arXiv:2610.04731
Sum-of-squares compilation of matrix-model dynamics for NISQ and fault-tolerant hardware
Ayanendu Dutta, Nabanita Sarkar
Abstract
The SU$(N)$ matrix-model potential $V=-\tfrac{g^2}{4}\sum_{I,J}Tr[X_I,X_J]^2$, the interaction of the BFSS/BMN family, whose thermal large-$N$ states are dual to black holes, is diagonal in the coordinate basis and is compiled in the literature as a phase polynomial with $O(d^2N^4Q^4)$ rotations per Trotter step. That cost is close to optimal for its gate class: we show that for a generic Trotter step the Walsh support $m_{\min}$ lower-bounds both the rotation count and the two-qubit count of every circuit over $\{CNOT,CZ,CPhase,R_{ZZ},SWAP,X,R_z\}$, ancillas included, and we build a parity network attaining $1.01$--$1.10\,m_{\min}$. Improvement therefore requires a gate outside that class, and the exact factorisation of $V$ into a sum of squared commutator entries supplies one in each era. For NISQ, Hadamard: each squared entry is added into a Fourier register and squared there, $Θ(d^2N^3Q^3)$ two-qubit gates. For fault tolerance, $T$: reversible arithmetic with one phase-gradient injection per squared entry, $Θ(d^2N^3Q^2)$ T gates and no arbitrary-angle rotation. Both are built and verified. With a native $R_{ZZ}$ the NISQ circuit uses fewer two-qubit gates than every parity network from $Q=6$ at SU(2) and $Q=2$ at SU(5); the oracle beats a precision-matched Hamming-weight-phasing incumbent at every point built except the smallest, by up to $11.9\times$ in T count and, at equal width, $25.5\times$ in surface-code spacetime. An ablation with the same arithmetic but no factorisation costs up to $21\times$ more and loses to that incumbent at $15$ of $17$ points, so the factorisation, not the arithmetic, produces the win. Both constructions survive the supersymmetric deformation to mini-BMN with fermions, and complete Trotter evolutions on a statevector simulator reproduce the exact product formula to $2\times10^{-12}$.
AI Summary
The paper studies the cost of compiling coordinate-grid Trotter steps for the commutator-square potential of BFSS/BMN-type matrix models. It uses the exact sum-of-squares factorisation into commutator entries rather than expanding the potential into quartic monomials. For generic step angles, it proves that the nonconstant Walsh support $m_{\min}$ lower-bounds rotations in CNOT–phase circuits and, when weight-one characters are absent, two-qubit gates in an extended diagonal/affine gate class, including clean ancillas. For $\mathrm{SU}(2)$ with two matrices, $m_{\min}=\frac92Q^4-3Q^3+\frac92Q^2-3Q$. A constructed parity network approaches this floor within 1–10% for $Q\ge4$. Two constructions leave that gate class: Fourier-register evaluation and quadratic phasing use $\Theta(d^2N^3Q^3)$ two-qubit gates; reversible arithmetic with phase-gradient kickback uses $\Theta(d^2N^3Q^2)$ T gates, versus the expanded compilation’s $\Theta(d^2N^4Q^4)$ scaling. Native-$R_{ZZ}$ counts beat the parity-network floor from $Q=6$ at $\mathrm{SU}(2)$ and $Q=2$ at $\mathrm{SU}(5)$. Built fault-tolerant examples achieve up to $11.9\times$ lower T counts than precision-matched Hamming-weight phasing. Gate-count checks, monomial-evaluation ablations, basis-state tests and small-system statevector evolutions support the comparisons; a mini-BMN extension includes fermions. Limitations include substantial ancilla requirements, NISQ crossovers beyond present circuit budgets, dyadic-step assumptions for exact fault-tolerant phases, and heuristic routing and surface-code estimates. Large-$N$ fermion layers, full BFSS/BMN circuits and truncation requirements remain unresolved; end-to-end simulations are noiseless and restricted to the smallest grid.
Error found: The abstract's uniform Θ(d²N³Q³) NISQ and Θ(d²N³Q²) fault-tolerant costs omit register widths growing as Θ(Q + log N), requiring logarithmic corrections or restrictions on N relative to Q. For the stated constructions at fixed Q, the register-width overhead changes the claimed linear-in-N advantage to asymptotic growth proportional to N/log N. Equation (leading) omits n_22/4 + n_112/2 from the leading Q⁴ Walsh-support coefficient, despite claiming exact N dependence.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 39 pages, 15 figures, 22 tables