Analysing Quantum Systems with sys-sage

sys-sage is part of an HPC software stack. The library processes data from supercomputers—such as information about processors and their performance, data transfer between compute nodes, and many other metrics used to manage these systems or improve code and algorithms. The tool delivers both static and dynamic data related to the architecture and operation of an HPC system.

However, these systems are becoming increasingly heterogeneous. In addition to graphics processing units (GPUs) and other accelerators, quantum computers are now being integrated: they are expected to accelerate specific computations and contribute new methods of data processing. The challenge, however, is that quantum computers function fundamentally differently from classical ones. They rely on the principles of quantum mechanics to perform calculations – for example, through quantum bits (qubits) that can become entangled or exist in a state of superposition. This enables them to solve specific yet important computation-heavy problems in fields such as chemistry, mathematics, or cryptography.

Moreover, there are various quantum technologies, including quantum computers based on superconducting circuits, ion traps, neutral atoms, or photons. Control and management mechanisms for these are still lacking – another reason for integrating them into supercomputers and managing them through these systems. To address this, the award-winning team extended sys-sage to include components for quantum computing and various quantum technologies with data paths and relationships that account for properties of quantum hardware. With this enhanced library, it is possible to, for instance, filter out the quality of qubits and quantum gates – so that only the connections that are accurate and largely noise-free are used for computation.

“Superconducting qubits, those based on ion traps, or neutral atoms each have specific limitations and characteristics. sys-sage characterizes these in a way that allows us to use them for computation even to enhance their practical accuracy,” explains Burak Mete, an algorithm specialist at LRZ who has already developed a compiler for integrated super and quantum computers. “sys-sage is a module that analyses the properties of qubits, optimises quantum algorithms for quantum backends, and generally improves the accuracy of quantum computations.”

The award-winning team has been collaborating for some time, complementing each other with strengths, experience, and knowledge, and dividing up their responsibilities for the Munich Quantum Software Stack (MQSS) accordingly. While Burak Mete focused on developing metrics for assessing qubit quality, Jorge Echavarria supported the team in working within the test environment, and Xialong Deng contributed his insights into neutral atom technologies. Just like compilers – which translate programming languages into machine code—or schedulers, which manage the scheduling of computational tasks, the extended sys-sage is part of MQSS. The stack is being developed as part of the Q-DESSI research project within the Munich Quantum Valley (MQV), involving TUM, LRZ, and other institutions and partners. “Although it’s still in an experimental stage,” says Jorge Echavarria, "it has already positioned itself as one of the most efficient software stacks worldwide. Its capabilities became evident when we achieved, for the first time in Germany, true HPC-QC integration by connecting a supercomputer with a 20-qubit quantum system from IQM based on superconducting circuits."

Now that the LRZ also operates quantum systems using an ion trap from AQT, neutral atom technology from planqc is announced, so the MQSS – and especially the new sys-sage library – plays an increasingly vital role. In pilot phases, the integrated HPC-QC system and other quantum technologies are already being made available to initial users. These users test, compare, and analyse the systems, suggest improvements, or begin developing early-stage software. “The most mature part of MQSS is certainly the interface for managing quantum hardware,” says Jorge Echavarria referencing the Quantum Device Management Interface (QDMI), the core element of the MQSS for the integration of divers quantum technologies. “We’re asking early users for feedback – they can let us know what such a system still needs.” This iterative approach allows both hardware and software to be gradually improved and made more practical for scientific use. But the team has even bigger ambitions: “The MQSS is connected to every quantum system we have at LRZ,” says Burak Mete. “It’s a complete software stack that we already use, but we aim to develop it further so that any HPC center can implement and adopt it.” (vs | LRZ)

Photo Credit: © ISC High Performance | Philipp Loeper