Theses and Dissertations
This collection includes both ASU Theses and Dissertations, submitted by graduate students, and the Barrett, Honors College theses submitted by undergraduate students.
Displaying 1 - 10 of 35
Filtering by
- Genre: Doctoral Dissertation
- Creators: Seo, Jae-Sun
Description
Photovoltaic (PV) systems are affected by converter losses, partial shading and other mismatches in the panels. This dissertation introduces a sub-panel maximum power point tracking (MPPT) architecture together with an integrated CMOS current sensor circuit on a chip to reduce the mismatch effects, losses and increase the efficiency of the PV system. The sub-panel MPPT increases the efficiency of the PV during the shading and replaces the bypass diodes in the panels with an integrated MPPT and DC-DC regulator. For the integrated MPPT and regulator, the research developed an integrated standard CMOS low power and high common mode range Current-to-Digital Converter (IDC) circuit and its application for DC-DC regulator and MPPT. The proposed charge based CMOS switched-capacitor circuit directly digitizes the output current of the DC-DC regulator without an analog-to-digital converter (ADC) and the need for high-voltage process technology. Compared to the resistor based current-sensing methods that requires current-to-voltage circuit, gain block and ADC, the proposed CMOS IDC is a low-power efficient integrated circuit that achieves high resolution, lower complexity, and lower power consumption. The IDC circuit is fabricated on a 0.7 um CMOS process, occupies 2mm x 2mm and consumes less than 27mW. The IDC circuit has been tested and used for boost DC-DC regulator and MPPT for photo-voltaic system. The DC-DC converter has an efficiency of 95%. The sub-module level power optimization improves the output power of a shaded panel by up to 20%, compared to panel MPPT with bypass diodes.
ContributorsMarti-Arbona, Edgar (Author) / Kiaei, Sayfe (Thesis advisor) / Bakkaloglu, Bertan (Committee member) / Kitchen, Jennifer (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2014
Description
Static CMOS logic has remained the dominant design style of digital systems for
more than four decades due to its robustness and near zero standby current. Static
CMOS logic circuits consist of a network of combinational logic cells and clocked sequential
elements, such as latches and flip-flops that are used for sequencing computations
over time. The majority of the digital design techniques to reduce power, area, and
leakage over the past four decades have focused almost entirely on optimizing the
combinational logic. This work explores alternate architectures for the flip-flops for
improving the overall circuit performance, power and area. It consists of three main
sections.
First, is the design of a multi-input configurable flip-flop structure with embedded
logic. A conventional D-type flip-flop may be viewed as realizing an identity function,
in which the output is simply the value of the input sampled at the clock edge. In
contrast, the proposed multi-input flip-flop, named PNAND, can be configured to
realize one of a family of Boolean functions called threshold functions. In essence,
the PNAND is a circuit implementation of the well-known binary perceptron. Unlike
other reconfigurable circuits, a PNAND can be configured by simply changing the
assignment of signals to its inputs. Using a standard cell library of such gates, a technology
mapping algorithm can be applied to transform a given netlist into one with
an optimal mixture of conventional logic gates and threshold gates. This approach
was used to fabricate a 32-bit Wallace Tree multiplier and a 32-bit booth multiplier
in 65nm LP technology. Simulation and chip measurements show more than 30%
improvement in dynamic power and more than 20% reduction in core area.
The functional yield of the PNAND reduces with geometry and voltage scaling.
The second part of this research investigates the use of two mechanisms to improve
the robustness of the PNAND circuit architecture. One is the use of forward and reverse body biases to change the device threshold and the other is the use of RRAM
devices for low voltage operation.
The third part of this research focused on the design of flip-flops with non-volatile
storage. Spin-transfer torque magnetic tunnel junctions (STT-MTJ) are integrated
with both conventional D-flipflop and the PNAND circuits to implement non-volatile
logic (NVL). These non-volatile storage enhanced flip-flops are able to save the state of
system locally when a power interruption occurs. However, manufacturing variations
in the STT-MTJs and in the CMOS transistors significantly reduce the yield, leading
to an overly pessimistic design and consequently, higher energy consumption. A
detailed analysis of the design trade-offs in the driver circuitry for performing backup
and restore, and a novel method to design the energy optimal driver for a given yield is
presented. Efficient designs of two nonvolatile flip-flop (NVFF) circuits are presented,
in which the backup time is determined on a per-chip basis, resulting in minimizing
the energy wastage and satisfying the yield constraint. To achieve a yield of 98%,
the conventional approach would have to expend nearly 5X more energy than the
minimum required, whereas the proposed tunable approach expends only 26% more
energy than the minimum. A non-volatile threshold gate architecture NV-TLFF are
designed with the same backup and restore circuitry in 65nm technology. The embedded
logic in NV-TLFF compensates performance overhead of NVL. This leads to the
possibility of zero-overhead non-volatile datapath circuits. An 8-bit multiply-and-
accumulate (MAC) unit is designed to demonstrate the performance benefits of the
proposed architecture. Based on the results of HSPICE simulations, the MAC circuit
with the proposed NV-TLFF cells is shown to consume at least 20% less power and
area as compared to the circuit designed with conventional DFFs, without sacrificing
any performance.
