Speed Record of AES-CTR and AES-ECB Bit-Sliced Implementation on GPUs

  • Lee, Wai-Kong; 
  • Seo, Seog Chung; 
  • Seo, Hwajeong; 
  • Kim, Dong Cheon; 
  • Hwang, Seong Oun
Citations

WEB OF SCIENCE

2
Citations

SCOPUS

4

초록

The advanced encryption standard (AES) has been widely used to protect digital data in various applications, such as secure IoT communication, files encryption, and pseudorandom number generation. The efficient implementation of AES on parallel architecture, such as graphics processing unit (GPU), has attracted considerable interest over the past decade. These prior studies mainly implemented the AES electronics code book (ECB) and counter (CTR) mode using the table-based approach. In this brief, we set a new speed record of AES-ECB and AES-CTR on GPU based on the proposed bit-sliced implementation techniques. Our implementation achieved 2.6% (ECB) and 9% (CTR) faster than the state-of-the-art table-based implementation on a RTX3080 GPU. Our work evaluated on an embedded GPU (Jetson Orin Nano) can also achieve high throughput at 60 Gb/s, which is 1.9% (ECB) and 7% (CTR) faster than state-of-the-art.

키워드

Graphics processing units; Registers; Encryption; Instruction sets; Throughput; Computer architecture; Standards; Advanced encryption standard (AES); bit-sliced implementation; graphics processing unit (GPU); ENCRYPTION
제목
Speed Record of AES-CTR and AES-ECB Bit-Sliced Implementation on GPUs
저자
Lee, Wai-Kong; Seo, Seog Chung; Seo, Hwajeong; Kim, Dong Cheon; Hwang, Seong Oun
DOI
10.1109/LES.2024.3409725
발행일
2024-12
유형
Article
저널명
IEEE Embedded Systems Letters
권
16
호
4
페이지
481 ~ 484