the Script
…(内容过长已截断)

------
导航页 打开网址