more than four decades due to its robustness and near zero standby current. Static
CMOS logic circuits consist of a network of combinational logic cells and clocked sequential
elements, such as latches and flip-flops that are used for sequencing computations
over time. The majority of the digital design techniques to reduce power, area, and
leakage over the past four decades have focused almost entirely on optimizing the
combinational logic. This work explores alternate architectures for the flip-flops for
improving the overall circuit performance, power and area. It consists of three main
sections.
First, is the design of a multi-input configurable flip-flop structure with embedded
logic. A conventional D-type flip-flop may be viewed as realizing an identity function,
in which the output is simply the value of the input sampled at the clock edge. In
contrast, the proposed multi-input flip-flop, named PNAND, can be configured to
realize one of a family of Boolean functions called threshold functions. In essence,
the PNAND is a circuit implementation of the well-known binary perceptron. Unlike
other reconfigurable circuits, a PNAND can be configured by simply changing the
assignment of signals to its inputs. Using a standard cell library of such gates, a technology
mapping algorithm can be applied to transform a given netlist into one with
an optimal mixture of conventional logic gates and threshold gates. This approach
was used to fabricate a 32-bit Wallace Tree multiplier and a 32-bit booth multiplier
in 65nm LP technology. Simulation and chip measurements show more than 30%
improvement in dynamic power and more than 20% reduction in core area.
The functional yield of the PNAND reduces with geometry and voltage scaling.
The second part of this research investigates the use of two mechanisms to improve
the robustness of the PNAND circuit architecture. One is the use of forward and reverse body biases to change the device threshold and the other is the use of RRAM
devices for low voltage operation.
The third part of this research focused on the design of flip-flops with non-volatile
storage. Spin-transfer torque magnetic tunnel junctions (STT-MTJ) are integrated
with both conventional D-flipflop and the PNAND circuits to implement non-volatile
logic (NVL). These non-volatile storage enhanced flip-flops are able to save the state of
system locally when a power interruption occurs. However, manufacturing variations
in the STT-MTJs and in the CMOS transistors significantly reduce the yield, leading
to an overly pessimistic design and consequently, higher energy consumption. A
detailed analysis of the design trade-offs in the driver circuitry for performing backup
and restore, and a novel method to design the energy optimal driver for a given yield is
presented. Efficient designs of two nonvolatile flip-flop (NVFF) circuits are presented,
in which the backup time is determined on a per-chip basis, resulting in minimizing
the energy wastage and satisfying the yield constraint. To achieve a yield of 98%,
the conventional approach would have to expend nearly 5X more energy than the
minimum required, whereas the proposed tunable approach expends only 26% more
energy than the minimum. A non-volatile threshold gate architecture NV-TLFF are
designed with the same backup and restore circuitry in 65nm technology. The embedded
logic in NV-TLFF compensates performance overhead of NVL. This leads to the
possibility of zero-overhead non-volatile datapath circuits. An 8-bit multiply-and-
accumulate (MAC) unit is designed to demonstrate the performance benefits of the
proposed architecture. Based on the results of HSPICE simulations, the MAC circuit
with the proposed NV-TLFF cells is shown to consume at least 20% less power and
area as compared to the circuit designed with conventional DFFs, without sacrificing
any performance.
ContributorsYang, Jinghua (Author) / Vrudhula, Sarma (Thesis advisor) / Barnaby, Hugh (Committee member) / Cao, Yu (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2018
Description
Over the past few decades, the silicon complementary-metal-oxide-semiconductor (CMOS) technology has been greatly scaled down to achieve higher performance, density and lower power consumption. As the device dimension is approaching its fundamental physical limit, there is an increasing demand for exploration of emerging devices with distinct operating principles from conventional CMOS. In recent years, many efforts have been devoted in the research of next-generation emerging non-volatile memory (eNVM) technologies, such as resistive random access memory (RRAM) and phase change memory (PCM), to replace conventional digital memories (e.g. SRAM) for implementation of synapses in large-scale neuromorphic computing systems.
