Considerations Of Integrating Computing-In-Memory And Processing-In-Sensor Into Convolutional Neural Network Accelerators For Low-Power Edge Devices

Tang, Kea-Tiong; Wei, Weichen; Yeh, Zuo-Wei; Hsu, Tzu-Hsiang; Chiu, Yen-Cheng; Xue, Cheng-Xin; Kuo, Yu-Chun; Wen, Tai-Hsing; Ho, Mon−Shu; Lo, Chung-Chuan; Liu, Ren-Shuo; Hsieh, Chih-Cheng; Chang, Meng‐Fan

doi:10.23919/vlsic.2019.8778074

Cited by 6 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Quantization parameters are optimized by analyzing the statistical inference information and comparing the impacts of different bit widths. The quantization process is denoted in (2), where 𝑎 represents the weight or input, and 𝑎 represents the quantized values. The parameter 𝑆 denotes the fractional length and a larger 𝑆 indicates higher quantization resolution.…”

Section: A Quantization Analyzer (1) Quantizermentioning

confidence: 99%

See 1 more Smart Citation

A Compilation Framework for SRAM Computing-in-Memory Systems with Optimized Weight Mapping and Error Correction

Bai,

Li,

Zhang

et al. 2024

Preprint

View full text Add to dashboard Cite

Deploying convolution-based algorithms into SRAM computing-in-memory (CIM) systems faces various challenges, such as operator incompatibility and intrinsic non-ideal error. This paper proposes a compilation framework to address this issue. Efficient weight mapping strategies are introduced to improve the utilization of SRAM-CIM macro. The intrinsic non-ideal errors of SRAM-CIM macro are also taken into consideration, and two efficient error correction schemes are proposed, which include calibration of computation voltage linear error (CCVLE) and the mitigation of analog-to-digital quantization error (MAQE). In addition, bit-width flexibility and signed-unsigned reconfigurability are also supported to facilitate the deployment of various convolution-based algorithms. ResNet18, finite impulse response (FIR) filtering, and Gaussian image filtering are deployed into a multi-macro SRAM-CIM system. These algorithms serve as deployment representatives of convolutional neural network (CNN), digital signal processing (DSP), and digital image processing (DIP), respectively. The results show that the introduced weight mapping strategies improve the macro utilization by 63.29% and 21.10% for two types of frequently used convolution layers compared to the traditional strategy. Moreover, the proposed error correction schemes achieve similar algorithm accuracy to the floating-point results, and the deployment result of ResNet18 achieves 66.3%~70.1% top-1 classification accuracy evaluated on the ImageNet dataset with different throughput tradeoffs.

show abstract

Section: A Quantization Analyzer (1) Quantizermentioning

confidence: 99%

“…xisting digital processors face the data-transmission bottleneck caused by the intrinsic Von-Neumann structure for computation-intensive tasks [1,2]. Recently, computing-in-memory (CIM) systems have been proposed to overcome this challenge by integrating computing logic into the memory macro [3].…”

Section: Introductionmentioning

confidence: 99%