With a combination of intelligent software, connected architectures and new modular and networked processors, the company is aiming to reach near-term quantum advantage.
IBM on Tuesday announced it will expand its roadmap for achieving large-scale, practical quantum computing with new modular architectures and networking. This will give IBM’s quantum systems up to hundreds of thousands of qubits, the company said during its annual Think conference.
To enable qubits with the speed and quality necessary for practical quantum computing, they will be orchestrated by what the company characterized as “an increasingly intelligent software layer to efficiently distribute workloads and abstract away infrastructure challenges.”
According to IBM, achieving practical quantum computing will depend upon three pillars: Robust and scalable quantum hardware, cutting-edge quantum software to orchestrate and enable accessible and powerful quantum programs, and a broad global ecosystem of quantum-ready organizations and communities.
The company first announced its quantum roadmap in 2020, starting with “Eagle,” a 127-qubit processor with quantum circuits that cannot be reliably simulated exactly on a classical computer, and whose architecture laid the groundwork for processors with more qubits.
IBM has delivered a 120x speedup in quantum runtimes delivered via the Qiskit Runtime software platform. Later this year, IBM said it expects to continue unveiling its 433-qubit processor, Osprey.
In 2023, IBM will further its goals to build a frictionless development experience with Qiskit Runtime and workflows built right in the cloud — bringing a serverless approach to the core quantum software stack. The company said the goal is to give developers advanced simplicity and flexibility.
On the hardware front, IBM will introduce Condor, which the company is touting as the first universal quantum processor with over 1,000 qubits.
SEE: Artificial Intelligence Ethics Policy (TechRepublic Premium)
Rolling out modular quantum computing
IBM is targeting three regimes of scalability for its quantum processors.
The first requires building capabilities to classically communicate and parallelize operations across multiple processors. This opens the avenue to a broader set of techniques necessary for practical quantum systems, such as improved error mitigation techniques and intelligent workload orchestration, by combining classical compute resources with quantum processors that can extend in size.
The next step in delivering scalable architecture will deploy short-range, chip-level couplers. These couplers will closely connect multiple chips together to effectively form a single and larger processor and will introduce fundamental modularity that is key to scaling.
The third component to reaching true scalability will be to provide quantum communication links between quantum processors. To do so, IBM has proposed quantum communication links to connect clusters together into a larger quantum system.
All three scalability techniques will be combined to reach IBM’s 2025 goal: A 4,000+ qubit processor built with multiple clusters of modularly scaled processors.
Building the fabric of quantum-centric supercomputing
In tandem with its hardware updates, IBM is targeting software milestones to improve error suppression and mitigation. These techniques are currently progressing the ability of quantum software to minimize the effect of noise on the users’ application and are paving the path toward the error-corrected quantum systems of the future.
Earlier this year, IBM launched ready-made primitive programs designed to minimize infrastructure requirements and allow developers to access and build larger quantum programs more easily. In 2023, IBM plans to expand these primitives, with capabilities that allow developers to run them on parallelized quantum processors thereby speeding up the user’s application.
“With quantum serverless operations and the advances in hardware, software, and theory … we will usher in an era of quantum-centric supercomputers that will open up large and powerful computational spaces for our partners and clients,” said Dario Gil, senior vice president and director of research at IBM, in a statement.
The company said this will help it deliver quantum serverless into its core software stack in 2023, which will enable developers to easily tap into flexible quantum and classical resources. Quantum serverless will also lay the groundwork for core functionality within IBM’s software stack to intelligently trade-off and switch between elastic classical and quantum resources, forming the fabric of quantum-centric supercomputing, the company said.
The new systems will be designed to work within IBM Quantum System Two, which IBM said will provide the infrastructure needed to link together multiple quantum processors. The prototype of the Quantum System Two is expected to be ready in 2023.