Bishnu Prasad Das
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Bishnu Prasad Das has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Aranya Gupta
Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee, Roorkee, India
3 shared publications - Priyamvada Sharma
Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee, Roorkee, India
3 shared publications - Sanjeev Kumar Manhas
Department of Electronics and Communication Engineering, IIT Roorkee, Roorkee, India
3 shared publications - Anshul Verma
Department of Electronics and Communication Engineering, IIT Roorkee, Roorkee, Uttarakhand, India
2 shared publications - Anu Verma
Department of Electronics and Communication Engineering, Indian Institute of Technology (IIT) Roorkee, Roorkee, Uttarakhand, India
2 shared publications - Poorvi Jain
Indian Inst Technol Roorkee, Dept Elect & Commun Engn, Roorkee 247667, Uttar Pradesh, India
2 shared publications
Latest Publications
- 18 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Microwave Theory and Techniques
- Vol. 73 (11) , 8917-8928
A Low-Jitter and Spur Dual-Edge Sub-Sampling Phase-Locked Loop With Double-Balanced Sampling Mixer-Adopted Phase Detector
Advancements in high-speed wireline and wireless applications are driving the demand for high-performance frequency synthesizers. The conventional sub-sampling phase detector (SSPD) suffers from poor reference spur performance, limiting the overall effectiveness of the sub-sampling phase-locked loop (SSPLL). This article presents a SSPLL incorporating a double-balanced sampling mixer-adopted SSPD (DBSMA-SSPD). The proposed design improves the reference spur by reducing reference feedthrough without requiring external calibration or RC networks. A dual-edge sampling technique is introduced to further enhance in-band phase noise performance. The proposed SSPLL employs an LC-voltage controlled oscillator (VCO) as its core to generate a 2.688 GHz frequency. The design is implemented and fabricated using an industrial 65-nm CMOS process, occupying a total active area of 0.39 mm2. The SSPLL successfully locks to a 2.688 GHz frequency from a reference frequency of 168 MHz, consuming around 4.1 mW from a 1.2-V supply. The design demonstrates state-of-the-art performance, achieving a measured rms jitter of 207.7 fs over an integration bandwidth of 10 kHz–100 MHz and a measured reference spur of −72 dBc, attributed to the proposed DBSMA-SSPD. Additionally, it achieves an in-band phase noise of −122.16 dBc/Hz at a 200-kHz frequency offset, with a jitter-power figure-of-merit (FOM) of −247.5 dB.
- 18 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Microwave Theory and Techniques
- Vol. 73 (11) , 8917-8928
A Low-Jitter and Spur Dual-Edge Sub-Sampling Phase-Locked Loop With Double-Balanced Sampling Mixer-Adopted Phase Detector
Advancements in high-speed wireline and wireless applications are driving the demand for high-performance frequency synthesizers. The conventional sub-sampling phase detector (SSPD) suffers from poor reference spur performance, limiting the overall effectiveness of the sub-sampling phase-locked loop (SSPLL). This article presents a SSPLL incorporating a double-balanced sampling mixer-adopted SSPD (DBSMA-SSPD). The proposed design improves the reference spur by reducing reference feedthrough without requiring external calibration or RC networks. A dual-edge sampling technique is introduced to further enhance in-band phase noise performance. The proposed SSPLL employs an LC-voltage controlled oscillator (VCO) as its core to generate a 2.688 GHz frequency. The design is implemented and fabricated using an industrial 65-nm CMOS process, occupying a total active area of 0.39 mm2. The SSPLL successfully locks to a 2.688 GHz frequency from a reference frequency of 168 MHz, consuming around 4.1 mW from a 1.2-V supply. The design demonstrates state-of-the-art performance, achieving a measured rms jitter of 207.7 fs over an integration bandwidth of 10 kHz–100 MHz and a measured reference spur of −72 dBc, attributed to the proposed DBSMA-SSPD. Additionally, it achieves an in-band phase noise of −122.16 dBc/Hz at a 200-kHz frequency offset, with a jitter-power figure-of-merit (FOM) of −247.5 dB.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Embedded Systems Letters
- Vol. 18 (2) , 140-143
Weak PUF-Based Variable Latency Obfuscation Technique for ML-Attack Resilient Arbiter PUFs
The conventional linear feedback shift register (LFSR)–based arbiter PUF (APUF) suffers from machine learning (ML) attacks, which is a challenge for employment in real-world authentication scenarios. To mitigate this vulnerability, our paper presents a weak PUF-assisted and challenge-dependent dynamic and non-linear feedback shift register (DNLFSR)-based APUF. The proposed technique introduces dynamic and variable iteration-based LFSR to make the generated challenge-response pair (CRP) space more non-linear, which effectively prevents attackers from modeling the APUF. The proposed DNLFSR-based APUF achieves ≈ 50% prediction accuracy (similar to random guessing) against various ML attacks carried out through exhaustive testing over 1 million CRPs. Additionally, the proposed technique reduces latency by ≈n/2 clock cycles for an n-stage LFSR compared to state-of-the-art techniques. Extensive performance evaluation of the proposed DNLFSR-based APUF on both Python simulation and FPGA hardware implementation shows near-ideal uniformity, uniqueness, and reliability for 64-bit and 128-bit DNLFSR-based APUF. Moreover, the proposed design consumes low hardware overhead, showing a minimum 43.2% reduction in LUT usage compared to the state-of-the-art techniques, making it a lightweight solution for IoT applications.
