Three block ciphers are considered to determine how well they can be
implemented on existing superscalar architectures such as the Intel Pentium.
An examination of the Pentium architecture suggests that substantial
performance increases can be achieved if particular rules are followed.
Software libraries are written in high-level C language and low-level assembly
language to produce a package of routines which achieve a near optimal
performance level on a current processor architecture. The structure of each
algorithm is studied to determine if it is possible to alternatively implement the
algorithm such that certain steps are reordered or reduced. Using the Intel
MMX architectural advances, it is observed that one algorithm benefits
dramatically from a new implementation that takes advantage of MMX
strengths. / Graduation date: 1999
| Identifer | oai:union.ndltd.org:ORGSU/oai:ir.library.oregonstate.edu:1957/33737 |
| Date | 23 November 1998 |
| Creators | Sessions, Julian Brently |
| Contributors | Koc, Cetin K. |
| Source Sets | Oregon State University |
| Language | en_US |
| Detected Language | English |
| Type | Thesis/Dissertation |
Page generated in 0.0024 seconds