Zikun Wei
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zikun Wei has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Chenchen Ding
Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, SAR, China
2 shared publications - Hao Yu
School of Microelectronics, Southern University of Science and Technology, Shenzhen, CHINA
2 shared publications - Dingbang Liu
Southern Univ. of Science and Tech., Shenzhen, China
1 shared publication - Haoxiang Zhou
Southern Univ. of Science and Tech., Shenzhen, China
1 shared publication - Jiaqi Yang
Peking University, Beijing, Beijing, CHINA
1 shared publication - Junyi Luo
School of Microelectronics, Southern University of Science and Technology, Shenzhen, CHINA
1 shared publication
Latest Publications
- Open Access
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 720-726
A Layer-wised Mixed-Precision CIM Accelerator with Bit-level Sparsity-aware ADCs for NAS-Optimized CNNs
Exploring multiple precisions as well as sparsities for a computingin-memory (CIM) based convolutional accelerators is challenging. To further improve energy efficiency with minimal accuracy loss, this paper develops a neural architecture search (NAS) method to identify precision for each layer of the CNN and further leverages bit-level sparsity. The results indicate that following this approach, ResNet-18 and VGG-16 not only maintain their accuracy but also implement layer-wised mixed-precision effectively. Furthermore, there is a substantial enhancement in the bit-level sparsity of weights within each layer, with an average bit-level sparsity exceeding 90% per bit, thus providing broader possibilities for hardware-level sparsity optimization. In terms of hardware design, a mixed-precision (2/4/8-bit) readout circuit as well as a bit-level sparsity-aware Analog-to-Digital Converter (ADC) are both proposed to reduce system power consumption. Based on bit-level sparsity mixed-precision CNNs benchmarks, post-layout simulation results in 28nm reveal that the proposed accelerator achieves up to 245.72 TOPS/W energy efficiency, which shows about 2.52 -- 6.57× improvement compared to the state-of-the-art SRAM-based CIM accelerators.
- 20 January 2025
- Published by Association for Computing Machinery (ACM)
- p. 720-726
Open AccessA Layer-wised Mixed-Precision CIM Accelerator with Bit-level Sparsity-aware ADCs for NAS-Optimized CNNs
Exploring multiple precisions as well as sparsities for a computingin-memory (CIM) based convolutional accelerators is challenging. To further improve energy efficiency with minimal accuracy loss, this paper develops a neural architecture search (NAS) method to identify precision for each layer of the CNN and further leverages bit-level sparsity. The results indicate that following this approach, ResNet-18 and VGG-16 not only maintain their accuracy but also implement layer-wised mixed-precision effectively. Furthermore, there is a substantial enhancement in the bit-level sparsity of weights within each layer, with an average bit-level sparsity exceeding 90% per bit, thus providing broader possibilities for hardware-level sparsity optimization. In terms of hardware design, a mixed-precision (2/4/8-bit) readout circuit as well as a bit-level sparsity-aware Analog-to-Digital Converter (ADC) are both proposed to reduce system power consumption. Based on bit-level sparsity mixed-precision CNNs benchmarks, post-layout simulation results in 28nm reveal that the proposed accelerator achieves up to 245.72 TOPS/W energy efficiency, which shows about 2.52 -- 6.57× improvement compared to the state-of-the-art SRAM-based CIM accelerators.
- 19 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 70 (12) , 5289-5301
An Integer-Only and Group-Vector Systolic Accelerator for Efficiently Mapping Vision Transformer on Edge
Transformer-like network has shown remarkable high performance in both natural language processing and computer vision. However, the huge computational demands in non-linear floating-point arithmetic and the irregular memory access requirement in self-attention mechanism make it still a challenge to deploy Transformer on edge. To address the above issues, we propose integer-only quantization scheme for the simplification of non-linear operations (such as LayerNorm, Softmax and Gelu), meanwhile algorithm-hardware co-design strategy is applied to guarantee both the high accuracy and high efficiency. Besides, we construct general-purpose group vector systolic array to efficiently accelerate the matrix multiplication operations including both regular matrix-multiplication/convolution and the irregular multi-head self-attention mechanism. Unified data-package strategy and flexible on-/off-chip data storage management strategy are also proposed to further improve the performance. The design has been deployed on Xilinx ZCU102 FPGA platform, achieving an overall inference latency of 4.077ms and 11.15ms per image for ViT-tiny and ViT-s, respectively. The average throughput can reach as high as 762.7 GOPs, which shows significant improvement over the previous state-of-the-art FPGA Transformer accelerator.
- 19 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 70 (12) , 5289-5301
An Integer-Only and Group-Vector Systolic Accelerator for Efficiently Mapping Vision Transformer on Edge
Transformer-like network has shown remarkable high performance in both natural language processing and computer vision. However, the huge computational demands in non-linear floating-point arithmetic and the irregular memory access requirement in self-attention mechanism make it still a challenge to deploy Transformer on edge. To address the above issues, we propose integer-only quantization scheme for the simplification of non-linear operations (such as LayerNorm, Softmax and Gelu), meanwhile algorithm-hardware co-design strategy is applied to guarantee both the high accuracy and high efficiency. Besides, we construct general-purpose group vector systolic array to efficiently accelerate the matrix multiplication operations including both regular matrix-multiplication/convolution and the irregular multi-head self-attention mechanism. Unified data-package strategy and flexible on-/off-chip data storage management strategy are also proposed to further improve the performance. The design has been deployed on Xilinx ZCU102 FPGA platform, achieving an overall inference latency of 4.077ms and 11.15ms per image for ViT-tiny and ViT-s, respectively. The average throughput can reach as high as 762.7 GOPs, which shows significant improvement over the previous state-of-the-art FPGA Transformer accelerator.