日本語フィールド
著者：Yoshiyuki Yamashita and Masato Tsuru題名：Software Pipelining for Packet Filters発表情報：Proc. 3rd Int. Conf. on High Performance Computing and Communication (HPCC07), Lecture Notes in Computer Science (LNCS) 巻: 4782 ページ: 446-459キーワード：Code Optimization, Packet Filters概要：パケットフィルターにコード最適化手法のひとつであるソフトウェアパイプライン化を適用する手法を開発し、処理速度に関する実験を行った、結果、tcpdumpインタプリタを約20倍高速化できた。抄録：Packet filters play an essential role in traffic management and
security management on the Internet. In order to create software-based
packet filters that are fast enough to work even under a DOS attack, it is
vital to effectively combine both the higher-level optimization related to
algorithmic structure and the lower-level optimization related to acceler-
ation techniques in compiler study. In the present paper, we focus on the
lower-level (machine code) optimization using software-pipelining, and
report experimental results that indicate the potential of our approach
for accelerating packet filter performance. The technical difficulty is that
the packet filter is a lump of conditional branches, so that standard opti-
mization techniques usually applied to basic blocks is not directly appli-
cable to this problem. Using predicated execution and enhanced modulo
scheduling, we solve this problem and achieve 20 times higher perfor-
mance compared with a conventional interpreter-based packet filter. We
also compare the proposed filters and compiler-based packet filters, and
obtain a better than two-fold increase in performance.英語フィールド
Author：Yoshiyuki Yamashita and Masato TsuruTitle：Software Pipelining for Packet FiltersAnnouncement information:Proc. 3rd Int. Conf. on High Performance Computing and Communication (HPCC07), Lecture Notes in Computer Science (LNCS) Vol: 4782 Page: 446-459Keyword：Code Optimization, Packet FiltersAn abstract：Packet filters play an essential role in traffic management and
security management on the Internet. In order to create software-based
packet filters that are fast enough to work even under a DOS attack, it is
vital to effectively combine both the higher-level optimization related to
algorithmic structure and the lower-level optimization related to acceler-
ation techniques in compiler study. In the present paper, we focus on the
lower-level (machine code) optimization using software-pipelining, and
report experimental results that indicate the potential of our approach
for accelerating packet filter performance. The technical difficulty is that
the packet filter is a lump of conditional branches, so that standard opti-
mization techniques usually applied to basic blocks is not directly appli-
cable to this problem. Using predicated execution and enhanced modulo
scheduling, we solve this problem and achieve 20 times higher perfor-
mance compared with a conventional interpreter-based packet filter. We
also compare the proposed filters and compiler-based packet filters, and
obtain a better than two-fold increase in performance.