How should problem-specific compilation be counted in the Q-CTRL Fermi-Hubbard quantum advantage claim?

How should problem-specific compilation be counted in the Q-CTRL Fermi-Hubbard quantum advantage claim? Ask Question Asked 4 days ago Modified 3 days ago Viewed 41 times 3 I am trying to understand how to fairly interpret the recent Q-CTRL Fermi-Hubbard practical quantum advantage claim https://arxiv.org/abs/2605.04025.My question is not whether the experiment is impressive. It is. What I am asking is how we should count the problem specific co design that makes the experiment work.The paper's performance seems to rely on several pieces of structure that are not generic black-box compilation: a pair-interleaved Jordan-Wigner ordering, a fermionic fSWAP network, with fSWAP = SWAP * CZ, a snake like embedding of the effective fermion line into the heavy-hex device, calibration aware layout selection and bad-qubit avoidance, a native circuit structure with very low two-qubit depth, and a hardware specific error suppression and mitigation stack. According to the paper this is the right thing to do for this problem. But it also makes the benchmarking question more subtle. Is the claimed advantage mainly an algorithmic advantage, a hardware execution time advantage, or a compiler/hardware/physics co-design advantage? I built a much smaller "poor man's" version to test the basic idea: https://github.com/BramDo/fermi-hubbard-60q-tdvp This is not a reproduction of the full paper. It is only: 60 qubits / 30 sites, 8 Trotter steps, dt = 0.2, total time t = 1.6, IBM hardware, native fswap,cz circuit family, local observables: occupation, charge, spin, double occupancy, comparison against ITensorMPS TDVP at chi = 64. Even at this smaller scale, the circuit can be run directly on IBM hardware and the speed separation is already easy to see. My local ITensorMPS TDVP chi64 reference took 821.277 seconds on a laptop/desktop CPU setup. The corresponding IBM main job used about 3 jobs of 2 quantum seconds. This is an order 100 QPU-execution-time separation for a simplified version of the same workflow. The best hardware route in my small run is DD + Pauli twirling + readout-corrected local observables. Against the TDVP chi64 reference I get: occupation RMSE: 0.02061, charge RMSE: 0.01471, spin RMSE: 0.03850. I am not calling my run quantum advantage. The reference is chi64, not chi4096; the run is 60q/8, not the paper's 120q/30 case; and end-to-end timing including classical preprocessing, calibration search, queueing, mitigation, and post-processing has to be treated carefully. But the small run makes the main issue concrete: a q qauntum enthousiast can now reproduce a simplified version where the quantum processor executes the relevant circuit much faster than a tensor-network calculation on a local machine, provided the fermionic mapping, fSWAP layout, transpilation choices, and mitigation stack are already packaged correctly. Some thoughts or questions to think over: For a practical quantum advantage claim in this setting, should the fSWAP/snake-layout construction be counted as part of the quantum algorithm, part of the compiler, or part of the benchmark setup? If a Qiskit Function / Fire Opal style service automates some of this, what information must be exposed for the result to be independently verifiable? For example, should a verification package include the exact fermion to qubit ordering, logical to physical layout, bad qubit avoidance. Main question More important, is this automation even possible for different sort of problems. It seems you need to know the specifics or the circuit to get the best result. quantum-algorithmshamiltonian-simulationquantum-advantage Share Improve this question Follow edited May 12 at 17:48 Martin Vesely 15.6k4 4 gold badges 34 34 silver badges 77 77 bronze badges asked May 12 at 13:32 Bram 1,6199 9 silver badges 10 10 bronze badges Add a comment Know someone who can answer? Share a link to this question via email, Twitter, or Facebook. Your Answer Sign up or log in Sign up using Google Sign up using Email and Password Post as a guest Name Email Required, but never shown Post Your Answer By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy. Start asking to get answers Find the answer to your question by asking. Ask question Explore related questions quantum-algorithmshamiltonian-simulationquantum-advantage See similar questions with these tags. The Overflow Blog Connecting the dots for accurate AI Observability and human intuition in an AI world Featured on Meta (Almost) One year of Challenges Related 7 Does a fully working quantum computer solve a specific problem at varying speeds every time? 10 Grover algorithm for a database search: where is the quantum advantage? 23 What does Google's claim of "Quantum Supremacy" mean for the question of BQP vs BPP vs NP? 3 What is the problem in demonstrating quantum supremacy? 11 What is the advantage of quantum machine learning over traditional machine learning? 