Computing systems are widely used in medical diagnosis, climate prediction, autonomous vehicles, etc. As the key part of electronics, the performance of computing systems is crucial in the intellectualization of the equipment. The conflict between performance, efficiency, and cost can be solved by choosing an appropriate computing system architecture. In order to provide useful advice and instructions for the designers to fabricate high-performance computing systems, this paper reviews the Chiplet-based computing system architectures, including computing architecture and memory architecture. Firstly, the computing architecture used for high-performance computing, mobile, and PC is presented and summarized. Secondly, the memory architecture based on mainstream memory and emerging non-volatile memory used for data storing and processing are introduced, and the key parameters of memory are compared and discussed. Finally, this paper is concluded, and the future perspectives of computing system architecture based on Chiplet are presented.

Microsystems are widely used in 5G, the Internet of Things, smart electronic devices and other fields, and signal integrity (SI) determines their performance. Establishing accurate and fast predictive models and intelligent optimization models for SI in microsystems is extremely essential. Recently, neural networks (NNs) and heuristic optimization algorithms have been widely used to predict the SI performance of microsystems. This paper systematically summarizes the neural network methods applied in the prediction of microsystem SI performance, including artificial neural network (ANN), deep neural network (DNN), recurrent neural network (RNN), convolutional neural network (CNN), etc., as well as intelligent algorithms applied in the optimization of microsystem SI, including genetic algorithm (GA), differential evolution (DE), deep partition tree Bayesian optimization (DPTBO), two stage Bayesian optimization (TSBO), etc., and compares and discusses the characteristics and application fields of the current applied methods. The future development prospects are also predicted. Finally, the article is summarized.

An accurate equivalent thermal model is proposed to calculate the equivalent thermal conductivity (ETC) of shield differential through-silicon via (SDTSV). The mathematical expressions of ETC in both horizontal and vertical directions are deduced by considering the anisotropy of SDTSV. The accuracy of the proposed model is verified by the finite element method (FEM), and the average errors of temperature along the X-axis, Y-axis, diagonal line, and vertical directions are 1.37%, 3.42%, 1.76%, and 0.40%, respectively. Compared with COMSOL, the proposed model greatly improves the computational efficiency. Moreover, the effects of different parameters on the thermal distribution of SDTSV are also investigated. The thermal conductivity is decreased with the increase in thickness of SiO₂. With the increase in pitch, the maximum temperature of SDTSV increases very slowly when β = 0$°$ , and decreases very slowly when β = 90$°$. The proposed model can be used to accurately and quickly describe the thermal distribution of SDTSV, which has a great prospect in the design of 3D IC.

An efficient multi-objective optimization method of temperature and stress for a microsystem based on particle swarm optimization (PSO) was established, which is used to map the relationship between through-silicon via (TSV) structural design parameters and performance objectives in the microsystem, and complete optimization temperature, stress and thermal expansion deformation efficiently. The relationship between the design and performance parameters is obtained by a finite element method (FEM) simulation model. The neural network is built and trained in order to understand the mapping relationship. Then, the design parameters are iteratively optimized using the PSO algorithm, and the FEM results are used to verify the efficiency and reliability of the optimization methods. When the optimization target of peak temperature, bump temperature, TSV temperature, maximum stress and maximum thermal deformation are set as 100 °C, 55 °C, 35 °C, 180 Mpa and 12 μm, the optimization results are as follows: the peak temperature is 97.90 °C, the bump temperature is 56.01 °C, the TSV temperature is 31.52 °C, the maximum stress is 247.4 Mpa and the maximum expansion deformation is 11.14 μm. The corresponding TSV structure design parameters are as follows: the radius of TSV is 10.28 μm, the pitch is 65 μm and the thickness of SiO₂ is 0.83 μm. The error between the optimization result and the target temperature is 2.1%, 1.8%, 9.9%, 37.4% and 7.2% respectively. The PSO method has been verified by regression analysis, and the difference between the temperature and deformation optimization results of the FEM method is not more than 3%. The stress error has been analyzed, and the reliability of the developed method has been verified. While ensuring the accuracy of the results, the proposed optimization method reduces the time consumption of a single simulation from 2 h to 70 s, saves a lot of time and human resources, greatly improves the efficiency of the optimization design of microsystems, and has great significance for the development of microsystems.