When the application context is ready to accept different levels of exactness in solutions and is supported
by human perception quality, then the term ‘Approximate Computing’ tossed before one decade will
become the first priority . Even though computer hardware and software are working to generate exact
results, approximate results are preferred whenever an error is in predefined bound and adaptive. It will
reduce power demand and critical path delay and improve other circuit metrics. When it comes to
traditional arithmetic circuits, those generating correct results with limitations on performance are rapidly
getting replaced by approximate arithmetic circuits which are the need of the hour, and so on about their
design.
An Efficient Approach Towards Mitigating Soft Errors Riskssipij
Smaller feature size, higher clock frequency and lower power consumption are of core concerns of today’s nano-technology, which has been resulted by continuous downscaling of CMOS technologies. The resultant‘device shrinking’ reduces the soft error tolerance of the VLSI circuits, as very little energy is needed to change their states. Safety critical systems are very sensitive to soft errors. A bit flip due to soft error can change the value of critical variable and consequently the system control flow can completely be changed which leads to system failure. To minimize soft error risks, a novel methodology is proposed to detect and recover from soft errors considering only ‘critical code blocks’ and ‘critical variables’ rather than considering all variables and/or blocks in the whole program. The proposed method shortens space and time overhead in comparison to existing dominant approaches.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to
around the operands to the closest exponent of 2. This way the machine intensive a part of the
multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is
evaluated by comparing its performance with those of some approximate and correct multipliers using
different design parameters. In this proposed approach combined the conventional RoBA multiplier with
Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA
multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved
the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the
DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to around the operands to the closest exponent of 2. This way the machine intensive a part of the multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is evaluated by comparing its performance with those of some approximate and correct multipliers using different design parameters. In this proposed approach combined the conventional RoBA multiplier with Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to around the operands to the closest exponent of 2. This way the machine intensive a part of the multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is evaluated by comparing its performance with those of some approximate and correct multipliers using different design parameters. In this proposed approach combined the conventional RoBA multiplier with Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to
around the operands to the closest exponent of 2. This way the machine intensive a part of the
multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is
evaluated by comparing its performance with those of some approximate and correct multipliers using
different design parameters. In this proposed approach combined the conventional RoBA multiplier with
Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA
multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved
the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the
DSP
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
An optimized cost-based data allocation model for heterogeneous distributed ...IJECEIAES
Continuous attempts have been made to improve the flexibility and effectiveness of distributed computing systems. Extensive effort in the fields of connectivity technologies, network programs, high processing components, and storage helps to improvise results. However, concerns such as slowness in response, long execution time, and long completion time have been identified as stumbling blocks that hinder performance and require additional attention. These defects increased the total system cost and made the data allocation procedure for a geographically dispersed setup difficult. The load-based architectural model has been strengthened to improve data allocation performance. To do this, an abstract job model is employed, and a data query file containing input data is processed on a directed acyclic graph. The jobs are executed on the processing engine with the lowest execution cost, and the system's total cost is calculated. The total cost is computed by summing the costs of communication, computation, and network. The total cost of the system will be reduced using a Swarm intelligence algorithm. In heterogeneous distributed computing systems, the suggested approach attempts to reduce the system's total cost and improve data distribution. According to simulation results, the technique efficiently lowers total system cost and optimizes partitioned data allocation.
An Efficient Approach Towards Mitigating Soft Errors Riskssipij
Smaller feature size, higher clock frequency and lower power consumption are of core concerns of today’s nano-technology, which has been resulted by continuous downscaling of CMOS technologies. The resultant‘device shrinking’ reduces the soft error tolerance of the VLSI circuits, as very little energy is needed to change their states. Safety critical systems are very sensitive to soft errors. A bit flip due to soft error can change the value of critical variable and consequently the system control flow can completely be changed which leads to system failure. To minimize soft error risks, a novel methodology is proposed to detect and recover from soft errors considering only ‘critical code blocks’ and ‘critical variables’ rather than considering all variables and/or blocks in the whole program. The proposed method shortens space and time overhead in comparison to existing dominant approaches.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to
around the operands to the closest exponent of 2. This way the machine intensive a part of the
multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is
evaluated by comparing its performance with those of some approximate and correct multipliers using
different design parameters. In this proposed approach combined the conventional RoBA multiplier with
Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA
multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved
the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the
DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to around the operands to the closest exponent of 2. This way the machine intensive a part of the multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is evaluated by comparing its performance with those of some approximate and correct multipliers using different design parameters. In this proposed approach combined the conventional RoBA multiplier with Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to around the operands to the closest exponent of 2. This way the machine intensive a part of the multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is evaluated by comparing its performance with those of some approximate and correct multipliers using different design parameters. In this proposed approach combined the conventional RoBA multiplier with Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the DSP.
