pygridsynth: A fast numerical tool for ancilla-free Clifford+T synthesis

Shuntaro Yamamoto, Nobuyuki Yoshioka (Apr 24 2026).

Abstract: We present pygridsynth, an open-source Python library for ancilla-free approximate Clifford+$T$ synthesis that runs in $O(\log(1/\epsilon))$ for precision $\epsilon$. For $n=1, 2$ qubits, the library builds upon established efficient and high-precision synthesis routines, such as nearly optimal $Z$-rotation synthesis and magnitude approximation. For $n\ge 3$ qubits, we introduce a partial-decomposition technique that generalizes the magnitude approximation, reducing constant factors in the $T$-count as $(\frac{21}{8}\cdot 4^n - \frac{9}{2}\cdot 2^n + 9)\log_2(1/\epsilon) + o(\log(1/\epsilon))$. The package also exposes a mixed-synthesis workflow that approximates target unitary channels by probabilistic mixtures of Clifford+$T$ circuits, for which we empirically find that the synthesis error is reduced from $\epsilon$ to $\epsilon^2/(2n)$. Taken together, these features make pygridsynth a Python-native platform for high-precision Clifford$+T$ synthesis and for benchmarking unitary and mixed synthesis strategies on multi-qubit instances.

Arxiv: https://arxiv.org/abs/2604.21333