Oak Ridge's LuGo algorithm cuts quantum gate counts by over 95 percent for fluid dynamics

[StarforgeSocket]

Scientists at Oak Ridge National Laboratory have introduced LuGo, an improved Quantum Phase Estimation framework aimed at quantum computational fluid dynamics, according to Quantum Computing Report. The approach uses classical preprocessing on Frontier, one of the world's most powerful supercomputers, before the quantum step. The result is "a greater than 95% reduction in total quantum gate operations." A circuit of about 2,000,000 logic gates shrinks to about 91,000

The test problem was the Hele-Shaw flow equation, which describes viscous fluid flowing between two closely spaced parallel plates. Fluid dynamics is one of the big hoped for applications of quantum computing, because simulating airflow, water flow and weather is extremely demanding for classical computers. Quantum Phase Estimation is a core quantum algorithm, but it normally needs deep circuits that today's noisy machines struggle with. Cutting the gate count tackles that problem directly

Interestingly, the preprocessing ran on a single Frontier node rather than the whole supercomputer. Through the Department of Energy's Quantum User Expansion for Science and Technology programme, LuGo was then run on several different machines. They included Quantinuum's H-1 trapped ion system, IBM's Marrakesh and Sherbrooke superconducting processors, and IQM's Garnet and Sirius processors. Testing across different hardware makes the result more convincing than a single vendor demonstration

The work has been published in Future Generation Computer Systems and won a 2026 R&D 100 Award. It fits a pattern we have seen this week with IQM's chemistry method, where classical computing does much of the heavy lifting and the quantum computer handles a smaller, more focused task. Supercomputing centres seem increasingly keen to position quantum computers as add-ons to their existing machines

Is fluid dynamics a realistic target for quantum computers, or one of the more overhyped applications? And does running on several vendors' hardware make results like this more trustworthy?

ORNL Develops LuGo Algorithm on Frontier Supercomputer, Reducing Quantum Circuit Gate Counts by 95% for Computational Fluid Dynamics - Quantum Computing Reportquantumcomputingreport.com

[Harry64]

Supercomputing centres positioning quantum as an accelerator makes a lot of sense. They already have the users, the scheduling systems and the cooling. Bolting a quantum processor on is easier than building a new kind of facility. Several European centres have already installed quantum machines alongside their supercomputers

[Bob92]

Running on Quantinuum, IBM and IQM hardware is the most impressive part for me. Vendor demos are easy to cherry pick. Cross platform results are much harder to fake. It also helps show where each type of hardware struggles, which is useful data for everyone

[RayOfLight]

Fluid dynamics is a long way off as a practical application. Real engineering simulations involve enormous grids and messy boundary conditions

A Hele-Shaw flow between plates is a sensible starting point, but it is a very simple case. Weather forecasting on quantum computers is a headline I expect to keep seeing for years before anything useful arrives