Developing of the method for optimizing the performance of architecture-independent hardware platforms

The object of research is the Raspberry Pi single-board computer. The work examines the optimization of architecture-independent hardware platforms using its example. The research is based on an integrated scientific approach based on a system-analytical, structural-functional, empirical and typological approach. It is emphasized that the entire Raspberry Pi line uses APM-architecture processors. The genesis of Raspberry Pi is given, the parameters of the last build are determined. It is noted that the latest version is dated November 2020. It is equipped with wireless WiFi and Bluetooth modules (2×USB 3.0 and 1×USB 2.0 ports type A, 5.0, BLE), which expand the boundaries of mini-PC application in the field of Ethernet technologies and has a frequency of 1.8 GHz. The appearance of one of the popular Raspberry Pi B+boards has been formed, with the separation of the main blocks. The basic principles of improving the performance of the Raspberry Pi single board computer are determined, each of which is based on a specific mechanism. The first is the addition of ZRAM as a compressed random access memory block device. The principle of ZRAM operation is described, the mechanism for activating ZRAM on the Raspberry Pi is given. To improve the performance of the Raspberry Pi single board computer, the use of an NVMe disk is justified. It is emphasized that the NVMe disk is reliable and has a high data transfer rate. Connecting it to the Raspberry Pi single board computer is the optimal solution to improve performance. The tuning sequence is presented, the numerical result of the NVMe disk operation based on the Raspberry Pi single-board computer is proposed. It is proposed, as a principle to improve performance, the installation of an ICE Tower CPU based on Raspberry Pi. It is noted that the ICE Tower CPU is a cooling system that is designed to cool the Raspberry Pi. The principles of tuning ICE Tower CPU and the result of fluctuations in temperature components using the rpi-monitor are described. As part of the study, performance improvements were obtained from 26 % to 34 %, which is mainly in line with the expected theoretical improvement of 34 %