The Cryptography and Security Engineering Laboratory at Kookmin University (Advisor: Professor Seo Seog Chung), PhD candidates Dong-Cheon Kim and Jun-Hyuk Choi, Dr. Wai Kong Lee of Universiti Tunku Abdul Rahman (UTAR) in Malaysia, and and Professor Seo Seog Chung have had their joint research paper, “Optimizing HQC with Enhanced Parallel Binary Polynomial Multiplication on GPUs,” accepted for publication in Issue 1, 2027, of *IACR Transactions on Cryptographic Hardware and Embedded Systems* (TCHES), the most prestigious academic journal in the field of cryptographic implementation. This research was conducted at the Glocal Lab Institute for Defense Technology Protection (Director: Professor Han Dong-guk), which is supported by the Ministry of Education.
 
This research was undertaken to achieve high-speed implementation on GPUs of HQC, a code-based quantum-resistant cryptosystem selected by the U.S. National Institute of Standards and Technology (NIST) as an additional standard in 2025. While server environments require rapid processing of large volumes of key exchange requests, HQC faced limitations in high-speed implementation because binary polynomial multiplication—which accounts for the majority of computations—acted as a major bottleneck.
 
To address this, the research team proposed two methods for binary polynomial multiplication in a GPU environment for the first time. These include a method that parallelizes the Frobenius Additive FFT on the GPU and a method that performs the multiplication using the GPU’s INT1 Tensor Core. Furthermore, the team parallelized Reed–Muller/Reed–Solomon encoding and decoding—which emerged as a new bottleneck after the multiplication optimization—as well as fixed-weight sampling at the warp level, enabling the entire KEM computation to be performed within the GPU. Furthermore, by presenting five implementation methods that set different performance trade-offs—taking into account not only throughput but also the latency of individual requests—the researchers enabled users to select an implementation tailored to their specific environment. As a result, on an RTX 4090 GPU, polynomial multiplication throughput improved by approximately 2.9 times and overall KEM throughput by up to 2.03 times compared to the latest GPU implementation, while single-request latency was reduced to a minimum of 254 µs.
 
Meanwhile, the Cryptography and Security Engineering Laboratory at Kookmin University has achieved the publication or acceptance of a total of five papers: two in TCHES 2026, Issue 2; one in Issue 4; and two in 2027, Issue 1. In particular, the lab achieved world-class research results across a variety of computing environments, ranging from the acceleration of GPU-based cryptographic operations to the optimization of quantum-resistant cryptography in ultra-lightweight embedded environments and the design of FPGA-based integrated hardware accelerators. These achievements demonstrate the lab’s broad cryptographic implementation capabilities spanning both software and hardware, and are expected to contribute to the nation’s transition to quantum-resistant cryptography and the strengthening of South Korea’s international competitiveness in cryptographic technology.

(From left: Professor Seo Seog Chung, Professor Han Dong-guk, Kim Dong-cheon, Ph.D. candidate Choi Jun-hyuk, and Dr. Wai Kong Lee)