Hash functions are essential in cryptographic primitives such as digital signatures, key exchanges, and blockchain technology. SM3, built upon the Merkle-Damgard structure, is a crucial element in Chinese commercial cryptographic schemes. Optimizing hash function performance is crucial given the growth of Internet of Things (IoT) devices and the rapid evolution of blockchain technology. In this paper, we introduce a high-performance implementation framework for accelerating the SM3 cryptography hash function, short for HI-SM3, using heterogeneous GPU (graphics processing unit) parallel computing devices. HI-SM3 enhances the implementation of hash functions across four dimensions: parallelism, register utilization, memory access, and instruction efficiency, resulting in significant performance gains across various GPU platforms. Leveraging the NVIDIA RTX 4090 GPU, HI-SM3 achieves a remarkable peak performance of 454.74 GB/s, surpassing OpenSSL on a high-end server CPU (E5-2699V3) with 16 cores by over 150 times. On the Hygon DCU accelerator, a Chinese domestic graphics card, it achieves 113.77 GB/s. Furthermore, compared with the fastest known GPU-based SM3 implementation, HI-SM3 on the same GPU platform exhibits a 3.12x performance improvement. Even on embedded GPUs consuming less than 40W, HI-SM3 attains a throughput of 5.90 GB/s, which is twice as high as that of a server-level CPU. In summary, HI-SM3 provides a significant performance advantage, positioning it as a compelling solution for accelerating hash operations.

With the rapid development of operating systems, attacks on system vulnerabilities are increasing. Dynamic link library (DLL) hijacking is prevalent in installers on freeware platforms and is highly susceptible to exploitation by malware attackers. However, existing studies are based solely on the load paths of DLLs, ignoring the attributes of installers and invocation modes, resulting in low accuracy and weak generality of vulnerability detection. In this paper, we propose a novel model, AB-DHD, which is based on an attention mechanism and a bi-directional gated recurrent unit (BiGRU) neural network for DLL hijacking vulnerability discovery. While BiGRU is an enhancement of GRU and has been widely applied in sequence data processing, a double-layer BiGRU network is introduced to analyze the internal features of installers with DLL hijacking vulnerabilities. Additionally, an attention mechanism is incorporated to dynamically adjust feature weights, significantly enhancing the ability of our model to detect vulnerabilities in new installers. A comprehensive “List of Easily Hijacked DLLs” is developed to serve a reference for future studies. We construct an EXEFul dataset and a DLLVul dataset, using data from two publicly available authoritative vulnerability databases, Common Vulnerabilities & Exposures (CVE) and China National Vulnerability Database (CNVD), and mainstream installer distribution platforms. Experimental results show that our model outperforms popular automated tools like Rattler and DLLHSC, achieving an accuracy of 97.79% and a recall of 94.72%. Moreover, 17 previously unknown vulnerabilities have been identified, and corresponding vulnerability certifications have been assigned.