The evaluation criteria for the Advanced Encryption Standard (AES) Round1 candidate algorithms, as specified in the Request for Comments (1), includes computional efficiency, among other criteria. Specifically, the Call For AES Candidate Algorithms (2) required both Reference ANSI1 C code and Optimized ANSI C code, as well as JavaTM2 code. Additionally, a reference hardware and software platform was specified for testing. NIST performed testingon this reference platform, as well as several others. Candidate algorithms were tested for computational efficiency using the Optimized ANSI C source code provided by the submitters.This paper describes the testing methodology used in ANSI C efficiency testing along with observations regarding the resulting measurements. Conclusions are provided regarding which algorithms have the most consistent performance across different platforms. This paper also includes an appendix containing tables of timing and cycle counting values obtained from testingthe algorithms. Some knowledge regarding compilation and processor architectures is useful in understanding how the data was derived. However, the raw data in the document can also be useful without necessarily understanding how it was derived.
Citation: NIST Interagency/Internal Report (NISTIR) - 6391
NIST Pub Series: NIST Interagency/Internal Report (NISTIR)
Pub Type: NIST Pubs
advanced encryption standard, ANSI C, cycle count, efficiency, performance, round1, timing