Essentially being compact and “analog”, these eNVM devices in a crossbar array can compute vector-matrix multiplication in parallel, significantly speeding up the machine/deep learning algorithms. However, non-ideal eNVM device and array properties may hamper the learning accuracy. To quantify their impact, the sparse coding algorithm was used as a starting point, where the strategies to remedy the accuracy loss were proposed, and the circuit-level design trade-offs were also analyzed. At architecture level, the parallel “pseudo-crossbar” array to prevent the write disturbance issue was presented. The peripheral circuits to support various parallel array architectures were also designed. One key component is the read circuit that employs the principle of integrate-and-fire neuron model to convert the analog column current to digital output. However, the read circuit is not area-efficient, which was proposed to be replaced with a compact two-terminal oscillation neuron device that exhibits metal-insulator-transition phenomenon.
To facilitate the design exploration, a circuit-level macro simulator “NeuroSim” was developed in C++ to estimate the area, latency, energy and leakage power of various neuromorphic architectures. NeuroSim provides a wide variety of design options at the circuit/device level. NeuroSim can be used alone or as a supporting module to provide circuit-level performance estimation in neural network algorithms. A 2-layer multilayer perceptron (MLP) simulator with integration of NeuroSim was demonstrated to evaluate both the learning accuracy and circuit-level performance metrics for the online learning and offline classification, as well as to study the impact of eNVM reliability issues such as data retention and write endurance on the learning performance.
Essentially being compact and “analog”, these eNVM devices in a crossbar array can compute vector-matrix multiplication in parallel, significantly speeding up the machine/deep learning algorithms. However, non-ideal eNVM device and array properties may hamper the learning accuracy. To quantify their impact, the sparse coding algorithm was used as a starting point, where the strategies to remedy the accuracy loss were proposed, and the circuit-level design trade-offs were also analyzed. At architecture level, the parallel “pseudo-crossbar” array to prevent the write disturbance issue was presented. The peripheral circuits to support various parallel array architectures were also designed. One key component is the read circuit that employs the principle of integrate-and-fire neuron model to convert the analog column current to digital output. However, the read circuit is not area-efficient, which was proposed to be replaced with a compact two-terminal oscillation neuron device that exhibits metal-insulator-transition phenomenon.
To facilitate the design exploration, a circuit-level macro simulator “NeuroSim” was developed in C++ to estimate the area, latency, energy and leakage power of various neuromorphic architectures. NeuroSim provides a wide variety of design options at the circuit/device level. NeuroSim can be used alone or as a supporting module to provide circuit-level performance estimation in neural network algorithms. A 2-layer multilayer perceptron (MLP) simulator with integration of NeuroSim was demonstrated to evaluate both the learning accuracy and circuit-level performance metrics for the online learning and offline classification, as well as to study the impact of eNVM reliability issues such as data retention and write endurance on the learning performance.
ContributorsChen, Pai-Yu (Author) / Yu, Shimeng (Thesis advisor) / Cao, Yu (Committee member) / Seo, Jae-Sun (Committee member) / Chakrabarti, Chaitali (Committee member) / Arizona State University (Publisher)
Created2018
Description
A 4-phase, quasi-current-mode hysteretic buck converter with digital frequency synchronization, online comparator offset-calibration and digital current sharing control is presented. The switching frequency of the hysteretic converter is digitally synchronized to the input clock reference with less than ±1.5% error in the switching frequency range of 3-9.5MHz. The online offset calibration cancels the input-referred offset of the hysteretic comparator and enables ±1.1% voltage regulation accuracy. Maximum current-sharing error of ±3.6% is achieved by a duty-cycle-calibrated delay line based PWM generator, without affecting the phase synchronization timing sequence. In light load conditions, individual converter phases can be disabled, and the final stage power converter output stage is segmented for high efficiency. The DC-DC converter achieves 93% peak efficiency for Vi = 2V and Vo = 1.6V.
