Software moves into the quantum spotlight
The basic principles of quantum computing and the physical challenges that it faces are well known. What is less understood is the critical role that quantum software has to play in the development of this new paradigm for computing.
Currently, the state-of-the-art for quantum computers is about 20 quantum bits (qubits). At around 50 to 60 qubits (equivalent to an exascale supercomputer), quantum computers will begin to perform operations that are out of the reach of classical computing technology. But even if a research lab was able to develop a 1-million-qubit quantum computer, people would be unable to use it because the quantum software is simply not ready.
That is one of the reasons why Atos has launched the Atos Quantum Learning Machine, a quantum simulator, along with a universal programming language, aQasm (Atos Quantum Assembly Language).
Learning by doing
The main purposes of the Atos QLM are to provide training in quantum software for our industrial partners and customers and to support the development of quantum hardware.
Since we launched the Atos QLM, I have been surprised by the amount of requests we have received from researchers who are building qubits and who want to calibrate their hardware using our platform. Because the AQLM simulates and computes quantum perfectly, researchers can apply it to their qubits to see if their computer can execute algorithms correctly. The AQLM serves as a benchmarking platform as well as a simulator.
Top academic labs with experimental qubits are already working or have requested to work, with the Atos QLM to see just how their qubits behave in a real-world computational flow.
Software and simulation must guide research into quantum hardware. Hardware researchers need the inputs from software. The future of quantum depends partly on software-guided hardware.
Keep the noise down
For example, currently one of the most challenging roadblocks in achieving quantum superiority is the phenomenon of quantum noise. Single quantum particles are extremely sensitive to noise, which means they can break down too fast and lose their state of superposition before the computation is complete.
Competing quantum technologies have different sensitivities to noise. Designers can use the Atos QLM to understand the robustness of their software to noise and to help decide on which qubit technology their algorithm will perform best. It is a platform on which researchers can develop and optimize quantum software with a higher level of tolerance for errors.
Noise is a major hardware challenge, but quantum software is leading the way out of the woods.
Software in the driving seat
The Atos QLM is also heavily focused on generating and evaluating the possible opportunities for applications for quantum.
Applications means software. And when it comes to quantum software, it can never be too early to learn and to think about the potential opportunities for applications. For example, there will be real opportunities for algorithms on quantum computers to significantly outperform classical computers in areas such as artificial intelligence, chemistry and materials science, simulating large molecules to solve new problems. The AQLM has been designed from the very beginning to support quantum software research.
I am convinced that one day quantum computing will help power our supercomputers to a new level of performance. We do not know yet what shape the hardware will take. It is only by looking at the potential applications and opportunities, and at how quantum software performs on different technologies, that we will find the signals we need to build the next generation of quantum supercomputers. That is the mission of the Atos QLM.