Frontier
A Hands-On High Performance Computing Intuition Engine
Explore High-Performance Computing
Select a module to build intuition about HPC concepts through interactive visualizations, simulations, and hands-on experiments — from resource sizing to GPU parallelism.
Drag-and-drop computing tasks onto laptops, workstations, or supercomputers to learn which resource fits each workload.
Configure job parameters, submit to a simulated cluster, and follow your job through the HPC pipeline in real time.
Explore Amdahl's Law in action — add workers and watch diminishing returns unfold through animated visualizations.
Race a CPU against a GPU rendering the same image pixel-by-pixel and discover why GPUs dominate parallel workloads.
Rightsizing
Match Tasks to Resources
HPC Workflow
From Click to Results
Scaling
Why Optimization Hits a Wall
CPU vs GPU
Parallel Processing Power
Rightsizing: Match Tasks to Resources
Not every task needs a supercomputer, and not every task fits on a laptop. Rightsizing means choosing the computing resource that best matches each job's requirements — saving time, money, and energy. Drag each task card onto the resource where it belongs best. Look at the CPU, memory, and GPU requirements to find the right match.
Task Deck
Job Configuration
Configure the resources your job will request from the cluster. Use the sliders below to set how much computing power, memory, and time you need.
HPC Job Pipeline
HPC stands for High-Performance Computing — powerful clusters of interconnected computers that solve problems too large for a single machine. This pipeline shows how a job submitted from your local computer travels through the cluster: from submission and scheduling, through execution, all the way to collecting your results.
Optimization Insights
Experiment Setup
Explore how adding more workers (processors) affects performance. Adjust the sliders to change the workload and see the limits of parallel scaling in action.
The serial portion cannot be parallelized. No matter how many workers you add, this part takes the same time!
Run an experiment to see insights.
Live Scaling Simulation
In HPC, throwing more processors at a problem doesn't always make it faster — understanding scaling is key to using resources wisely. Watch below as work is distributed in real time: the serial portion runs on a single worker first, then the parallel portion is split among all available workers simultaneously.
Work Queue
Tasks waiting to be processed
How Your Work Flows
Results
Render Settings
CPUs and GPUs process pixels very differently. Pick an image, choose a resolution, and watch both processors race side by side to see why GPUs dominate graphics workloads.
Why is GPU faster?
CPU
8 powerful cores
Sequential processing
GPU
1000s of simple cores
Massive parallelism
Live Render Comparison
A Central Processing Unit (CPU) processes pixels one by one using a few powerful cores, while a Graphics Processing Unit (GPU) computes thousands of pixels simultaneously using massively parallel architecture.
GPU Advantage
GPU Time Savings
Now Imagine: Animating Your Neon Plasma
You just rendered a single Neon Plasma image. But what if you wanted to create an animated GIF or video? Each frame requires the same computation. The CPU vs GPU gap becomes catastrophic.