A METHODOLOGY FOR IMPROVEMENT OF ROBA MULTIPLIER FOR ELECTRONIC APPLICATIONSVLSICS Design
In this paper, propose an approximate multiplier that is high speed yet energy efficient. The approach is to
around the operands to the closest exponent of 2. This way the machine intensive a part of the
multiplication is omitted up speed and energy consumption. The potency of the planned multiplier factor is
evaluated by comparing its performance with those of some approximate and correct multipliers using
different design parameters. In this proposed approach combined the conventional RoBA multiplier with
Kogge-stone parallel prefix adder. The results revealed that, in most (all) cases, the newly designed RoBA
multiplier architectures outperformed the corresponding approximate (exact) multipliers. Thus improved
the parameters of RoBA multiplier which can be used in the voice or image smoothing applications in the
DSP
International Journal of Engineering Research and Applications (IJERA) is an open access online peer reviewed international journal that publishes research and review articles in the fields of Computer Science, Neural Networks, Electrical Engineering, Software Engineering, Information Technology, Mechanical Engineering, Chemical Engineering, Plastic Engineering, Food Technology, Textile Engineering, Nano Technology & science, Power Electronics, Electronics & Communication Engineering, Computational mathematics, Image processing, Civil Engineering, Structural Engineering, Environmental Engineering, VLSI Testing & Low Power VLSI Design etc.
An optimized cost-based data allocation model for heterogeneous distributed ...IJECEIAES
Continuous attempts have been made to improve the flexibility and effectiveness of distributed computing systems. Extensive effort in the fields of connectivity technologies, network programs, high processing components, and storage helps to improvise results. However, concerns such as slowness in response, long execution time, and long completion time have been identified as stumbling blocks that hinder performance and require additional attention. These defects increased the total system cost and made the data allocation procedure for a geographically dispersed setup difficult. The load-based architectural model has been strengthened to improve data allocation performance. To do this, an abstract job model is employed, and a data query file containing input data is processed on a directed acyclic graph. The jobs are executed on the processing engine with the lowest execution cost, and the system's total cost is calculated. The total cost is computed by summing the costs of communication, computation, and network. The total cost of the system will be reduced using a Swarm intelligence algorithm. In heterogeneous distributed computing systems, the suggested approach attempts to reduce the system's total cost and improve data distribution. According to simulation results, the technique efficiently lowers total system cost and optimizes partitioned data allocation.
Multilevel Hybrid Cognitive Load Balancing Algorithm for Private/Public Cloud...IDES Editor
Cloud computing is an emerging computing
paradigm. It aims to share data, resources and services
transparently among users of a massive grid. Although the
industry has started selling cloud-computing products,
research challenges in various areas, such as architectural
design, task decomposition, task distribution, load
distribution, load scheduling, task coordination, etc. are still
unclear. Therefore, we study the methods to reason and model
cloud computing as a step towards identifying fundamental
research questions in this paradigm. In this paper, we propose
a model for load distribution on cloud computing by modeling
them as cognitive systems and using aspects which not only
depend on the present state of the system, but also, on a set of
predefined transitions and conditions. The entirety of this
model is then bundled to cater the task of job distribution
using the concept of application metadata. Later, we draw a
qualitative and simulation based summarization for the
proposed model. We finally evaluate the results and draw up
a series of key conclusions in cloud computing for future
exploration.
Comparative Study of Neural Networks Algorithms for Cloud Computing CPU Sched...IJECEIAES
Cloud Computing is the most powerful computing model of our time. While the major IT providers and consumers are competing to exploit the benefits of this computing model in order to thrive their profits, most of the cloud computing platforms are still built on operating systems that uses basic CPU (Core Processing Unit) scheduling algorithms that lacks the intelligence needed for such innovative computing model. Correspdondingly, this paper presents the benefits of applying Artificial Neural Networks algorithms in regards to enhancing CPU scheduling for Cloud Computing model. Furthermore, a set of characteristics and theoretical metrics are proposed for the sake of comparing the different Artificial Neural Networks algorithms and finding the most accurate algorithm for Cloud Computing CPU Scheduling.
