Mixed computing unlocks new boundaries

Scientists have demonstrated how a method of combining quantum computers with supercomputers solves complex molecular problems, neither of which can be solved separately, potentially advancing drug design and materials science.

Researchers at the Cleveland Clinic have successfully used hybrid quantum classical calculation methods to calculate the ground state energy of a molecule – a fundamental property that determines the stability and behavior of a molecule. These findings, published in the Journal of Chemical Theory and Computing, mark the progress of the practical application of extended quantum computing in chemistry.

“By combining the power of quantum computers with the error correction capabilities of supercomputers, we can begin to simulate and predict the behavior of molecules, enhancing our ability to understand and treat diseases,” explains Dr. Kenneth Merz.

Divide unsolvable items into manageable items

The team’s approach uses a method called density matrix embedding theory (DMET), which breaks down large and complex molecules into smaller fragments that can be analyzed more efficiently. The fragments were processed using a technique called sample-based quantum diagonalization (SQD) used on the IBM quantum system at the Cleveland Clinic.

Quantum computers perform calculations on the electronic configuration of each molecular fragment, while supercomputers process error corrections and combine results to determine the overall properties of the molecule.

• Compared to running the entire simulation on a quantum computer, this method greatly reduces the required number of buildings
• Researchers tested the method on various conformations of 18 atoms of hydrogen rings and cyclohexane
• The hybrid system accurately predicts the relative stability of different molecular structures

Overcome the current quantum limitations

Current quantum computers face significant limitations in error correction and QUIT count. This hybrid approach takes advantage of each system: the ability of quantum computers to effectively explore multiple electronic configurations simultaneously and the accuracy and reliability of supercomputers.

“Current state quantum computers are very powerful, but have no error correction yet,” Dr. Merz noted. The team’s approach works around these limitations while still taking advantage of the unique advantages of quantum computing.

From simple rings to complex proteins

Although the current research focuses on relatively simple molecules, the researchers believe their method can eventually scale to analyze more complex biological molecules. Accurate modeling of proteins and other large biomolecules can accelerate drug discovery and gain a deeper understanding of biological processes.

The researchers’ method achieves subkilocalorie in predicting energy differences between cyclohexane conformations, an impressive accuracy of the quantum-based method. This accuracy is critical for applications like drug-bound prediction, where small energy differences can determine whether a potential drug is effective.

The study shows how scientists pave the way for advances in computational chemistry by combining them with classical computing resources, neither approach can be achieved independently.

Related