ContributorsSun, Ming (Author) / Bakkaloglu, Bertan (Thesis advisor) / Garrity, Douglas (Committee member) / Seo, Jae-Sun (Committee member) / Kitchen, Jennifer (Committee member) / Arizona State University (Publisher)
Created2017
Description
Complex electronic systems include multiple power domains and drastically varying dynamic power consumption patterns, requiring the use of multiple power conversion and regulation units. High frequency switching converters have been gaining prominence in the DC-DC converter market due to smaller solution size (higher power density) and higher efficiency. As the filter components become smaller in value and size, they are unfortunately also subject to higher process variations and worse degradation profiles jeopardizing stable operation of the power supply. This dissertation presents techniques to track changes in the dynamic loop characteristics of the DC-DC converters without disturbing the normal mode of operation. A digital pseudo-noise (PN) based stimulus is used to excite the DC-DC system at various circuit nodes to calculate the corresponding closed-loop impulse response. The test signal energy is spread over a wide bandwidth and the signal analysis is achieved by correlating the PN input sequence with the disturbed output generated, thereby
accumulating the desired behavior over time. A mixed-signal cross-correlation circuit is used to derive on-chip impulse responses, with smaller memory and lower computational requirement in comparison to a digital correlator approach. Model reference based parametric and non-parametric techniques are discussed to analyze the impulse response results in both time and frequency domain. The proposed techniques can extract open-loop phase margin and closed-loop unity-gain frequency within 5.2% and 4.1% error, respectively, for the load current range of 30-200mA. Converter parameters such as natural frequency (ω_n ), quality factor (Q), and center frequency (ω_c ) can be estimated within 3.6%, 4.7%, and 3.8% error respectively, over load inductance of 4.7-10.3µH, and filter capacitance of 200-400nF. A 5-MHz switching frequency, 5-8.125V input voltage range, voltage-mode controlled DC-DC buck converter is designed for the proposed built-in self-test (BIST) analysis. The converter output voltage range is 3.3-5V and the supported maximum
load current is 450mA. The peak efficiency of the converter is 87.93%. The proposed converter is fabricated on a 0.6µm 6-layer-metal Silicon-On-Insulator (SOI) technology with a die area of 9mm^2 . The area impact due to the system identification blocks including related I/O structures is 3.8% and they consume 530µA quiescent current during operation.
accumulating the desired behavior over time. A mixed-signal cross-correlation circuit is used to derive on-chip impulse responses, with smaller memory and lower computational requirement in comparison to a digital correlator approach. Model reference based parametric and non-parametric techniques are discussed to analyze the impulse response results in both time and frequency domain. The proposed techniques can extract open-loop phase margin and closed-loop unity-gain frequency within 5.2% and 4.1% error, respectively, for the load current range of 30-200mA. Converter parameters such as natural frequency (ω_n ), quality factor (Q), and center frequency (ω_c ) can be estimated within 3.6%, 4.7%, and 3.8% error respectively, over load inductance of 4.7-10.3µH, and filter capacitance of 200-400nF. A 5-MHz switching frequency, 5-8.125V input voltage range, voltage-mode controlled DC-DC buck converter is designed for the proposed built-in self-test (BIST) analysis. The converter output voltage range is 3.3-5V and the supported maximum
load current is 450mA. The peak efficiency of the converter is 87.93%. The proposed converter is fabricated on a 0.6µm 6-layer-metal Silicon-On-Insulator (SOI) technology with a die area of 9mm^2 . The area impact due to the system identification blocks including related I/O structures is 3.8% and they consume 530µA quiescent current during operation.
ContributorsBeohar, Navankur (Author) / Bakkaloglu, Bertan (Thesis advisor) / Ozev, Sule (Committee member) / Ayyanar, Raja (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2017
Description
Portable devices often require multiple power management IC (PMIC) to power different sub-modules, Li-ion batteries are well suited for portable devices because of its small size, high energy density and long life cycle. Since Li-ion battery is the major power source for portable device, fast and high-efficiency battery charging solution has become a major requirement in portable device application.
In the first part of dissertation, a high performance Li-ion switching battery charger is proposed. Cascaded two loop (CTL) control architecture is used for seamless CC-CV transition, time based technique is utilized to minimize controller area and power consumption. Time domain controller is implemented by using voltage controlled oscillator (VCO) and voltage controlled delay line (VCDL). Several efficiency improvement techniques such as segmented power-FET, quasi-zero voltage switching (QZVS) and switching frequency reduction are proposed. The proposed switching battery charger is able to provide maximum 2 A charging current and has an peak efficiency of 93.3%. By configure the charger as boost converter, the charger is able to provide maximum 1.5 A charging current while achieving 96.3% peak efficiency.
The second part of dissertation presents a digital low dropout regulator (DLDO) for system on a chip (SoC) in portable devices application. The proposed DLDO achieve fast transient settling time, lower undershoot/overshoot and higher PSR performance compared to state of the art. By having a good PSR performance, the proposed DLDO is able to power mixed signal load. To achieve a fast load transient response, a load transient detector (LTD) enables boost mode operation of the digital PI controller. The boost mode operation achieves sub microsecond settling time, and reduces the settling time by 50% to 250 ns, undershoot/overshoot by 35% to 250 mV and 17% to 125 mV without compromising the system stability.
In the first part of dissertation, a high performance Li-ion switching battery charger is proposed. Cascaded two loop (CTL) control architecture is used for seamless CC-CV transition, time based technique is utilized to minimize controller area and power consumption. Time domain controller is implemented by using voltage controlled oscillator (VCO) and voltage controlled delay line (VCDL). Several efficiency improvement techniques such as segmented power-FET, quasi-zero voltage switching (QZVS) and switching frequency reduction are proposed. The proposed switching battery charger is able to provide maximum 2 A charging current and has an peak efficiency of 93.3%. By configure the charger as boost converter, the charger is able to provide maximum 1.5 A charging current while achieving 96.3% peak efficiency.