Cloud computing is a new computing paradigm that, just as electricity was firstly generated at home and
evolved to be supplied from a few utility providers, aims to transform computing into a utility. It is a mapping
strategy that efficiently equilibrates the task load into multiple computational resources in the network based on the
system status to improve performance. The objective of this research paper is to show the results of Hybrid DEGA,
in which GA is implemented after DE
Fault-Tolerance Aware Multi Objective Scheduling Algorithm for Task Schedulin...csandit
Computational Grid (CG) creates a large heterogeneous and distributed paradigm to manage and execute the applications which are computationally intensive. In grid scheduling tasks are assigned to the proper processors in the grid system to for its execution by considering the execution policy and the optimization objectives. In this paper, makespan and the faulttolerance of the computational nodes of the grid which are the two important parameters for the task execution, are considered and tried to optimize it. As the grid scheduling is considered to be NP-Hard, so a meta-heuristics evolutionary based techniques are often used to find a solution for this. We have proposed a NSGA II for this purpose. The performance estimation ofthe proposed Fault tolerance Aware NSGA II (FTNSGA II) has been done by writing program in Matlab. The simulation results evaluates the performance of the all proposed algorithm and the results of proposed model is compared with existing model Min-Min and Max-Min algorithm which proves effectiveness of the model.
Proactive Scheduling in Cloud ComputingjournalBEEI
Autonomic fault aware scheduling is a feature quite important for cloud computing and it is related to adoption of workload variation. In this context, this paper proposes an fault aware pattern matching autonomic scheduling for cloud computing based on autonomic computing concepts. In order to validate the proposed solution, we performed two experiments one with traditional approach and other other with pattern recognition fault aware approach. The results show the effectiveness of the scheme.
Fpga based efficient multiplier for image processing applications using recur...VLSICS Design
The Digital Image processing applications like medical imaging, satellite imaging, Biometric trait images
etc., rely on multipliers to improve the quality of image. However, existing multiplication techniques
introduce errors in the output with consumption of more time, hence error free high speed multipliers has
to be designed. In this paper we propose FPGA based Recursive Error Free Mitchell Log Multiplier
(REFMLM) for image Filters. The 2x2 error free Mitchell log multiplier is designed with zero error by
introducing error correction term is used in higher order Karastuba-Ofman Multiplier (KOM)
Architectures. The higher order KOM multipliers is decomposed into number of lower order multipliers
using radix 2 till basic multiplier block of order 2x2 which is designed by error free Mitchell log multiplier.
The 8x8 REFMLM is tested for Gaussian filter to remove noise in fingerprint image. The Multiplier is
synthesized using Spartan 3 FPGA family device XC3S1500-5fg320. It is observed that the performance
parameters such as area utilization, speed, error and PSNR are better in the case of proposed architecture
compared to existing architectures.
PERFORMANCE FACTORS OF CLOUD COMPUTING DATA CENTERS USING [(M/G/1) : (∞/GDM O...ijgca
The ever-increasing status of the cloud computing h
ypothesis and the budding concept of federated clou
d
computing have enthused research efforts towards in
tellectual cloud service selection aimed at develop
ing
techniques for enabling the cloud users to gain max
imum benefit from cloud computing by selecting
services which provide optimal performance at lowes
t possible cost. Cloud computing is a novel paradig
m
for the provision of computing infrastructure, whic
h aims to shift the location of the computing
infrastructure to the network in order to reduce th
e maintenance costs of hardware and software resour
ces.
Cloud computing systems vitally provide access to l
arge pools of resources. Resources provided by clou
d
computing systems hide a great deal of services fro
m the user through virtualization. In this paper, t
he
cloud data center is modelled as
queuing system with a single task arrivals
and a task request buffer of infinite capacity.
