Jie-Seok Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jie-Seok Kim has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Jong-Wook Lee
Department of Electronics and Information Convergence Engineering, Information and Communication System-on-Chip (SoC) Research Center, Kyung Hee University, Yongin, South Korea
2 shared publications Van Truong Nguyen
Department of Electronics and Information Convergence Engineering, Information and Communication System-on-Chip (SoC) Research Center, Kyung Hee University, Yongin, South Korea
1 shared publication- Van-Nhan Nguyen
Department of Electronics and Information Convergence Engineering, Information and Communication System-on-Chip (SoC) Research Center, Kyung Hee University, Yongin, Republic of Korea
1 shared publication - Xuan Thanh Pham
School of Electrical and Electronic Engineering, Hanoi University of Industry, Hanoi, Vietnam
1 shared publication
Latest Publications
- Open Access
- 12 May 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 71262-71276
10T SRAM Computing-in-Memory Macros for Binary and Multibit MAC Operation of DNN Edge Processors
Computing-in-memory (CIM) is a promising approach to reduce latency and improve the energy efficiency of the multiply-and-accumulate (MAC) operation under a memory wall constraint for artificial intelligence (AI) edge processors. This paper proposes an approach focusing on scalable CIM designs using a new ten-transistor (10T) static random access memory (SRAM) bit-cell. Using the proposed 10T SRAM bit-cell, we present two SRAM-based CIM (SRAM-CIM) macros supporting multibit and binary MAC operations. The first design achieves fully parallel computing and high throughput using 32 parallel binary MAC operations. Advanced circuit techniques such as an input-dependent dynamic reference generator and an input-boosted sense amplifier are presented. Fabricated in 28 nm CMOS process, this design achieves 409.6 GOPS throughput, 1001.7 TOPS/W energy efficiency, and a 169.9 TOPS/mm 2 throughput area efficiency. The proposed approach effectively solves previous problems such as writing disturb, throughput, and the power consumption of an analog to digital converter (ADC). The second design supports multibit MAC operation (4-b weight, 4-b input, and 8-b output) to increase the inference accuracy. We propose an architecture that divides 4-b weight and 4-b input multiplication to four 2-b multiplication in parallel, which increases the signal margin by 16× compared to conventional 4-b multiplication. Besides, the capacitive digital-to-analog converter (CDAC) area issue is effectively addressed using the intrinsic bit-line capacitance existing in the SRAM-CIM architecture. The proposed approach of realizing four 2-b parallel multiplication using the CDAC is successfully demonstrated with a modified LeNet-5 neural network. These results demonstrate that the proposed 10T bit-cell is promising for realizing robust and scalable SRAM-CIM designs, which is essential for realizing fully parallel edge computing.
- 12 May 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 71262-71276
Open Access10T SRAM Computing-in-Memory Macros for Binary and Multibit MAC Operation of DNN Edge Processors
Computing-in-memory (CIM) is a promising approach to reduce latency and improve the energy efficiency of the multiply-and-accumulate (MAC) operation under a memory wall constraint for artificial intelligence (AI) edge processors. This paper proposes an approach focusing on scalable CIM designs using a new ten-transistor (10T) static random access memory (SRAM) bit-cell. Using the proposed 10T SRAM bit-cell, we present two SRAM-based CIM (SRAM-CIM) macros supporting multibit and binary MAC operations. The first design achieves fully parallel computing and high throughput using 32 parallel binary MAC operations. Advanced circuit techniques such as an input-dependent dynamic reference generator and an input-boosted sense amplifier are presented. Fabricated in 28 nm CMOS process, this design achieves 409.6 GOPS throughput, 1001.7 TOPS/W energy efficiency, and a 169.9 TOPS/mm 2 throughput area efficiency. The proposed approach effectively solves previous problems such as writing disturb, throughput, and the power consumption of an analog to digital converter (ADC). The second design supports multibit MAC operation (4-b weight, 4-b input, and 8-b output) to increase the inference accuracy. We propose an architecture that divides 4-b weight and 4-b input multiplication to four 2-b multiplication in parallel, which increases the signal margin by 16× compared to conventional 4-b multiplication. Besides, the capacitive digital-to-analog converter (CDAC) area issue is effectively addressed using the intrinsic bit-line capacitance existing in the SRAM-CIM architecture. The proposed approach of realizing four 2-b parallel multiplication using the CDAC is successfully demonstrated with a modified LeNet-5 neural network. These results demonstrate that the proposed 10T bit-cell is promising for realizing robust and scalable SRAM-CIM designs, which is essential for realizing fully parallel edge computing.
