Simulators vs Real Hardware

Overview

This page introduces the programming concept of choosing the right backend.

It answers: What do simulators do, why are they idealized, and when should you use hardware?

Conceptual Mapping

From earlier tracks:

• In Foundations, circuits are described as ideal operations.
• In Noise & Errors, you learned real devices have decoherence, gate errors, and readout errors.

In programming terms:

• Simulator backend: executes the ideal circuit model (or a chosen noise model) on a classical computer.
• Hardware backend: runs the circuit on a physical device with real constraints and variability.

So “backend choice” is not just convenience. It changes what assumptions are being tested.

Code Walkthrough

A minimal pattern showing the idea of switching backends:

from qiskit import QuantumCircuit
 
qc = QuantumCircuit(1, 1)
qc.h(0)
qc.measure(0, 0)
 
# Conceptually:
# - a simulator backend runs an idealized model
# - a hardware backend runs on a real device
# (We’ll keep hardware setup out of this beginner section.)

Line by line:

• You build the same circuit either way.
• The difference is where you execute it and what effects are present.

What Happens Under the Hood

What simulators do

At a high level, a simulator:

• represents the circuit’s evolution in a classical data structure
• applies ideal operations (or a specified noise model)
• samples measurements to produce counts

Simulators are great for:

• debugging circuit logic
• checking expected distributions
• learning measurement statistics

Why simulators are idealized

Even when they sample measurement outcomes, simulators often assume:

• perfect gates
• no drift
• no crosstalk
• perfect measurement (unless you add a noise model)

That idealization is useful for learning, but it can hide real-world constraints.

Why hardware behaves differently

On hardware:

• operations are approximations of ideal gates
• noise accumulates with circuit depth
• connectivity constraints can add extra routing operations
• calibration drift can change results over time

This is exactly what you learned in Hardware Models & Metrics. Programming is where you start seeing those effects in your own outputs.

Turtle Tip

Common Pitfalls

Quick Check

What’s Next

You now have the basic workflow: build a circuit, run it, interpret counts, and understand backend differences. Next we can start writing slightly larger circuits (multi-qubit), introduce basic visualization, and connect programming patterns to the algorithms you already learned.