In addition to standard computer hardware, Big Blue’s quantum computing technique includes error avoidance. IBM is bullish on quantum computing but concedes there’s a long way to go.
Big Blue said during its Think! Program this week that it hopes to build realistic quantum-computing platforms with an intelligent software orchestration layer by 2025.
IBM plans to prototype quantum software solutions for specific use cases by 2019. Their first test case will be machine learning, and they will work with partners to define these services.
The main goal is to build a 4,000-qubit quantum computer made up of clusters of quantum processors. A 433-qubit Osprey processor is expected by the end of the year, followed by a 1,121-qubit Condor processor in 2023.
IBM envisions combining three 1,386 qubit Kookaburra multi-chip processors for 4,158 qubits. So IBM and its partners will need to design a lot of new software to regulate and connect these systems while avoiding faults that hinder quantum processing.
Researchers stated in a blog post that IBM’s goal is to build quantum-centric supercomputers. The quantum-centric supercomputer will mix quantum and classical processors, quantum communication networks, and classical networks to compute.
According to IBM, a serverless programming architecture that allows quantum and classical computers to work together without friction is needed to tackle the scaling difficulty of quantum processors. IBM plans to improve its Qiskit Runtime software, allowing users to create and manage quantum programs.
IBM expects to deliver direct Qiskit Runtime and cloud-based workflows to simplify and expand developer options in 2023. According to IBM, this serverless method will enable intelligent and efficient problem distribution across quantum and classical systems.
The business will also enable quantum processors to run in parallel within that timeframe. To construct a larger quantum processor, IBM will create short-range chip-level couplers.
IBM stated it will provide error mitigation and suppression approaches to Qiskit Runtime in 2024 and 2025 so clients may focus on improving quantum technology results. A future framework for quantum mistake correction will be built on these ideas.
Next year, the firm plans to start prototyping quantum software solutions for specific use cases, starting with machine learning. A quantum computer will be able to study machine learning, optimisation, natural sciences, and more by 2025.