Compute

Compute is the processing capacity and work supplied by hardware to train, run, and operate digital models and systems.

Plan for this page

Understand one concept in plain language and connect it to nearby ideas.

You will leave with: It can be described through operations, processor time, memory, networking, throughput, energy, cost, or available hardware. The useful measure depends on the task, architecture, software efficiency, precision, and utilization.
Time: 3–8 min
Before you begin: No prerequisite. Keep one real example in mind.
Do this now: Read the explanation, then compare the example with the common misunderstanding.

Prepared next stepPlace AI inside physical infrastructure

ComputeCompute is the processing capacity and work supplied by hardware to train, run, and operate digital models and systems.

Training
Accelerator
Data center

Plain explanation

It can be described through operations, processor time, memory, networking, throughput, energy, cost, or available hardware. The useful measure depends on the task, architecture, software efficiency, precision, and utilization.

Example

Training a large model may use many accelerators together for weeks, while one inference request uses a small share of deployed capacity for a short time.

Why it matters

Compute shapes cost, speed, access, concentration, infrastructure, and environmental effects. A raw hardware count does not reveal useful work without efficiency and utilization context.

Common misunderstanding

Misconception: more compute always produces a proportionally better system. Correction: data, method, architecture, objective, bottlenecks, and diminishing returns also determine results.

Related concepts

Follow relationships, not only the alphabet

TrainingTraining is the process of adjusting a model to improve performance on an objective using data, feedback, or interaction.AcceleratorAn accelerator is specialized hardware designed to perform selected computations faster or more efficiently than a general-purpose processor.Data centerA data center is a facility and operating environment that houses computing, storage, networking, power, cooling, and security systems.

Place AI inside physical infrastructure

Sources and scope

What supports this object

Energy and AIInternational Energy Agency

Provides measured and scenario-based evidence about AI, data centers and energy; projections depend on geography, workload, efficiency, system boundary and demand assumptions.

Open source

Loading page content...

…