A Defect Prediction Model for Software Product based on ANFISIJSRD
Artificial intelligence techniques are day by day getting involvement in all the classification and prediction based process like environmental monitoring, stock exchange conditions, biomedical diagnosis, software engineering etc. However still there are yet to be simplify the challenges of selecting training criteria for design of artificial intelligence models used for prediction of results. This work focus on the defect prediction mechanism development using software metric data of KC1.We have taken subtractive clustering approach for generation of fuzzy inference system (FIS).The FIS rules are generated at different radius of influence of input attribute vectors and the developed rules are further modified by ANFIS technique to obtain the prediction of number of defects in software project using fuzzy logic system.
A Defect Prediction Model for Software Product based on ANFISIJSRD
Artificial intelligence techniques are day by day getting involvement in all the classification and prediction based process like environmental monitoring, stock exchange conditions, biomedical diagnosis, software engineering etc. However still there are yet to be simplify the challenges of selecting training criteria for design of artificial intelligence models used for prediction of results. This work focus on the defect prediction mechanism development using software metric data of KC1.We have taken subtractive clustering approach for generation of fuzzy inference system (FIS).The FIS rules are generated at different radius of influence of input attribute vectors and the developed rules are further modified by ANFIS technique to obtain the prediction of number of defects in software project using fuzzy logic system.
CONFIGURABLE TASK MAPPING FOR MULTIPLE OBJECTIVES IN MACRO-PROGRAMMING OF WIR...ijassn
Macro-programming is the new generation advanced method of using Wireless Sensor Network (WSNs), where application developers can extract data from sensor nodes through a high level abstraction of the system. Instead of developing the entire application, task graph representation of the WSN model presents simplified approach of data collection. However, mapping of tasks onto sensor nodes highlights several problems in energy consumption and routing delay. In this paper, we present an efficient hybrid approach of task mapping for WSN – Hybrid Genetic Algorithm, considering multiple objectives of optimization – energy consumption, routing delay and soft real time requirement. We also present a method to configure the algorithm as per user's need by changing the heuristics used for optimization. The trade-off analysis between energy consumption and delivery delay was performed and simulation results are presented. The algorithm is applicable during macro-programming enabling developers to choose a better mapping according to their application requirements.
CONFIGURABLE TASK MAPPING FOR MULTIPLE OBJECTIVES IN MACRO-PROGRAMMING OF WIR...ijassn
Macro-programming is the new generation advanced method of using Wireless Sensor Network (WSNs),
where application developers can extract data from sensor nodes through a high level abstraction of the
system. Instead of developing the entire application, task graph representation of the WSN model presents
simplified approach of data collection.
THE REMOVAL OF NUMERICAL DRIFT FROM SCIENTIFIC MODELSIJSEA
Computer programs often behave differently under different compilers or in different computing
environments. Relative debugging is a collection of techniques by which these differences are analysed.
Differences may arise because of different interpretations of errors in the code, because of bugs in the
compilers or because of numerical drift, and all of these were observed in the present study. Numerical
drift arises when small and acceptable differences in values computed by different systems are
integrated, so that the results drift apart. This is well understood and need not degrade the validity of the
program results. Coding errors and compiler bugs may degrade the results and should be removed. This
paper describes a technique for the comparison of two program runs which removes numerical drift and
therefore exposes coding and compiler errors. The procedure is highly automated and requires very little
intervention by the user. The technique is applied to the Weather Research and Forecasting model, the
most widely used weather and climate modelling code.
Student information management system project report ii.pdfKamal Acharya
Our project explains about the student management. This project mainly explains the various actions related to student details. This project shows some ease in adding, editing and deleting the student details. It also provides a less time consuming process for viewing, adding, editing and deleting the marks of the students.
Multilevel Hybrid Cognitive Load Balancing Algorithm for Private/Public Cloud...IDES Editor
Cloud computing is an emerging computing
paradigm. It aims to share data, resources and services
transparently among users of a massive grid. Although the
industry has started selling cloud-computing products,
research challenges in various areas, such as architectural
design, task decomposition, task distribution, load
distribution, load scheduling, task coordination, etc. are still
unclear. Therefore, we study the methods to reason and model
cloud computing as a step towards identifying fundamental
research questions in this paradigm. In this paper, we propose
a model for load distribution on cloud computing by modeling
them as cognitive systems and using aspects which not only
depend on the present state of the system, but also, on a set of
predefined transitions and conditions. The entirety of this
model is then bundled to cater the task of job distribution
using the concept of application metadata. Later, we draw a
qualitative and simulation based summarization for the
proposed model. We finally evaluate the results and draw up
a series of key conclusions in cloud computing for future
exploration.
