상세정보
학술지
Subquadratic Space Complexity Multiplier Using Even Type GNB Based on Efficient Toeplitz Matrix-Vector Product
- 저자
- 박선미, 장구영, 홍도원, 서창호
- 발행일
- 201812
- 출처
- IEEE Transactions on Computers, v.67 no.12, pp.1794-1805
- ISSN
- 0018-9340
- 출판사
- IEEE
- 협약과제
- 18ZH1200, 데이터 안심사회를 위한 트러스트 데이터 커넥톰 핵심 원천 기술 개발, 박종대
- 초록
- Multiplication schemes based on Toeplitz matrix-vector product (TMVP) have been proposed by many researchers. TMVP can be computed using the recursive two-way and three-way split methods, which are composed of four blocks. Among them, we improve the space complexity of the component matrix formation (CMF) block. This result derives the improvements of multiplication schemes based on TMVP. Also, we present a subquadratic space complexity GF(2m) multiplier with even type Gaussian normal basis (GNB). In order to design the multiplier, we formulate field multiplication as a sum of two TMVPs and efficiently compute the sum. As a result, for type 2 and 4 GNBs, the proposed multipliers outperform other similar schemes. The proposed type 6 GNB is the first subquadrtic space complexity multiplier with its explicit complexity formula.
- KSP 제안 키워드
- Field multiplication, Gaussian normal basis, Space Complexity, Three-way, Toeplitz matrix-vector product, Type 2, matrix formation, two-way