A new algorithm for efficient arithmetic in an optimal extension field is proposed. The new algorithm improves the speeds of multiplication, squaring, and inversion by performing two subfield multiplications simultaneously within a single integer multiplication instruction of a CPU. Our algorithm is used to improve throughputs of elliptic curve operations.

NTRU is a public key cryptosystem based on hard problems over lattices. In this paper, we present efficient methods for convolution product computation which is a dominant operation of NTRU. The new methods are based on the observation that repeating patterns in coefficients of an NTRU polynomial can be used for the construction of look-up tables, which is a similar approach to the sliding window methods for exponentiation. We provide efficient convolution algorithms to implement this idea, and we make a comprehensive analysis of the complexity of the new algorithms. We also give software implementations over a Pentium IV CPU, a MICAz mote, and a CUDA-based GPGPU platform. According to our analyses and experimental results, the new algorithms speed up the NTRU encryption and decryption operations by up to 41%.

Succinct representation is a space-efficient method to represent n discrete objects using space close to the information-theoretic lower bound. In order to directly access the ith object of succinctly represented data structures in constant time, two fundamental functions, rank and select, are commonly used. In this paper we propose two implementations supporting rank and select in constant time for non-compressed bit strings. One uses O(n log log n / ) bits of extra space and the other uses n+O(n log log n / log n) bits of extra space in the worst case. The former is rather a theoretical algorithm and the latter is a practical algorithm which works faster and uses less space in practice.