Chapter 7 AHPCC: Arkansas High Performance Computing Cluster

For advanced computational work, the University of Arkansas provides access to the Arkansas High Performance Computing Cluster (AHPCC). This chapter shows you how to set up your account for future use.

What you’ll accomplish in this chapter:

• ✅ Run RStudio interactively on AHPCC (primary focus)
• ✅ Verify your AHPCC account access
• ✅ Log into the AHPCC Portal
• ✅ Understand what AHPCC is and when to use it
• ✅ Explore available computing resources

Most Important for This Course:

The key feature you’ll use is RStudio on AHPCC—a cloud-based RStudio environment with powerful computing resources. Skip to Step 3 if you just want to launch RStudio!

7.1 What is AHPCC?

The Arkansas High Performance Computing Cluster is a university-wide computing resource that provides:

• 🖥️ Massively parallel computing: Run large computations much faster
• 📊 Big data processing: Analyze large datasets efficiently
• 💾 Shared storage: 1TB+ project storage for collaborative work
• 🔧 Pre-installed software: R, Python, MATLAB, COMSOL, and more
• 🌐 Web portal access: Access from any browser (no SSH required)
• 🎓 Free for university members: Included in your tuition

For This Course:

While Posit Cloud is perfect for typical coursework, AHPCC becomes valuable if you: - Run computationally intensive simulations - Analyze very large datasets - Need to collaborate on complex projects - Explore advanced statistical methods

For most students, Posit Cloud will be sufficient for this course.

7.2 Why You Have AHPCC Access

As a University of Arkansas student, you have a free AHPCC account. Your HPC ID is typically the first part of your email address (e.g., if your email is user@uark.edu, your HPC ID is user).

7.3 Step 1: Verify Your AHPCC Account

7.3.1 1.1: Access the AHPCC Portal

Navigate to:

https://hpc-portal2.hpc.uark.edu

You’ll see the Posit Open OnDemand portal login page.

7.3.2 1.2: Log In

Username: Your UARK email first part (e.g., user not user@uark.edu)

Password: Your UARK email password

Important Security & Access:

• Use your UARK email password (same as your university email)
• First-time users: Your supervisor may need to approve your AHPCC account access before you can log in
• If login fails, contact your instructor or the AHPCC support team at hpc@uark.edu
• Never share your credentials

Click “Sign In”

7.4 Step 2: Navigate the AHPCC Portal

7.4.1 2.1: Portal Dashboard

After login, you’ll see the Posit Open OnDemand dashboard:

Main sections you’ll see:

1. Files: Browse your home directory and project storage
2. Jobs: View queued and running compute jobs
3. Clusters: SSH terminal access to compute nodes
4. Interactive Apps: Launch graphical applications (including RStudio!)
5. My Interactive Sessions: Manage currently running applications

7.5 Step 3: Access RStudio on AHPCC

One of the most useful features of AHPCC is interactive RStudio, which gives you access to powerful computing with the familiar RStudio interface.

7.5.1 3.1: Launch Interactive RStudio

In the portal dashboard:

1. Click “Interactive Apps” (top menu)
2. Look for “RStudio Server” option
3. Click “Launch a new session”

You’ll see configuration options:

• Cluster: Choose pinnacle (main cluster) or karpinski (CSCE cluster)
• Partition: Leave as default
• Number of cores: Start with 4 for this course
• Amount of memory: Start with 16GB
• Walltime: Duration needed (e.g., 1:00 for 1 hour)

Resource Limits:

Computational resources are shared, so be considerate: - Request only what you need - Keep walltime short (typically max 4-24 hours) - Monitor your jobs to avoid waste

Click “Launch”

7.5.2 3.2: Wait for RStudio to Boot

The portal will show “Your session is being prepared…”

This takes 1-5 minutes depending on system load. You’ll see a “Connect to RStudio Server” button when ready.

Click the button to launch RStudio in a new browser tab.