- 25 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Embedded Systems Letters
- Vol. 18 (2) , 140-143
Weak PUF-Based Variable Latency Obfuscation Technique for ML-Attack Resilient Arbiter PUFs
The conventional linear feedback shift register (LFSR)–based arbiter PUF (APUF) suffers from machine learning (ML) attacks, which is a challenge for employment in real-world authentication scenarios. To mitigate this vulnerability, our paper presents a weak PUF-assisted and challenge-dependent dynamic and non-linear feedback shift register (DNLFSR)-based APUF. The proposed technique introduces dynamic and variable iteration-based LFSR to make the generated challenge-response pair (CRP) space more non-linear, which effectively prevents attackers from modeling the APUF. The proposed DNLFSR-based APUF achieves ≈ 50% prediction accuracy (similar to random guessing) against various ML attacks carried out through exhaustive testing over 1 million CRPs. Additionally, the proposed technique reduces latency by ≈n/2 clock cycles for an n-stage LFSR compared to state-of-the-art techniques. Extensive performance evaluation of the proposed DNLFSR-based APUF on both Python simulation and FPGA hardware implementation shows near-ideal uniformity, uniqueness, and reliability for 64-bit and 128-bit DNLFSR-based APUF. Moreover, the proposed design consumes low hardware overhead, showing a minimum 43.2% reduction in LUT usage compared to the state-of-the-art techniques, making it a lightweight solution for IoT applications.
- Open Access
- 15 July 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
A Unified Approach to a Secure and Lightweight Mutual Authentication Protocol Using Pre-Characterized COTS SRAM ICs for IoT Applications
Traditional Physical Unclonable Function (PUF)-based authentication protocols are vulnerable to machine learning attacks and evolving cyber threats. Moreover, these protocols lack suitability for resource-constrained IoT devices due to the involvement of heavy cryptographic primitives, error correction modules, and significant computational overhead. This article proposes a mutual authentication protocol and session key agreement utilizing commercial-off-the-shelf (COTS) SRAM integrated circuits (ICs) to extract a secret key. We introduce a block-based lightweight fuzzy extractor to minimize the overhead associated with error correction modules on IoT devices. Our protocol relies only hash, XOR and masking functions for the identity verification for both parties and stores only one challenge-response pair (CRP) on the server, reducing memory overhead on the authentication server. In addition, we perform a rigorous informal security analysis against well-known attacks and formal security analysis using Verifpal tool considering an active attacker in the communication link. Furthermore, the performance evaluation and comparative analysis indicate that the proposed protocol significantly outperforms the state-of-the-art protocols in terms of communication, computational overhead, storage overhead, and energy consumption by up to 69%, 81%, 87.5% and 76.6% respectively. We have implemented the proposed authentication protocol on ESP32 and Raspberry Pi 3 boards to show its applicability and scalability in a real-world IoT framework.
