6 February 2020 in IEEE Transactions on Components, Packaging and Manufacturing Technology
1 October 2021 in IEEE Transactions on Components, Packaging and Manufacturing Technology
In this article, we propose a processing-in-memory of high bandwidth memory (PIM-HBM) scheme including system architecture and hardware structure. The proposed scheme embeds processing units into the logic layer of the high bandwidth memory (HBM) to expose an excess dynamic random access memory (DRAM) bandwidth. With parallelized DRAM architecture and a high-speed through-silicon via (TSV) structure, the proposed scheme successfully extends the DRAM bandwidth of processing-in-memory (PIM). Also, the total energy consumption is decreased by the reduced interconnection and capacitance-reduced channel structure. We designed the overall architecture and structure with physical feasibility for application to the current HBM. The logic layer and DRAM layers in the HBM are configured to embed the processing units and parallelize the DRAM channels. For high-speed data transfer with low interconnect energy, the TSV and silicon interposer channels are designed and analyzed in consideration of signal integrity (SI). Based on the physical design, we obtained the interconnect length in detail. The interconnect energy and delay of the silicon interposer and on-chip interconnect were modeled through a SPICE simulation. We analyzed the accurate effects of interconnect reduction caused by PIM. For overall system performance and efficiency analysis, a cycle-level architectural simulation was conducted. We successfully evaluated and analyzed the system performance for memory-intensive applications. As a result, the proposed PIM-HBM achieves 53% and 10.4% improvement on average in computing performance and energy efficiency compared to the conventional graphic processing unit of high bandwidth memory (GPU-HBM).
9 December 2020 in IEEE Transactions on Electromagnetic Compatibility
In this article, we propose a near field analytical model of an electromagnetic interference reduction method using an nth harmonic frequency shielding coil with efficiency enhancement in a loosely coupled automotive wireless power transfer system. In general, in the loosely coupled automotive wireless power transfer system, the power transfer efficiency is relatively lower than a tightly coupled wireless power transfer system due to the low coupling. In addition, the electromagnetic interference at the nth harmonic frequency from the power source can interfere with nearby electronic devices. To reduce electromagnetic interference level and enhance efficiency, we proposed a shielding coil between a transmitter and a receiver coil. By using a near field analytical model of the magnetic field intensity and efficiency equations, we analyze based on the equations and obtain a proper value of the resonant frequency of the shielding coil. Therefore, we can achieve low levels of electromagnetic interference at the specific nth harmonic frequency and higher value of the power transfer efficiency in the loosely coupled automotive wireless power transfer system.
8 July 2021 in IEEE Transactions on Electromagnetic Compatibility
In this article, for the first time, we propose channel characteristic-based deep neural network (DNN) models for accurate eye-height (EH) and eye-width (EW) estimation of high bandwidth memory (HBM) silicon interposer channels. The proposed models preprocess the design parameters that are highly relevant to the characteristics of the HBM channels. By taking account of the contribution of each design parameter to the eye diagram, the proposed models can accurately estimate the EH and EW of the channels even with a limited number of datasets. For verification, the proposed DNN models were applied to the microstrip and stripline channels of an HBM silicon interposer. Only redistributed layer (RDL) was used to clearly see the effect of reflecting the channel characteristics of the proposed method. We compared the proposed DNN models with various regression methods and a conventional fully connected multilayer DNN model. As a result, the proposed DNN models reduced the EH and EW error rates by 22.7 and 43.9% compared to the other regression methods. In addition, the proposed DNN models not only reduced the error rates by 22.0–28.4% but also reduced the computing cost by 8.0–9.4%, compared to the conventional DNN model. Moreover, we compared the proposed models with various DNN models having other preprocessing structures. By showing 26.7 and 28.8% lower EH and EW error rates than the other DNN models, we validated that the proposed models properly consider the most dominant design factors in preprocessing.
17 January 2024 in IEEE Transactions on Electromagnetic Compatibility
In this article, a policy-based reinforcement learning (RL) method for optimizing through silicon via (TSV) array design in high-bandwidth memory (HBM) considering signal integrity is proposed. The proposed method can provide an optimal TSV-array signal/ground pattern design to maximize the eye opening (EO), which determines the bandwidth of the high-speed TSV channel. The proposed method adopts the proximal policy optimization algorithm, which directly trains the optimal policy, providing efficient handling of large action spaces rather than value-based RL. The convolutional neural network is used as a feature extractor to extract the location information of the TSV-array. To overcome the computational cost of the reward estimation, a fast EO estimation method is developed based on the equivalent circuit modeling and peak distortion analysis. The proposed method is applied to optimize 1-byte of TSV-array in a 16-high HBM and showed an 18.2% increase in EO compared with the initial design. The optimality performance of the proposed method is compared with deep q-network and random search algorithm, and the proposed method shows 3.4% and 9.6% better optimality, respectively.
19 November 2021 in IEEE Transactions on Components, Packaging and Manufacturing Technology
In this article, we, for the first time, propose a reinforcement learning (RL) model to design an optimal 3-D cross-point (X-Point) array structure considering signal integrity issues. The interconnection design problem is modeled to the Markov decision process (MDP). The proposed RL model designs the 3-D X-Point array structure based on three reward factors: the number of bits, the crosstalk, and the IR drop. We applied multilayer perceptron (MLP) and long short-term memory (LSTM) to parameterize the policy. Proximal policy optimization (PPO) is used to optimize the parameters to train the policy. The reward of the proposed RL model is well-converged with variations in the array structure size and hyperparameters of the reward factors. We verified the scalability and sensitivity of the proposed RL model. With the optimal 3-D X-Point array structure design, we analyzed the reward factor and signal integrity issues. The optimal design of the 3-D X-Point array structure shows 17%–26.5% better signal integrity performance than the conventional design in finer process technology. In addition, we suggest a range of possible directions for improvement of the proposed model with variations in MDP tuples, reward factors, and learning algorithms, among other factors. Using the proposed model, we can easily design an optimal 3-D X-Point array structure with a certain size, performance capabilities, and specifications based on reward factors and hyperparameters.