A recent study conducted by the Jülich research center in Germany compared 19 quantum processors from five manufacturers. The tests measured the stability and reliability of these systems for high-performance computing. The results, available on
arXiv, show varying performance across models and configurations.
IBM processors stood out for their ability to handle deep quantum circuits. Quantinuum, on the other hand, excelled in tests involving a large number of qubits. These performances reflect recent technological advances, particularly in hardware and software.
The study also reveals that IBM's latest model, named Marrakesh, didn't achieve the expected improvements. Despite reducing quantum gate errors, its performance remained similar to its predecessor. This result highlights persistent obstacles in quantum technology development.
Quantinuum scored an important point with its H2-1 processor, capable of handling 56-qubit problems. This performance surpasses classical computers' capabilities in certain situations. Researchers used a specific algorithm to evaluate this capability, demonstrating the practical potential of quantum computers.
The benchmark developed by the team is based on the MaxCut problem, known for its computational complexity. This method allows standardized and cost-effective evaluation of different systems' performance. The obtained results provide valuable insights into the current state of quantum technology.
Despite its advantages, the benchmark has limitations, such as the absence of dynamic parameter adjustments during computation. Scientists call for developing additional complementary tests to better assess quantum processors' performance. This pluralistic approach could accelerate progress in this field.
The study's results open interesting perspectives for the future of quantum computing. They highlight progress made by various industry players while identifying necessary improvement areas to achieve industrially usable quantum supremacy.
How does the MaxCut benchmark evaluate quantum processors?
The MaxCut problem is an optimization problem used to test quantum processors' performance. It involves dividing a graph into two subsets to maximize the number of edges between them. This problem is chosen as a benchmark because it's difficult for classical computers to solve and its complexity can be adjusted by changing the graph size.
The MaxCut benchmark measures a quantum processor's ability to execute quantum circuits of different depths and widths. Depth refers to the number of successive operations, while width corresponds to the number of qubits used. These two parameters are essential for evaluating a quantum system's power and flexibility.
A system fails the test when its results become indistinguishable from those produced by a random generator. This threshold determines how far a quantum processor can provide meaningful results before noise and errors take over.
The MaxCut benchmark is particularly useful because it's simple to implement and scalable. It can be adapted to a wide range of quantum systems, making it a valuable tool for comparing different quantum technologies' performance. However, like any benchmark, it has limitations and should be complemented by other tests for a complete evaluation.
What is a qubit and why is it fundamental in quantum computing?
A qubit is the basic unit of quantum information, analogous to a classical bit. Unlike a bit that can be either 0 or 1, a qubit can exist in a superposition state, meaning a combination of 0 and 1 simultaneously. This property allows quantum computers to process exponentially more information than classical computers for certain tasks.
Quantum entanglement is another key property of qubits. When two qubits are entangled, one's state instantly influences the other, regardless of the distance between them. This phenomenon enables the creation of extremely powerful quantum circuits for solving complex problems.
However, qubits are also very fragile. They're sensitive to their environment, which can cause calculation errors. Current research focuses on developing more stable qubits and error correction methods to make quantum computers more reliable.
Advances in qubit manipulation, such as the introduction of fractional gates, have significantly improved quantum processors' performance. These breakthroughs are crucial for realizing quantum computing's potential in fields like cryptography, molecular modeling, and optimization.