1 What is the main advantage of using the Variational Quantum Eigensolver over a classical algorithm? 6 Proven quantum advantage (in the algorithmic sense) without error correction (for specific algo, or noise models) 1 Problem with the mathematical definition of the eigenvalue algorithm on a specific exercise 3 What is the intuition behind achieving Quantum advantage in simulating non-hermitian dynamics using Quantum computer? 7 What is the relation between magic states, contextuality and quantum advantage? Hot Network Questions Where does the "group" come from in the naming of quantum groups? Cyclotron radiation of a charged solid Bash/jq hydrate() — transparent JSON-to-bash type converter Regarding Dawid's defense that it is reasonable to accept string theory as credible without experimental evidence Is there a way to prove the quadratic formula by direct substitution? Which languages has Sir Terry Pratchett been translated into? Clear and Keep Transform behavior if parent object is also parented Does the number of eyes change a race's vision? is it wrong to say "I have one pence"? A criterion of cyclic group nfs4_setfacl fails with username@domain (AD/SSSD) even though nfs4_getfacl prints ACEs in that format (Debian 13, Kerberos, OneFS) Does the sunk costs fallacy require an unprofitable outcome? Structural elucidation of olefin from hydrogenation and oxidative cleavage What is an appropriate role for LLMs in early mathematical research training? How can I sell old shares? How to shrink with a certain amount of offset? What type of domain should a function have to define its derivative? Rear hub can be separated into left and right part, can it be fixed? Does the Doppler effect apply to frequency of pitch modulation as well? SQL Server 2017 Synchronous Database Mirroring Latency Caused Principal DB to Hang and Application Timeouts Why does a resistor physical temperature of 290 K make it emit -174 dBm/Hz of noise power, but other circuit elements at 290 K don't emit -174 dBm/Hz? Is There A Limit For After Tax 401K Contributions? Books with multiple proofs of one statement How to turn "Alt+l+e+up-key+enter" combination into a shortcut? Question feed By continuing to use this website, you agree Stack Exchange can store cookies on your device and disclose information in accordance with our Cookie Policy. By exiting this window, default cookies will be accepted. To reject cookies, select an option from below. Customize settings Cookie Consent Preference Center When you visit any of our websites, it may store or retrieve information on your browser, mostly in the form of cookies. This information might be about you, your preferences, or your device and is mostly used to make the site work as you expect it to. The information does not usually directly identify you, but it can give you a more personalized experience. Because we respect your right to privacy, you can choose not to allow some types of cookies. Click on the different category headings to find out more and manage your preferences. Please note, blocking some types of cookies may impact your experience of the site and the services we are able to offer. Cookie Policy Accept all cookies Manage Consent Preferences Strictly Necessary Cookies Always Active These cookies are necessary for the website to function and cannot be switched off in our systems. They are usually only set in response to actions made by you which amount to a request for services, such as setting your privacy preferences, logging in or filling in forms. You can set your browser to block or alert you about these cookies, but some parts of the site will not then work. These cookies do not store any personally identifiable information. Targeting Cookies Targeting Cookies These cookies are used to make advertising messages more relevant to you and may be set through our site by us or by our advertising partners. They may be used to build a profile of your interests and show you relevant advertising on our site or on other sites. They do not store directly personal information, but are based on uniquely identifying your browser and internet device. Performance Cookies Performance Cookies These cookies allow us to count visits and traffic sources so we can measure and improve the performance of our site. They help us to know which pages are the most and least popular and see how visitors move around the site. All information these cookies collect is aggregated and therefore anonymous. If you do not allow these cookies we will not know when you have visited our site, and will not be able to monitor its performance. Functional Cookies Functional Cookies These cookies enable the website to provide enhanced functionality and personalisation. They may be set by us or by third party providers whose services we have added to our pages. If you do not allow these cookies then some or all of these services may not function properly. Cookie List Clear checkbox label label Apply Cancel Consent Leg.Interest checkbox label label checkbox label label checkbox label label Necessary cookies only Confirm My Choices