What you’ll see:

A full RStudio interface connected to AHPCC’s computational power! Your files are in /home/username.

7.5.3 3.3: Using AHPCC RStudio

Your AHPCC RStudio works exactly like Posit Cloud or local RStudio, but with:

• ✅ More computing power: Multi-core processors for faster computation
• ✅ More memory: 16GB+ available (vs. ~2GB on personal laptops)
• ✅ Pre-installed software: R, Quarto, Git, and many others already installed
• ✅ Shared storage: Easy access to collaborative project files

7.5.4 3.4: End Your Session

When done, click “Delete” in the portal dashboard under “My Interactive Sessions” to free up resources for other users.

7.6 Step 4: Command-Line Access (Advanced)

For advanced users, you can also access AHPCC via command line (SSH).

7.6.1 4.1: SSH Access

From your terminal/command prompt:

ssh user@hpc.uark.edu

Replace user with your HPC username.

This gives you direct terminal access for: - Running long-running batch jobs - Submitting jobs to the scheduler - Advanced scripting

For This Course:

Command-line access is optional. The interactive portal is easier for most students.

7.7 Step 5: Submit Batch Jobs (Advanced)

For very long-running computations, you can submit batch jobs to the scheduler.

7.7.1 5.1: Create an R Script

In your home directory, create analysis.R:

# This script runs on AHPCC
library(tidyverse)

# Simulate large dataset
data <- tibble(
  x = rnorm(1000000),
  y = rnorm(1000000)
)

# Perform analysis
result <- data |>
  mutate(z = x + y) |>
  summarize(mean_z = mean(z))

print(result)

7.7.2 5.2: Create a SLURM Job Script

Create run_analysis.sh:

#!/bin/bash
#SBATCH --job-name=r_analysis
#SBATCH --cpus-per-task=4
#SBATCH --mem=16G
#SBATCH --time=01:00:00
#SBATCH --output=results.log

module load r
Rscript analysis.R

7.7.3 5.3: Submit the Job

sbatch run_analysis.sh

Your job enters the queue and runs when resources become available.

For This Course:

Batch job submission is advanced and optional. Interactive RStudio is sufficient for coursework.

7.8 Troubleshooting AHPCC Access

7.8.1 “RStudio Won’t Launch”

• Check that you have compute hours available (quota)
• Try with fewer resources (e.g., 2 cores instead of 4)
• Wait a few minutes—queues may be busy
• Try the karpinski cluster instead of pinnacle

7.9 AHPCC Resources and Documentation

For more information, visit the AHPCC Wiki:

https://hpcwiki.uark.edu/

Useful pages:

• Portal Login
• Getting Started
• Job Submission
• Software Modules

Contact HPC support: hpc@uark.edu

7.10 When to Use Each Platform

Platform	Best For	Resources
Posit Cloud	Getting started, typical coursework	25 hrs/month free
Local RStudio	Fast work, offline, long-term projects	Your computer’s specs
AHPCC RStudio	Heavy computation, big data, collaboration	Unlimited university resources

Recommendation:

For AGST 50104 Experimental Design: 1. Start with Posit Cloud (immediate, no setup) 2. Optionally use local RStudio (Chapter 5) for consistency 3. Use AHPCC if your simulations/analyses need more power

All three are compatible—you can switch between them seamlessly!

7.11 Summary Checklist

Before moving to the next chapter, make sure you have:

• ✅ Accessed the AHPCC Portal at least once
• ✅ Verified your login credentials work
• ✅ Explored the dashboard and file structure
• ✅ (Optional) Launched interactive RStudio on AHPCC
• ✅ Reviewed the AHPCC Wiki documentation

Great! 🎉

Your AHPCC account is ready for advanced work. For now, focus on Posit Cloud or local RStudio. You’ll come back to AHPCC if your analyses need more computational power.

---

Next: Chapter 8: GitHub Copilot Setup