The second part of dissertation presents a digital low dropout regulator (DLDO) for system on a chip (SoC) in portable devices application. The proposed DLDO achieve fast transient settling time, lower undershoot/overshoot and higher PSR performance compared to state of the art. By having a good PSR performance, the proposed DLDO is able to power mixed signal load. To achieve a fast load transient response, a load transient detector (LTD) enables boost mode operation of the digital PI controller. The boost mode operation achieves sub microsecond settling time, and reduces the settling time by 50% to 250 ns, undershoot/overshoot by 35% to 250 mV and 17% to 125 mV without compromising the system stability.
ContributorsLim, Chai Yong (Author) / Kiaei, Sayfe (Thesis advisor) / Bakkaloglu, Bertan (Committee member) / Ogras, Umit Y. (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2018
Description
The increased adoption of Internet-of-Things (IoT) for various applications like smart home, industrial automation, connected vehicles, medical instrumentation, etc. has resulted in a large scale distributed network of sensors, accompanied by their power supply regulator modules, control and data transfer circuitry. Depending on the application, the sensor location can be virtually anywhere and therefore they are typically powered by a localized battery. To ensure long battery-life without replacement, the power consumption of the sensor nodes, the supply regulator and, control and data transmission unit, needs to be very low. Reduction in power consumption in the sensor, control and data transmission is typically done by duty-cycled operation such that they are on periodically only for short bursts of time or turn on only based on a trigger event and are otherwise powered down. These approaches reduce their power consumption significantly and therefore the overall system power is dominated by the consumption in the always-on supply regulator.
Besides having low power consumption, supply regulators for such IoT systems also need to have fast transient response to load current changes during a duty-cycled operation. Supply regulation using low quiescent current low dropout (LDO) regulators helps in extending the battery life of such power aware always-on applications with very long standby time. To serve as a supply regulator for such applications, a 1.24 µA quiescent current NMOS low dropout (LDO) is presented in this dissertation. This LDO uses a hybrid bias current generator (HBCG) to boost its bias current and improve the transient response. A scalable bias-current error amplifier with an on-demand buffer drives the NMOS pass device. The error amplifier is powered with an integrated dynamic frequency charge pump to ensure low dropout voltage. A low-power relaxation oscillator (LPRO) generates the charge pump clocks. Switched-capacitor pole tracking (SCPT) compensation scheme is proposed to ensure stability up to maximum load current of 150 mA for a low-ESR output capacitor range of 1 - 47µF. Designed in a 0.25 µm CMOS process, the LDO has an output voltage range of 1V – 3V, a dropout voltage of 240 mV, and a core area of 0.11 mm2.
Besides having low power consumption, supply regulators for such IoT systems also need to have fast transient response to load current changes during a duty-cycled operation. Supply regulation using low quiescent current low dropout (LDO) regulators helps in extending the battery life of such power aware always-on applications with very long standby time. To serve as a supply regulator for such applications, a 1.24 µA quiescent current NMOS low dropout (LDO) is presented in this dissertation. This LDO uses a hybrid bias current generator (HBCG) to boost its bias current and improve the transient response. A scalable bias-current error amplifier with an on-demand buffer drives the NMOS pass device. The error amplifier is powered with an integrated dynamic frequency charge pump to ensure low dropout voltage. A low-power relaxation oscillator (LPRO) generates the charge pump clocks. Switched-capacitor pole tracking (SCPT) compensation scheme is proposed to ensure stability up to maximum load current of 150 mA for a low-ESR output capacitor range of 1 - 47µF. Designed in a 0.25 µm CMOS process, the LDO has an output voltage range of 1V – 3V, a dropout voltage of 240 mV, and a core area of 0.11 mm2.
ContributorsMagod Ramakrishna, Raveesh (Author) / Bakkaloglu, Bertan (Thesis advisor) / Garrity, Douglas (Committee member) / Kitchen, Jennifer (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2018
Description
Semiconductor memory is a key component of the computing systems. Beyond the conventional memory and data storage applications, in this dissertation, both mainstream and eNVM memory technologies are explored for radiation environment, hardware security system and machine learning applications.
In the radiation environment, e.g. aerospace, the memory devices face different energetic particles. The strike of these energetic particles can generate electron-hole pairs (directly or indirectly) as they pass through the semiconductor device, resulting in photo-induced current, and may change the memory state. First, the trend of radiation effects of the mainstream memory technologies with technology node scaling is reviewed. Then, single event effects of the oxide based resistive switching random memory (RRAM), one of eNVM technologies, is investigated from the circuit-level to the system level.