- 15 July 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
Open AccessA Unified Approach to a Secure and Lightweight Mutual Authentication Protocol Using Pre-Characterized COTS SRAM ICs for IoT Applications
Traditional Physical Unclonable Function (PUF)-based authentication protocols are vulnerable to machine learning attacks and evolving cyber threats. Moreover, these protocols lack suitability for resource-constrained IoT devices due to the involvement of heavy cryptographic primitives, error correction modules, and significant computational overhead. This article proposes a mutual authentication protocol and session key agreement utilizing commercial-off-the-shelf (COTS) SRAM integrated circuits (ICs) to extract a secret key. We introduce a block-based lightweight fuzzy extractor to minimize the overhead associated with error correction modules on IoT devices. Our protocol relies only hash, XOR and masking functions for the identity verification for both parties and stores only one challenge-response pair (CRP) on the server, reducing memory overhead on the authentication server. In addition, we perform a rigorous informal security analysis against well-known attacks and formal security analysis using Verifpal tool considering an active attacker in the communication link. Furthermore, the performance evaluation and comparative analysis indicate that the proposed protocol significantly outperforms the state-of-the-art protocols in terms of communication, computational overhead, storage overhead, and energy consumption by up to 69%, 81%, 87.5% and 76.6% respectively. We have implemented the proposed authentication protocol on ESP32 and Raspberry Pi 3 boards to show its applicability and scalability in a real-world IoT framework.
- 25 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (10) , 5861-5871
Light-Weight Low-Latency Reconfigurable CORDIC Architecture With a New Non-Power-of-2 Angle Set of Microrotations
Trigonometric functions are widely used in various scientific and engineering applications, including digital signal processing. This article proposes a low-complexity low-latency reconfigurable CORDIC (COordinate Rotation Digital Computer) algorithm that computes trigonometric functions, which is based on a novel angle set of microrotations. The CORDIC algorithm is an iterative algorithm that reduces the performance due to the high latency of computation. The proposed pipelined scale-free CORDIC algorithm computes the functions in only four iterations with an optimized angle set of microrotations with high accuracy. It can be configured in two modes with different trajectories 1) rotation-mode; and 2) vectoring-mode for circular and hyperbolic trajectories. On-board FPGA implementation of the proposed architecture consumes 34.95% less register, 48.12% less occupied slices, 19.06% less LUTs, 66.51% less slice-latency product (SLP), 59.29% less LUT-latency product (LLP), and 71.23% less register-latency product (RLP) compared to the best existing state-of-the-art CORDIC architectures. Moreover, the latency of the proposed architecture is 72.34% less than the best of existing designs. Besides, the parasitic-extracted layout simulation results of the proposed design across different word lengths are reported in TSMC 65-nm CMOS technology for the validation of the proposed architecture.
- 25 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 72 (10) , 5861-5871
Light-Weight Low-Latency Reconfigurable CORDIC Architecture With a New Non-Power-of-2 Angle Set of Microrotations
Trigonometric functions are widely used in various scientific and engineering applications, including digital signal processing. This article proposes a low-complexity low-latency reconfigurable CORDIC (COordinate Rotation Digital Computer) algorithm that computes trigonometric functions, which is based on a novel angle set of microrotations. The CORDIC algorithm is an iterative algorithm that reduces the performance due to the high latency of computation. The proposed pipelined scale-free CORDIC algorithm computes the functions in only four iterations with an optimized angle set of microrotations with high accuracy. It can be configured in two modes with different trajectories 1) rotation-mode; and 2) vectoring-mode for circular and hyperbolic trajectories. On-board FPGA implementation of the proposed architecture consumes 34.95% less register, 48.12% less occupied slices, 19.06% less LUTs, 66.51% less slice-latency product (SLP), 59.29% less LUT-latency product (LLP), and 71.23% less register-latency product (RLP) compared to the best existing state-of-the-art CORDIC architectures. Moreover, the latency of the proposed architecture is 72.34% less than the best of existing designs. Besides, the parasitic-extracted layout simulation results of the proposed design across different word lengths are reported in TSMC 65-nm CMOS technology for the validation of the proposed architecture.
