DiVincenzo Criteria & Benchmarks
What a quantum computer needs, and how we grade one.
In 2000 David DiVincenzo wrote down what any quantum computer needs: five requirements, plus two more for sending quantum information between machines. Today the hard questions are quantitative. How good is good enough, and how do you measure a gate's error when your state preparation and measurement are themselves imperfect?
What you’ll learn
- State DiVincenzo's five criteria and two networking criteria, and attach a measurable figure of merit to each.
- Define average gate fidelity and process fidelity, convert between them, and compute both for depolarizing noise.
- Model / decoherence and derive the coherence limit on single-qubit gate fidelity.
- Explain SPAM errors and why randomized benchmarking is insensitive to them; extract an error per Clifford from a decay curve.
- Say what holistic benchmarks (quantum volume, cross-entropy, layer fidelity) measure and what they hide.
The DiVincenzo criteria
DiVincenzo's 2000 paper turned "build a quantum computer" into a checklist. Five criteria concern computation; two more concern communication, moving quantum states between processors or across a network. Each has since become a set of numbers that experiments report.
| # | Criterion (DiVincenzo 2000) | Modern figures of merit |
|---|---|---|
| 1 | A scalable physical system with well-characterized qubits | qubit count and yield; known Hamiltonian, couplings and leakage levels; parameter spread between qubits |
| 2 | The ability to initialize the state of the qubits to a simple fiducial state, such as | reset fidelity and duration, including mid-circuit reset of ancillas |
| 3 | Long relevant decoherence times, much longer than the gate operation time | , , gate durations, and the resulting coherence-limited error |
| 4 | A "universal" set of quantum gates | single- and two-qubit gate fidelities (e.g. from randomized benchmarking), crosstalk, leakage |
| 5 | A qubit-specific measurement capability | single-shot readout fidelity and duration; disturbance of neighboring qubits |
| 6 | The ability to interconvert stationary and flying qubits | efficiency and fidelity of transferring a qubit onto a photon and back |
| 7 | The ability faithfully to transmit flying qubits between specified locations | channel loss; remote entanglement rate and fidelity |
Think first Which DiVincenzo criterion is the main bottleneck for photonic qubits, and which for topological qubits?
Photonic: criterion 4. Photons barely interact, so entangling gates are probabilistic and built from measurements, and loss eats into every operation. Topological: criterion 1. A well-characterized topological qubit has not yet been unambiguously demonstrated. (Superconducting and ion platforms mostly fight criterion 1 at scale: wiring and variability for chips, crowded chains and slow gates for ions.)
Fidelity: how close is close enough?
To grade a gate, compare the operation that the hardware actually performs, a quantum channel, with the ideal unitary . For pure states the fidelity is the overlap ; for gates there are two standard averages.
The depolarizing channel: with probability the state is replaced by the maximally mixed state. Take .
The overlap is the same for every pure input, so it equals the average.
For a qubit : a state replaced by still passes the fidelity test half the time.
Acting on half of , whose reduced state is .
The theorem checks out.
A two-qubit gate has process infidelity . What is its average gate error ?
Show the answer
D.
with : . Always check which fidelity a paper quotes: for two qubits the two differ by 20%.
Decoherence and the coherence limit
Two processes dominate idle errors. Amplitude damping at rate relaxes . Pure dephasing at rate randomizes the phase without exchanging energy. Together they set the decay rate of the coherence :
Measured in a Ramsey experiment (two pulses separated by a wait), the coherence time is called and includes slow, quasi-static frequency fluctuations. A Hahn echo (a pulse halfway through the wait) refocuses those and gives the longer . Run the two standard experiments:
During a gate of duration , model the noise as / decay of the Bloch vector (the ideal rotation is undone in the comparison).
The fidelity of a pure input (Bloch vector , ) with the output.
Average over the sphere of pure states.
The exact result for this noise model.
To first order in . With this is .
A single-qubit gate lasts ns on a qubit with μs. What is its coherence-limited error , in units of ?
Show the answer
2.5 × 10⁻⁴
and , so . Measured errors above this limit come from control imperfections, leakage or crosstalk, which tells you what to fix next.
SPAM: when the ruler is bent
Every characterization experiment starts with state preparation and ends with measurement, and errors in either (SPAM errors) contaminate the result. For readout the basic object is the assignment matrix , often summarized by the assignment fidelity
Readout errors are usually asymmetric: a qubit in can relax during the measurement window, so .
Think first How could you see a gate error through a readout error?
Make the gate error accumulate while SPAM happens only once: apply the gate (or a sequence of gates) many times before measuring. After gates the gate contribution grows roughly like , while SPAM stays fixed, so fitting the rate of decay versus separates the two. That is the idea behind randomized benchmarking; gate set tomography takes the other route and estimates everything self-consistently.
Randomized benchmarking
Randomized benchmarking (RB) turns "repeat many times" into a robust protocol:
- Pick gates uniformly at random from the Clifford group.
- Append the one Clifford that inverts the whole sequence, so that the ideal circuit is the identity.
- Prepare , run the sequence, and record whether you measure again (the survival probability).
- Average over many random sequences, for several lengths , and fit the decay.
Single-qubit RB yields . What is the average error per Clifford?
Show the answer
A.