Physical Unclonable Function (PUF) has been widely investigated as a promising hardware security primitive, which employs the inherent randomness in a physical system (e.g. the intrinsic semiconductor manufacturing variability). In the dissertation, two RRAM-based PUF implementations are proposed for cryptographic key generation (weak PUF) and device authentication (strong PUF), respectively. The performance of the RRAM PUFs are evaluated with experiment and simulation. The impact of non-ideal circuit effects on the performance of the PUFs is also investigated and optimization strategies are proposed to solve the non-ideal effects. Besides, the security resistance against modeling and machine learning attacks is analyzed as well.
Deep neural networks (DNNs) have shown remarkable improvements in various intelligent applications such as image classification, speech classification and object localization and detection. Increasing efforts have been devoted to develop hardware accelerators. In this dissertation, two types of compute-in-memory (CIM) based hardware accelerator designs with SRAM and eNVM technologies are proposed for two binary neural networks, i.e. hybrid BNN (HBNN) and XNOR-BNN, respectively, which are explored for the hardware resource-limited platforms, e.g. edge devices.. These designs feature with high the throughput, scalability, low latency and high energy efficiency. Finally, we have successfully taped-out and validated the proposed designs with SRAM technology in TSMC 65 nm.
Overall, this dissertation paves the paths for memory technologies’ new applications towards the secure and energy-efficient artificial intelligence system.
In the radiation environment, e.g. aerospace, the memory devices face different energetic particles. The strike of these energetic particles can generate electron-hole pairs (directly or indirectly) as they pass through the semiconductor device, resulting in photo-induced current, and may change the memory state. First, the trend of radiation effects of the mainstream memory technologies with technology node scaling is reviewed. Then, single event effects of the oxide based resistive switching random memory (RRAM), one of eNVM technologies, is investigated from the circuit-level to the system level.
Physical Unclonable Function (PUF) has been widely investigated as a promising hardware security primitive, which employs the inherent randomness in a physical system (e.g. the intrinsic semiconductor manufacturing variability). In the dissertation, two RRAM-based PUF implementations are proposed for cryptographic key generation (weak PUF) and device authentication (strong PUF), respectively. The performance of the RRAM PUFs are evaluated with experiment and simulation. The impact of non-ideal circuit effects on the performance of the PUFs is also investigated and optimization strategies are proposed to solve the non-ideal effects. Besides, the security resistance against modeling and machine learning attacks is analyzed as well.
Deep neural networks (DNNs) have shown remarkable improvements in various intelligent applications such as image classification, speech classification and object localization and detection. Increasing efforts have been devoted to develop hardware accelerators. In this dissertation, two types of compute-in-memory (CIM) based hardware accelerator designs with SRAM and eNVM technologies are proposed for two binary neural networks, i.e. hybrid BNN (HBNN) and XNOR-BNN, respectively, which are explored for the hardware resource-limited platforms, e.g. edge devices.. These designs feature with high the throughput, scalability, low latency and high energy efficiency. Finally, we have successfully taped-out and validated the proposed designs with SRAM technology in TSMC 65 nm.
Overall, this dissertation paves the paths for memory technologies’ new applications towards the secure and energy-efficient artificial intelligence system.
ContributorsLiu, Rui (Author) / Yu, Shimeng (Thesis advisor, Committee member) / Cao, Yu (Committee member) / Barnaby, Hugh (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2018
Description
The rapid improvement in computation capability has made deep convolutional neural networks (CNNs) a great success in recent years on many computer vision tasks with significantly improved accuracy. During the inference phase, many applications demand low latency processing of one image with strict power consumption requirement, which reduces the efficiency of GPU and other general-purpose platform, bringing opportunities for specific acceleration hardware, e.g. FPGA, by customizing the digital circuit specific for the deep learning algorithm inference. However, deploying CNNs on portable and embedded systems is still challenging due to large data volume, intensive computation, varying algorithm structures, and frequent memory accesses. This dissertation proposes a complete design methodology and framework to accelerate the inference process of various CNN algorithms on FPGA hardware with high performance, efficiency and flexibility.
As convolution contributes most operations in CNNs, the convolution acceleration scheme significantly affects the efficiency and performance of a hardware CNN accelerator. Convolution involves multiply and accumulate (MAC) operations with four levels of loops. Without fully studying the convolution loop optimization before the hardware design phase, the resulting accelerator can hardly exploit the data reuse and manage data movement efficiently. This work overcomes these barriers by quantitatively analyzing and optimizing the design objectives (e.g. memory access) of the CNN accelerator based on multiple design variables. An efficient dataflow and hardware architecture of CNN acceleration are proposed to minimize the data communication while maximizing the resource utilization to achieve high performance.