- 2 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4586-4596
Process-Variation-Aware In-Memory Computation With Improved Linearity Using On-Chip Configurable Current-Steering Thermometric DAC
The in-memory computation (IMC) is a potential technique to improve the speed and energy efficiency of data-intensive designs. However, the scalability of IMC to large systems is hindered by the non-linearities of analog multiply-and-accumulate (MAC) operations and process variation, which impacts the precision of high bit-width MAC operations. In this paper, we present an IMC architecture that is capable of performing multi-bit MAC operations with improved speed, linearity, and computational accuracy. To improve the speed/linearity of the IMC-MAC operations, the image and weight data are applied by using the pulse amplitude modulation (PAM) and thermometric techniques, respectively. Although the PAM technique improves the speed of the IMC-MAC operations, it has linearity issues that need to be addressed. Based on the detailed linearity analysis of the IMC-MAC circuit, we proposed two approaches to improve the linearity and the signal margin (SM) of the IMC architecture. The proposed configurable current steering thermometric digital-to-analog converter (CST-DAC) array is employed to provide the PAM signals with various dynamic ranges and non-linear gaps that are required to improve the linearity/SM. The proposed combined PAM and thermometric IMC (PT-IMC) architecture is designed and fabricated in the TSMC 180-nm CMOS process. The post-silicon calibration of the design point mitigates the process-variation issues and provides the maximum SM (close to the simulation results). Furthermore, the proposed PT-IMC architecture performs MNIST/CIFAR-10 data set classification with an accuracy of 98 $\%$ /88 $\%$ . In addition, the PT-IMC architecture achieves a peak throughput of 12.41 GOPS, a normalized energy efficiency of 30.64 TOPS/W, a normalized figure-of-merit (FOM) of 3039, a loss in the SM of 8.3 $\%$ with respect to the ideal SM, and a computational error of 0.41 $\%$ .
- 2 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4586-4596
Process-Variation-Aware In-Memory Computation With Improved Linearity Using On-Chip Configurable Current-Steering Thermometric DAC
The in-memory computation (IMC) is a potential technique to improve the speed and energy efficiency of data-intensive designs. However, the scalability of IMC to large systems is hindered by the non-linearities of analog multiply-and-accumulate (MAC) operations and process variation, which impacts the precision of high bit-width MAC operations. In this paper, we present an IMC architecture that is capable of performing multi-bit MAC operations with improved speed, linearity, and computational accuracy. To improve the speed/linearity of the IMC-MAC operations, the image and weight data are applied by using the pulse amplitude modulation (PAM) and thermometric techniques, respectively. Although the PAM technique improves the speed of the IMC-MAC operations, it has linearity issues that need to be addressed. Based on the detailed linearity analysis of the IMC-MAC circuit, we proposed two approaches to improve the linearity and the signal margin (SM) of the IMC architecture. The proposed configurable current steering thermometric digital-to-analog converter (CST-DAC) array is employed to provide the PAM signals with various dynamic ranges and non-linear gaps that are required to improve the linearity/SM. The proposed combined PAM and thermometric IMC (PT-IMC) architecture is designed and fabricated in the TSMC 180-nm CMOS process. The post-silicon calibration of the design point mitigates the process-variation issues and provides the maximum SM (close to the simulation results). Furthermore, the proposed PT-IMC architecture performs MNIST/CIFAR-10 data set classification with an accuracy of 98 $\%$ /88 $\%$ . In addition, the PT-IMC architecture achieves a peak throughput of 12.41 GOPS, a normalized energy efficiency of 30.64 TOPS/W, a normalized figure-of-merit (FOM) of 3039, a loss in the SM of 8.3 $\%$ with respect to the ideal SM, and a computational error of 0.41 $\%$ .
- 27 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 71 (1) , 308-315
On-Chip Characterization of Ultra-Low Voltage Standard Cell Library Considering Input Slew and Output Capacitance
The propagation delay of a logic gate in a standard cell library (SCL) primarily depends on the input transition time (slew) and output load capacitance. In this article, we proposed an area-efficient test structure, which can measure the delay of the logic gate considering input slew and output load conditions at ultra-low voltage. The proposed circuit consists of a reconfigurable ring oscillator (RRO), which can configure the input slew and output load of each logic gate in the ring oscillator (RO) during the delay measurement. A test chip has been fabricated in an industrial 130 n m process node to validate the proposed concept. Within-die and die-to-die variations of gate delay across different slew and load conditions are studied and analyzed for sub-threshold regions of operation. Ten chips are tested and measured across supply voltages with output loads and input slew conditions. The measured results agree with the simulation results.