- 17 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 68 (6) , 1793-1797
A 0.52 μW, 38 nV/√Hz Chopper Amplifier With a Low-Noise DC Servo Loop, an Embedded Ripple Reduction Loop, and a Squeezed Inverter Stage
To realize an ultralow-power, low-noise amplifier for bio-potential recording, we investigate three design techniques. The first technique uses a noise-efficient squeezed inverter (SQI) stage biased at the supply of the 2VDSAT saturation limit. The challenge of interfacing the SQI stage with such a low supply is addressed by proposing a new capacitively-coupled chopper instrumentation amplifier (CCIA) with a low-noise DC servo loop (L-DSL) and an embedded ripple reduction loop (E-RRL). The second technique is reducing the high noise contribution of the conventional DSL. The proposed L-DSL reduces the inputreferred noise (IRN) by removing the charge dividing effect. The proposed E-RRL inserted inside the gain stage cancels the offset and achieves a ripple suppression without loading the output. The CCIA is implemented using a 0.18 μm CMOS process. The measured results show that the L-DSL successfully creates a sub-Hz high-pass corner needed to block the electrode offset. The ERRL achieves a ripple suppression of 39 dB. The CCIA achieves 32 nV/Hz noise density, which is slightly increased to 38 nV/Hz when the L-DSL is enabled. The integrated noise over 800 Hz bandwidth is 0.9 and 1.1 μVrms without and with the L-DSL, respectively. The mid-band gain of 39.6 dB is achieved by consuming 0.52 μW. These results correspond to a favorable noise efficiency factor (NEF) of 2.1 and a power efficiency factor (PEF) of 1.2.
- 17 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems II: Express Briefs
- Vol. 68 (6) , 1793-1797
A 0.52 μW, 38 nV/√Hz Chopper Amplifier With a Low-Noise DC Servo Loop, an Embedded Ripple Reduction Loop, and a Squeezed Inverter Stage
To realize an ultralow-power, low-noise amplifier for bio-potential recording, we investigate three design techniques. The first technique uses a noise-efficient squeezed inverter (SQI) stage biased at the supply of the 2VDSAT saturation limit. The challenge of interfacing the SQI stage with such a low supply is addressed by proposing a new capacitively-coupled chopper instrumentation amplifier (CCIA) with a low-noise DC servo loop (L-DSL) and an embedded ripple reduction loop (E-RRL). The second technique is reducing the high noise contribution of the conventional DSL. The proposed L-DSL reduces the inputreferred noise (IRN) by removing the charge dividing effect. The proposed E-RRL inserted inside the gain stage cancels the offset and achieves a ripple suppression without loading the output. The CCIA is implemented using a 0.18 μm CMOS process. The measured results show that the L-DSL successfully creates a sub-Hz high-pass corner needed to block the electrode offset. The ERRL achieves a ripple suppression of 39 dB. The CCIA achieves 32 nV/Hz noise density, which is slightly increased to 38 nV/Hz when the L-DSL is enabled. The integrated noise over 800 Hz bandwidth is 0.9 and 1.1 μVrms without and with the L-DSL, respectively. The mid-band gain of 39.6 dB is achieved by consuming 0.52 μW. These results correspond to a favorable noise efficiency factor (NEF) of 2.1 and a power efficiency factor (PEF) of 1.2.