Although great performance and efficiency can be achieved by customizing the FPGA hardware for each CNN model, significant efforts and expertise are required leading to long development time, which makes it difficult to catch up with the rapid development of CNN algorithms. In this work, we present an RTL-level CNN compiler that automatically generates customized FPGA hardware for the inference tasks of various CNNs, in order to enable high-level fast prototyping of CNNs from software to FPGA and still keep the benefits of low-level hardware optimization. First, a general-purpose library of RTL modules is developed to model different operations at each layer. The integration and dataflow of physical modules are predefined in the top-level system template and reconfigured during compilation for a given CNN algorithm. The runtime control of layer-by-layer sequential computation is managed by the proposed execution schedule so that even highly irregular and complex network topology, e.g. GoogLeNet and ResNet, can be compiled. The proposed methodology is demonstrated with various CNN algorithms, e.g. NiN, VGG, GoogLeNet and ResNet, on two different standalone FPGAs achieving state-of-the art performance.
Based on the optimized acceleration strategy, there are still a lot of design options, e.g. the degree and dimension of computation parallelism, the size of on-chip buffers, and the external memory bandwidth, which impact the utilization of computation resources and data communication efficiency, and finally affect the performance and energy consumption of the accelerator. The large design space of the accelerator makes it impractical to explore the optimal design choice during the real implementation phase. Therefore, a performance model is proposed in this work to quantitatively estimate the accelerator performance and resource utilization. By this means, the performance bottleneck and design bound can be identified and the optimal design option can be explored early in the design phase.
As convolution contributes most operations in CNNs, the convolution acceleration scheme significantly affects the efficiency and performance of a hardware CNN accelerator. Convolution involves multiply and accumulate (MAC) operations with four levels of loops. Without fully studying the convolution loop optimization before the hardware design phase, the resulting accelerator can hardly exploit the data reuse and manage data movement efficiently. This work overcomes these barriers by quantitatively analyzing and optimizing the design objectives (e.g. memory access) of the CNN accelerator based on multiple design variables. An efficient dataflow and hardware architecture of CNN acceleration are proposed to minimize the data communication while maximizing the resource utilization to achieve high performance.
Although great performance and efficiency can be achieved by customizing the FPGA hardware for each CNN model, significant efforts and expertise are required leading to long development time, which makes it difficult to catch up with the rapid development of CNN algorithms. In this work, we present an RTL-level CNN compiler that automatically generates customized FPGA hardware for the inference tasks of various CNNs, in order to enable high-level fast prototyping of CNNs from software to FPGA and still keep the benefits of low-level hardware optimization. First, a general-purpose library of RTL modules is developed to model different operations at each layer. The integration and dataflow of physical modules are predefined in the top-level system template and reconfigured during compilation for a given CNN algorithm. The runtime control of layer-by-layer sequential computation is managed by the proposed execution schedule so that even highly irregular and complex network topology, e.g. GoogLeNet and ResNet, can be compiled. The proposed methodology is demonstrated with various CNN algorithms, e.g. NiN, VGG, GoogLeNet and ResNet, on two different standalone FPGAs achieving state-of-the art performance.
Based on the optimized acceleration strategy, there are still a lot of design options, e.g. the degree and dimension of computation parallelism, the size of on-chip buffers, and the external memory bandwidth, which impact the utilization of computation resources and data communication efficiency, and finally affect the performance and energy consumption of the accelerator. The large design space of the accelerator makes it impractical to explore the optimal design choice during the real implementation phase. Therefore, a performance model is proposed in this work to quantitatively estimate the accelerator performance and resource utilization. By this means, the performance bottleneck and design bound can be identified and the optimal design option can be explored early in the design phase.
ContributorsMa, Yufei (Author) / Vrudhula, Sarma (Thesis advisor) / Seo, Jae-Sun (Thesis advisor) / Cao, Yu (Committee member) / Barnaby, Hugh (Committee member) / Arizona State University (Publisher)
Created2018
Description
Deep learning (DL) has proved itself be one of the most important developements till date with far reaching impacts in numerous fields like robotics, computer vision, surveillance, speech processing, machine translation, finance, etc. They are now widely used for countless applications because of their ability to generalize real world data, robustness to noise in previously unseen data and high inference accuracy. With the ability to learn useful features from raw sensor data, deep learning algorithms have out-performed tradinal AI algorithms and pushed the boundaries of what can be achieved with AI. In this work, we demonstrate the power of deep learning by developing a neural network to automatically detect cough instances from audio recorded in un-constrained environments. For this, 24 hours long recordings from 9 dierent patients is collected and carefully labeled by medical personel. A pre-processing algorithm is proposed to convert event based cough dataset to a more informative dataset with start and end of coughs and also introduce data augmentation for regularizing the training procedure. The proposed neural network achieves 92.3% leave-one-out accuracy on data captured in real world.