- 27 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 71 (1) , 308-315
On-Chip Characterization of Ultra-Low Voltage Standard Cell Library Considering Input Slew and Output Capacitance
The propagation delay of a logic gate in a standard cell library (SCL) primarily depends on the input transition time (slew) and output load capacitance. In this article, we proposed an area-efficient test structure, which can measure the delay of the logic gate considering input slew and output load conditions at ultra-low voltage. The proposed circuit consists of a reconfigurable ring oscillator (RRO), which can configure the input slew and output load of each logic gate in the ring oscillator (RO) during the delay measurement. A test chip has been fabricated in an industrial 130 n m process node to validate the proposed concept. Within-die and die-to-die variations of gate delay across different slew and load conditions are studied and analyzed for sub-threshold regions of operation. Ten chips are tested and measured across supply voltages with output loads and input slew conditions. The measured results agree with the simulation results.
- 26 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 31 (8) , 1167-1177
An Efficient Scaling-Free Folded Hyperbolic CORDIC Design Using a Novel Low-Complexity Power-of-2 Taylor Series Approximation
Hyperbolic trigonometric functions are widely used in several engineering and scientific applications, including digital signal processing (DSP), communication systems, and many others. In this article, we propose a scaling-free hyperbolic coordinate rotation digital computer (CORDIC) algorithm and its architecture based on a novel power-of-2 coefficient low-complexity Taylor series approximation to implement $\sinh$ and $\cosh$ functions. CORDIC architectures are generally slow due to their high latency of computation. The proposed architecture reduces the latency and achieves the desired precision with only four iterations where an optimized angle set comprised of six CORDIC microrotations are mapped into a four-stage folded-pipeline structure leveraging mutually exclusive behavior of two pairs of microrotations. The proposed design is implemented on field-programmable gate arrays (FPGAs) Xilinx Zedboard using 65.38% less registers with $\sim$ 63.63% less latency and 48.97% less power consumption compared with the best of the existing designs. The proposed design is synthesized by Synopsys Design Compiler and place and route (PnR) tool using Taiwan Semiconductor Manufacturing Company (TSMC) 65-nm CMOS process. It consumes $\sim$ 76.31% less area, 68.75% less computational delay, and 68.92% less power consumption compared with the best of the existing designs. Moreover, the proposed architecture involves 46.89% less energy per output (EPO) than the best of the existing designs. The error–energy performance (EEP) and the error–area performance (EAP) of the proposed design are, respectively, $\sim$ 1.25 times and $\sim$ 2.8 times better than that of the best of the existing designs. Besides, the proposed architecture is also implemented and verified on a silicon chip in the TSMC 180-nm CMOS process for the validation of the algorithm and architecture.
- 26 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Very Large Scale Integration (VLSI) Systems
- Vol. 31 (8) , 1167-1177
An Efficient Scaling-Free Folded Hyperbolic CORDIC Design Using a Novel Low-Complexity Power-of-2 Taylor Series Approximation
Hyperbolic trigonometric functions are widely used in several engineering and scientific applications, including digital signal processing (DSP), communication systems, and many others. In this article, we propose a scaling-free hyperbolic coordinate rotation digital computer (CORDIC) algorithm and its architecture based on a novel power-of-2 coefficient low-complexity Taylor series approximation to implement $\sinh$ and $\cosh$ functions. CORDIC architectures are generally slow due to their high latency of computation. The proposed architecture reduces the latency and achieves the desired precision with only four iterations where an optimized angle set comprised of six CORDIC microrotations are mapped into a four-stage folded-pipeline structure leveraging mutually exclusive behavior of two pairs of microrotations. The proposed design is implemented on field-programmable gate arrays (FPGAs) Xilinx Zedboard using 65.38% less registers with $\sim$ 63.63% less latency and 48.97% less power consumption compared with the best of the existing designs. The proposed design is synthesized by Synopsys Design Compiler and place and route (PnR) tool using Taiwan Semiconductor Manufacturing Company (TSMC) 65-nm CMOS process. It consumes $\sim$ 76.31% less area, 68.75% less computational delay, and 68.92% less power consumption compared with the best of the existing designs. Moreover, the proposed architecture involves 46.89% less energy per output (EPO) than the best of the existing designs. The error–energy performance (EEP) and the error–area performance (EAP) of the proposed design are, respectively, $\sim$ 1.25 times and $\sim$ 2.8 times better than that of the best of the existing designs. Besides, the proposed architecture is also implemented and verified on a silicon chip in the TSMC 180-nm CMOS process for the validation of the algorithm and architecture.
