In recent years, the increase in attacks on IoT devices has become a problem, and security measures for IoT devices are urgently needed. We have proposed a malware detection mechanism using processor information and confirmed that mechanism can detect malware and its variants. In some past research, a malware detection mechanism using processor information was proposed and confirmed that the mechanism can detect malware and its variants. By offloading the proposed mechanism to hardware, we aim to realize a lightweight malware detection mechanism for IoT devices in the future. In this paper, we consider reduction of the classifier size and acceleration of the classification algorithm, which are issues in the our proposed mechanism.