Space and Time-Efficient Quantum Multiplier in Post Quantum Cryptography Era
Dedy Septono Catur Putranto, Rini Wisnu Wardhani, Harashta Tatimma Larasati, Howon Kim
Abstract
This paper examines the asymptotic performance of multiplication and the cost of quantum implementation for the Naive schoolbook, Karatsuba, and Toom-Cook methods in the classical and quantum cases and provides insights into multiplication roles in the post-quantum cryptography (PQC) era. Further, considering that the lattice-based PQC algorithm is based on polynomial multiplication algorithms, including the Toom-Cook 4-way multiplier as its fundamental building block, we propose a higher-degree multiplier, the Toom-Cook 8-way multiplier, which has the lowest asymptotic performance and implementation cost. Additionally, the designed multiplication will include additional sub-operations to complete the multiplication of large integers in order to prevent side-channel attacks. To design our Toom-Cook 8-way in detail, we employ detailed step computations such as splitting, evaluation, point-wise multiplication, interpolation, and recomposition, as well as several strategies to reduce space and time requirements. Existing asymptotic performance and quantum implementation cost multipliers are compared with our 2-way, 4-way, and 8-way Toom-Cook multiplier designs. Our Toom-Cook 8-way quantum multiplier has the lowest asymptotic performance analysis or qubit count in terms of space efficiency, with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> ( 15/8 ) <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">log 15 /(2 log 15–log 8) log<sub>8</sub> <i>n</i></sup> or asymptotically <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">O</i> ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1.245</sup> ). The design has the lowest logical Toffoli counts bound at 112 <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">log<sub>8</sub> 15</sup> – 128 <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> and Toffoli depth of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> ( 15/8 ) <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1– log 15 /(2 log 15–log 8) log<sub>8</sub> <i>n</i></sup> , asymptotically close to <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">O</i> ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1.0569</sup> ), which corresponds to a space- and time-efficient multiplication.