Deep neural networks are composed of multiple layers that are compute/memory intensive. This makes it difficult to execute these algorithms real-time with low power consumption using existing general purpose computers. In this work, we propose hardware accelerators for a traditional AI algorithm based on random forest trees and two representative deep convolutional neural networks (AlexNet and VGG). With the proposed acceleration techniques, ~ 30x performance improvement was achieved compared to CPU for random forest trees. For deep CNNS, we demonstrate that much higher performance can be achieved with architecture space exploration using any optimization algorithms with system level performance and area models for hardware primitives as inputs and goal of minimizing latency with given resource constraints. With this method, ~30GOPs performance was achieved for Stratix V FPGA boards.
Hardware acceleration of DL algorithms alone is not always the most ecient way and sucient to achieve desired performance. There is a huge headroom available for performance improvement provided the algorithms are designed keeping in mind the hardware limitations and bottlenecks. This work achieves hardware-software co-optimization for Non-Maximal Suppression (NMS) algorithm. Using the proposed algorithmic changes and hardware architecture
With CMOS scaling coming to an end and increasing memory bandwidth bottlenecks, CMOS based system might not scale enough to accommodate requirements of more complicated and deeper neural networks in future. In this work, we explore RRAM crossbars and arrays as compact, high performing and energy efficient alternative to CMOS accelerators for deep learning training and inference. We propose and implement RRAM periphery read and write circuits and achieved ~3000x performance improvement in online dictionary learning compared to CPU.
This work also examines the realistic RRAM devices and their non-idealities. We do an in-depth study of the effects of RRAM non-idealities on inference accuracy when a pretrained model is mapped to RRAM based accelerators. To mitigate this issue, we propose Random Sparse Adaptation (RSA), a novel scheme aimed at tuning the model to take care of the faults of the RRAM array on which it is mapped. Our proposed method can achieve inference accuracy much higher than what traditional Read-Verify-Write (R-V-W) method could achieve. RSA can also recover lost inference accuracy 100x ~ 1000x faster compared to R-V-W. Using 32-bit high precision RSA cells, we achieved ~10% higher accuracy using fautly RRAM arrays compared to what can be achieved by mapping a deep network to an 32 level RRAM array with no variations.
Deep neural networks are composed of multiple layers that are compute/memory intensive. This makes it difficult to execute these algorithms real-time with low power consumption using existing general purpose computers. In this work, we propose hardware accelerators for a traditional AI algorithm based on random forest trees and two representative deep convolutional neural networks (AlexNet and VGG). With the proposed acceleration techniques, ~ 30x performance improvement was achieved compared to CPU for random forest trees. For deep CNNS, we demonstrate that much higher performance can be achieved with architecture space exploration using any optimization algorithms with system level performance and area models for hardware primitives as inputs and goal of minimizing latency with given resource constraints. With this method, ~30GOPs performance was achieved for Stratix V FPGA boards.
Hardware acceleration of DL algorithms alone is not always the most ecient way and sucient to achieve desired performance. There is a huge headroom available for performance improvement provided the algorithms are designed keeping in mind the hardware limitations and bottlenecks. This work achieves hardware-software co-optimization for Non-Maximal Suppression (NMS) algorithm. Using the proposed algorithmic changes and hardware architecture
With CMOS scaling coming to an end and increasing memory bandwidth bottlenecks, CMOS based system might not scale enough to accommodate requirements of more complicated and deeper neural networks in future. In this work, we explore RRAM crossbars and arrays as compact, high performing and energy efficient alternative to CMOS accelerators for deep learning training and inference. We propose and implement RRAM periphery read and write circuits and achieved ~3000x performance improvement in online dictionary learning compared to CPU.
This work also examines the realistic RRAM devices and their non-idealities. We do an in-depth study of the effects of RRAM non-idealities on inference accuracy when a pretrained model is mapped to RRAM based accelerators. To mitigate this issue, we propose Random Sparse Adaptation (RSA), a novel scheme aimed at tuning the model to take care of the faults of the RRAM array on which it is mapped. Our proposed method can achieve inference accuracy much higher than what traditional Read-Verify-Write (R-V-W) method could achieve. RSA can also recover lost inference accuracy 100x ~ 1000x faster compared to R-V-W. Using 32-bit high precision RSA cells, we achieved ~10% higher accuracy using fautly RRAM arrays compared to what can be achieved by mapping a deep network to an 32 level RRAM array with no variations.
ContributorsMohanty, Abinash (Author) / Cao, Yu (Thesis advisor) / Seo, Jae-Sun (Committee member) / Vrudhula, Sarma (Committee member) / Chakrabarti, Chaitali (Committee member) / Arizona State University (Publisher)
Created2018