GPUにおける多倍長演算アルゴリズムとその高速化に関する研究

書誌事項

タイトル別名
  • ACCELERATION OF MULTI-PRECISION ARITHMETIC ALGORITHUM ON GPU

この論文をさがす

抄録

GPU is one of the parts in charge of image processing, and has a high parallel computing power when compared to need blank after CPU. GPGPU technique using a high computational power of GPU to general purpose computing have been studied. In this paper, we consider to accelerate operations on multi-precision integers used when dealing with large values exceeding 32 bits or 64 bits. We implemented a carry look-ahead addition algorithm, a carry save addition algorithm, and a redundant binary addition algorithm on the GPU. The implemented algorithm compares the MPIR of the multi-precision arithmetic library with the calculation time. As a result, multiple length addition was faster than MPIR under limited conditions, and with multiple length multiplication it was two to three times slower than MPIR.

収録刊行物

詳細情報 詳細情報について

  • CRID
    1390572174778232576
  • NII論文ID
    120006587725
  • NII書誌ID
    AA12677220
  • DOI
    10.15002/00021593
  • HANDLE
    10114/00021593
  • ISSN
    21879923
  • 本文言語コード
    ja
  • データソース種別
    • JaLC
    • IRDB
    • CiNii Articles
  • 抄録ライセンスフラグ
    使用可

問題の指摘

ページトップへ