- 25 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nanotechnology
- Vol. 22 (1536125X) , 222-227
A Configurable 10T SRAM-Based IMC Accelerator With Scaled-Voltage-Based Pulse Count Modulation for MAC and High-Throughput XAC
This work proposes a 10 T SRAM-based in-memory computation (IMC) architecture that can be configured to perform linear multiply-and-accumulate (MAC) operations and high-throughput XNOR-and-accumulate (XAC) operations. The IMC-MAC operation is performed by using the proposed scaled-voltage-based pulse count modulation (PCM) technique, which improves the linearity and signal margin of the MAC operation. The IMC-XAC operation is performed by using the proposed single capacitor discharge (SCD) approach with various advantages such as no deterministic error in XAC output, low latency, and less variation compared to the traditional charge sharing (TCS)-based XAC operation. The post-layout simulations of the proposed IMC architecture in a 65-nm CMOS process shows that the IMC architecture achieves a signal margin of 44.3 mV using the proposed scaled-voltage-based PCM approach in IMC-MAC mode whereas 37% less variation using the proposed SCD approach in IMC-XAC mode. In IMC-MAC mode, we achieve a 54.6 GOPS, 273 TOPS/W, and 98.67%/88.72% classification accuracy on MNIST/CIFAR-10 dataset using the convolution neural network (CNN)/ResNet20 algorithm. In IMC-XAC mode, we achieve a 3276.8 GOPS, 1092.2 TOPS/W, and 97.12% classification accuracy on the MNIST dataset using the binary neural network (BNN) algorithm.
- 25 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Nanotechnology
- Vol. 22 (1536125X) , 222-227
A Configurable 10T SRAM-Based IMC Accelerator With Scaled-Voltage-Based Pulse Count Modulation for MAC and High-Throughput XAC
This work proposes a 10 T SRAM-based in-memory computation (IMC) architecture that can be configured to perform linear multiply-and-accumulate (MAC) operations and high-throughput XNOR-and-accumulate (XAC) operations. The IMC-MAC operation is performed by using the proposed scaled-voltage-based pulse count modulation (PCM) technique, which improves the linearity and signal margin of the MAC operation. The IMC-XAC operation is performed by using the proposed single capacitor discharge (SCD) approach with various advantages such as no deterministic error in XAC output, low latency, and less variation compared to the traditional charge sharing (TCS)-based XAC operation. The post-layout simulations of the proposed IMC architecture in a 65-nm CMOS process shows that the IMC architecture achieves a signal margin of 44.3 mV using the proposed scaled-voltage-based PCM approach in IMC-MAC mode whereas 37% less variation using the proposed SCD approach in IMC-XAC mode. In IMC-MAC mode, we achieve a 54.6 GOPS, 273 TOPS/W, and 98.67%/88.72% classification accuracy on MNIST/CIFAR-10 dataset using the convolution neural network (CNN)/ResNet20 algorithm. In IMC-XAC mode, we achieve a 3276.8 GOPS, 1092.2 TOPS/W, and 97.12% classification accuracy on the MNIST dataset using the binary neural network (BNN) algorithm.
- 17 December 2022
- Published by Springer Nature
- p. 185-198
Low Power Dual-Band Current Reuse-based LC-Voltage Controlled Oscillator with Shared Inductor for IoT Applications
- 17 December 2022
- Published by Springer Nature
- p. 185-198
Low Power Dual-Band Current Reuse-based LC-Voltage Controlled Oscillator with Shared Inductor for IoT Applications
- 17 December 2022
- Published by Springer Nature
- p. 234-248
Highly Non-linear Feed-Forward Arbiter PUF Against Machine Learning Attacks
- 17 December 2022
- Published by Springer Nature
- p. 234-248
Highly Non-linear Feed-Forward Arbiter PUF Against Machine Learning Attacks