. To quote an error per native gate, divide by the average number of native gates per Clifford, which only makes sense if all native gates have similar errors.
import numpy as np
from scipy.optimize import curve_fit
# Sequence lengths and survival probabilities, averaged over random sequences
m = np.array([1, 10, 25, 50, 100, 200, 400, 800])
F = np.array([0.979, 0.971, 0.957, 0.934, 0.893, 0.822, 0.716, 0.597])
def decay(m, A, p, B):
return A * p**m + B
(A, p, B), cov = curve_fit(decay, m, F, p0=[0.5, 0.99, 0.5])
d = 2 # one qubit
r = (d - 1) * (1 - p) / d
print(f"p = {p:.4f}, error per Clifford r = {r:.1e}")Illustrative data following , , (rounded), so the fit should recover . Note that : the 2% shortfall at short lengths is SPAM, and it does not bias .
Holistic benchmarks, and going deeper
Gate-level numbers miss what happens when everything runs at once: crosstalk, compilation overheads, limited connectivity, drift. Holistic benchmarks run whole circuits instead.
| Benchmark | What it reports | Main caveat |
|---|---|---|
| Randomized benchmarking | average error per gate, free of SPAM | an average over a gate set; hides coherent-error structure, leakage, and crosstalk unless run simultaneously |
| Quantum volume | for the largest -qubit, depth- random circuits whose "heavy" outputs appear more than of the time | needs classical simulation of every circuit, so it cannot certify very large |
| Cross-entropy benchmarking (XEB) | fidelity of sampling from random circuits; about for Google's 53-qubit, 20-cycle circuits in 2019 | also needs classical simulation; can mislead when noise is not well described as uniform depolarization |
| Layer fidelity | fidelity of full layers of simultaneous two-qubit gates across many qubits | specific to the chosen layers and gate set |
Key takeaways
- DiVincenzo's five criteria (plus two for networking) are now quantitative: reset fidelity, coherence relative to gate time, gate and readout fidelities, and scalability.
- ; depolarizing noise with parameter has .
- , and decoherence alone limits a single-qubit gate of duration to .
- SPAM errors contaminate one-shot characterization; randomized benchmarking extracts a SPAM-free average error from , with .
- Holistic benchmarks (quantum volume, XEB, layer fidelity) capture crosstalk and compilation, but every single number hides structure.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
A platform has excellent gates, but its only readout destroys the entire register at once. Which criterion does it fail, and why does error correction care?
- Criterion 5 (qubit-specific measurement): error correction must repeatedly measure ancilla qubits mid-circuit without disturbing the data qubits
- Criterion 2 (initialization), because measuring resets the qubits
- Criterion 3 (coherence), because measurement causes decoherence
- None: any measurement suffices as long as it happens at the end
Show the answer
A. Criterion 5 (qubit-specific measurement): error correction must repeatedly measure ancilla qubits mid-circuit without disturbing the data qubits
Error correction extracts syndromes by measuring ancillas every cycle while the data qubits keep their encoded state. A readout that cannot address individual qubits, or that disturbs the rest, fails criterion 5.
The single-qubit depolarizing channel has average gate error
Show the answer
C.
for : a state replaced by still overlaps the input with probability .
A qubit has μs and μs. What is its pure-dephasing time ?
- μs
- μs
- μs
- about μs
Show the answer
D. about μs
per μs, so μs.
Why is the RB decay parameter insensitive to state-preparation and measurement errors?
- Because RB uses only Clifford gates, which cannot produce SPAM errors
- Because SPAM occurs once per sequence, independently of , so it changes only and
- Because the final inverting gate corrects SPAM errors
- Because RB measures in a random basis each time
Show the answer
B. Because SPAM occurs once per sequence, independently of , so it changes only and
The survival probability is with an imperfect and . Expanding gives , where and depend on and , but depends only on the gate noise.
Device A has 50 qubits with 1% two-qubit error; device B has 20 qubits with 0.1% error and all-to-all connectivity. Which has the larger quantum volume, and why?
- A, because quantum volume grows with qubit count
- Neither: quantum volume ignores error rates
- B, because quantum volume requires depth- circuits on qubits to succeed, and A's errors cap at a handful of qubits
- They tie, because both have more than 16 qubits
Show the answer
C. B, because quantum volume requires depth- circuits on qubits to succeed, and A's errors cap at a handful of qubits
An model circuit contains roughly two-qubit gates. With 1% error per gate, circuits wider than about 7 qubits lose too much fidelity, even before routing overheads; with 0.1% error, all 20 of B's qubits can pass. Quality beats quantity.
Go deeper
- Randomized Benchmarking · Robust average gate fidelities from random sequences.
- Quantum Volume, XEB & Holistic Benchmarks · Single numbers for whole machines, and their caveats.
- Tomography & Gate Set Tomography · Self-consistent characterization of noisy operations.
Further reading
- D. P. DiVincenzo, "The physical implementation of quantum computation", Fortschritte der Physik 48, 771 (2000).
- M. A. Nielsen, "A simple formula for the average gate fidelity of a quantum dynamical operation", Physics Letters A 303, 249 (2002).
- E. Magesan, J. M. Gambetta & J. Emerson, "Scalable and robust randomized benchmarking of quantum processes", Physical Review Letters 106, 180504 (2011).
- J. Eisert, D. Hangleiter, N. Walk, I. Roth, D. Markham, R. Parekh, U. Chabaud & E. Kashefi, "Quantum certification and benchmarking", Nature Reviews Physics 2, 382 (2020).