IBM Corp., Advanced Micro Devices Inc. and the Quantum Science Center at Oak Ridge National Laboratory today announced a new quantum error-correction platform that integrates quantum and high-performance computing.
Gaining high performance and correctness in quantum computing requires protecting quantum information from noise. In many cases, researchers have paired quantum computers with classical computers to achieve superior performance for running industry-relevant problems. The challenge today is scaling up these advances with classical computing.
Today’s physical quantum qubits make errors roughly once every few hundred operations. Without correction, errors compound rapidly, rendering long or complex calculations completely untenable. Game-changing algorithms such as Shor’s algorithm for factoring or for molecular-scale simulations in chemistry require billions of error-free operations.
This launch uses a co-design platform for quantum error correction using IBM’s quantum computers, including the 120-qubit Nighthawk with ORNL’s Umbra High-Performance Computing system powered by AMD’s Instinct MI355X graphics processing units and 5th Gen EPYC central processing units.
“ORNL’s strength in high-performance computing, combined with the expertise of our QSC partners – like IBM and AMD – creates an opportunity to push the boundaries of what is possible,” said Director of the Quantum Science Center Travis Humble. “With Umbra, we will now test new ideas, evaluate approaches at scale, and build the foundation for more reliable quantum computers.”
Jay Gambetta, an IBM Fellow and director of IBM Research, added that the biggest contributor to today’s research is no longer whether error correction will work on quantum hardware, but how to accelerate and optimize it to build systems scientists can use to tackle problems.
IBM said the Nighthawk’s square-lattice connectivity enables industry-leading speed and real-time control to test error-correction techniques. GPU accelerators, such as AMD’s nodes in the Umbra system, can help with this fine-tuning. The co-design platform connected at Oak Ridge will help explore quantum as part of a broader artificial intelligence and scientific computing infrastructure.
“This new co-design platform for quantum error correction allows us to evaluate different approaches and determine the architectures, software and workflows that will scale,” said Gambetta.
Researchers can measure qubits mid-calculation and act on the results in real time, set the exact timing for the signals that protect the qubits from noise, and see exactly what the hardware ran while working with raw measurement signals. This will provide error-correction research with the raw data needed for the next generation of approaches to build sophisticated capabilities for large systems running real code.
Earlier this year, IBM and ORNL unveiled results showing how workloads running on both quantum and classical computers and the Frontier supercomputer can simulate real magnetic materials and calculate molecular configurations of a promising material for fusion fuel.
The simulation of magnetic materials could lead to breakthroughs in superconductors, more efficient batteries and novel drugs. A promising fusion fuel could provide a stepping stone to abundant energy superior to nuclear fission, which powers current nuclear power plants.