Skip to main content

FAQ

Helpful information

CPU GPU
Purpose The CPU is often referred to as the “brain” of the computer. It handles all the general-purpose processing tasks. Originally designed to handle rendering of images and video, GPUs are now also used for tasks that require parallel processing.
Architecture Typically has a few cores (ranging from 2 to 64 in high-end models) designed for serial processing, meaning they handle tasks one at a time but very quickly. Contains hundreds to thousands of smaller cores designed for parallel processing, meaning they can handle many tasks simultaneously.
Tasks  Ideal for tasks that require high interactivity and quick switching between tasks, such as running the operating system, and applications, and managing input/output operations. Excellent for tasks that can be broken down into smaller, parallel operations, such as rendering graphics, machine learning, and scientific computations.
Latency / Throughput Low latency, meaning it can quickly respond to individual tasks High throughput, meaning it can process a large number of operations simultaneously
Use Examples Running your operating system, browsing the web, and managing files. Playing video games, rendering 3D models, and training machine learning models

 

 

In High Performance Computing (HPC) clusters, login nodes and compute nodes serve distinct purposes:

Login Nodes

Compute Nodes
Purpose Login nodes are the entry points to the HPC cluster. Users log in to these nodes to prepare their work. Compute nodes are where the actual computational work is performed.
Functions They are used for tasks such as editing scripts, compiling code, and submitting jobs to the scheduler. They execute the jobs submitted by users from the login nodes. These nodes handle the heavy lifting of processing data and running simulations.
Access Users can directly access login nodes via SSH. Users do not log in directly to compute nodes. Instead, jobs are dispatched to these nodes by the scheduler.
Resources: Typically have fewer resources compared to compute nodes, as they are not intended for heavy computational tasks. Equipped with high-performance CPUs, large amounts of memory, and sometimes GPUs, designed to handle intensive computational tasks123.

Workflow

 You log in, prepare your job script, and submit it to the scheduler. The scheduler assigns your job to a compute node, where it runs until completion.

 

 

PuTTY

Command Line (SSH)
  • Graphical Interface: PuTTY provides a graphical user interface (GUI), making it easier for users who prefer not to work solely with command-line interfaces
  • Customization: It offers various customization options for the terminal window, such as font size, colors, and window behavior.
  • Session Management: You can save sessions, which means you don’t have to re-enter the server details every time you connect.
  • Key Management: PuTTY includes tools like PuTTYgen for generating SSH keys and Pageant for managing SSH key
  • Compatibility: Primarily used on Windows, but versions are available for Unix and macOS as well
  • Built-in Utility: On Unix-based systems (like Linux and macOS), SSH is typically built-in, so you don’t need to install additional software.
  • Simplicity: Using the command line can be faster and more straightforward for users comfortable with text-based interfaces.
  • Scriptability: Command-line SSH can be easily integrated into scripts and automated tasks, making it powerful for advanced users.
  • Native Integration: On Windows 10 and later, OpenSSH is integrated into PowerShell, allowing for native SSH connections without third-party software

 

 

The salloc command is used to allocate resources for a job within the SLURM workload manager.

Resource Allocation: salloc requests a set of resources (such as nodes, CPUs, and memory) from the SLURM schedule.

Interactive Sessions: It is often used to start interactive sessions, allowing users to run commands directly on the allocated resources.

Basic Command:    salloc [options] [command]

If no command is specified, salloc will start a new shell on the allocated resources

 

SLURM (Simple Linux Utility for Resource Management) is a cluster management and job scheduling system that used in many of the world’s supercomputers and computing clusters.

  1. Resource Allocation: SLURM allocates exclusive or non-exclusive access to resources (compute nodes) for users to perform work.
  2. Job Management: It provides a framework for starting, executing, and monitoring jobs on the allocated nodes.
  3. Resource Contention: SLURM manages a queue of pending jobs and arbitrates contention for resources
  4. Scalability: Can schedule up to 100,000 jobs on large clusters.

1. Portability

  • Containers encapsulate all the dependencies and libraries required by an application, ensuring it runs consistently across different environments
  • Applications can be easily moved between different HPC systems or from development to production environments without compatibility issues

2. Reproducibility

  • Containers help in creating reproducible research environments, which is crucial for scientific experiments and simulations
  • They allow for precise version control of software and dependencies, making it easier to replicate and validate results

3. Efficiency

  • Containers are more lightweight compared to traditional virtual machines, leading to lower overhead and better performance
  • They enable better resource utilization by allowing multiple isolated applications to run on the same hardware without interference

4. Flexibility

  • Users can create custom software environments tailored to their specific needs without affecting the host system or other users
  • Containers support DevOps practices, enabling continuous integration and continuous deployment (CI/CD) workflows

5. Scalability

  • Containers can be easily scaled up or down to meet the demands of different workloads
  • They integrate well with cluster management and orchestration tools like Kubernetes, enhancing the scalability of HPC applications

6. Security

  • Containers provide a level of isolation between applications, reducing the risk of conflicts and enhancing security
  • They allow for controlled access to resources, ensuring that applications run in a secure and confined environment

 

How-Tos

1. Make sure you have Putty downloaded and installed: https://www.putty.org/

2. Open up Putty and you will be prompted with a dialog window (see below). Enter the host name: hopper001

3. You will be prompted to "login as", please enter your utep credentials username

4. You will be prompted to enter a password, please use your utep account password (when you type, the password you type will not be visible, don't worry, its a security measure)

5. You should be getting a message from the university about authorization and compliance (see below)

6. And you're in!

putty1.png

putt2.png

Before you start: Make sure you are either on the utep wifi or using the Utep VPN (global protect) to connect to the utep network.

WIndows 10 / Windows 11:

  1. Go to terminal (click Start Button or Search Button on the taskbar and search "cmd")
  2. In the terminal window (see below), type "ssh your-username@utep.edu@hopper001" (enter your utep username)
  3. You will be prompted to enter a password, please use your utep account password (when you type, the password you type will not be visible, don't worry, its a security measure)
  4. You should be getting a message from the university about authorization and compliance (see below)
  5. And you're in!

windows-ssh.png

Mac OS:

  1. Bring up the terminal (search "terminal" in the Spotlight search function)
  2. In the terminal window, enter "ssh hopper001"
  3. You will be prompted to enter your username, use your utep credentials username
  4. You will then be prompted to enter a password, please use your utep account password (when you type, the password you type will not be visible, don't worry, its a security measure)
  5. You should be getting a message from the university about authorization and compliance (see below)
  6. And you're in!

Answer

Answer

Connect With Us

The University of Texas at El Paso
College
Building, Room
500 W University
El Paso, Texas 79902

E: cchopper@utep.edu
P: (915) 747-5000