Comparative Study of Neural Networks Algorithms for Cloud Computing CPU Sched...IJECEIAES
Cloud Computing is the most powerful computing model of our time. While the major IT providers and consumers are competing to exploit the benefits of this computing model in order to thrive their profits, most of the cloud computing platforms are still built on operating systems that uses basic CPU (Core Processing Unit) scheduling algorithms that lacks the intelligence needed for such innovative computing model. Correspdondingly, this paper presents the benefits of applying Artificial Neural Networks algorithms in regards to enhancing CPU scheduling for Cloud Computing model. Furthermore, a set of characteristics and theoretical metrics are proposed for the sake of comparing the different Artificial Neural Networks algorithms and finding the most accurate algorithm for Cloud Computing CPU Scheduling.
Cloud computing is a new computing paradigm that, just as electricity was firstly generated at home and
evolved to be supplied from a few utility providers, aims to transform computing into a utility. It is a mapping
strategy that efficiently equilibrates the task load into multiple computational resources in the network based on the
system status to improve performance. The objective of this research paper is to show the results of Hybrid DEGA,
in which GA is implemented after DE
Fault-Tolerance Aware Multi Objective Scheduling Algorithm for Task Schedulin...csandit
Computational Grid (CG) creates a large heterogeneous and distributed paradigm to manage and execute the applications which are computationally intensive. In grid scheduling tasks are assigned to the proper processors in the grid system to for its execution by considering the execution policy and the optimization objectives. In this paper, makespan and the faulttolerance of the computational nodes of the grid which are the two important parameters for the task execution, are considered and tried to optimize it. As the grid scheduling is considered to be NP-Hard, so a meta-heuristics evolutionary based techniques are often used to find a solution for this. We have proposed a NSGA II for this purpose. The performance estimation ofthe proposed Fault tolerance Aware NSGA II (FTNSGA II) has been done by writing program in Matlab. The simulation results evaluates the performance of the all proposed algorithm and the results of proposed model is compared with existing model Min-Min and Max-Min algorithm which proves effectiveness of the model.
Proactive Scheduling in Cloud ComputingjournalBEEI
Autonomic fault aware scheduling is a feature quite important for cloud computing and it is related to adoption of workload variation. In this context, this paper proposes an fault aware pattern matching autonomic scheduling for cloud computing based on autonomic computing concepts. In order to validate the proposed solution, we performed two experiments one with traditional approach and other other with pattern recognition fault aware approach. The results show the effectiveness of the scheme.
Fpga based efficient multiplier for image processing applications using recur...VLSICS Design
The Digital Image processing applications like medical imaging, satellite imaging, Biometric trait images
etc., rely on multipliers to improve the quality of image. However, existing multiplication techniques
introduce errors in the output with consumption of more time, hence error free high speed multipliers has
to be designed. In this paper we propose FPGA based Recursive Error Free Mitchell Log Multiplier
(REFMLM) for image Filters. The 2x2 error free Mitchell log multiplier is designed with zero error by
introducing error correction term is used in higher order Karastuba-Ofman Multiplier (KOM)
Architectures. The higher order KOM multipliers is decomposed into number of lower order multipliers
using radix 2 till basic multiplier block of order 2x2 which is designed by error free Mitchell log multiplier.
The 8x8 REFMLM is tested for Gaussian filter to remove noise in fingerprint image. The Multiplier is
synthesized using Spartan 3 FPGA family device XC3S1500-5fg320. It is observed that the performance
parameters such as area utilization, speed, error and PSNR are better in the case of proposed architecture
compared to existing architectures.
PERFORMANCE FACTORS OF CLOUD COMPUTING DATA CENTERS USING [(M/G/1) : (∞/GDM O...ijgca
The ever-increasing status of the cloud computing h
ypothesis and the budding concept of federated clou
d
computing have enthused research efforts towards in
tellectual cloud service selection aimed at develop
ing
techniques for enabling the cloud users to gain max
imum benefit from cloud computing by selecting
services which provide optimal performance at lowes
t possible cost. Cloud computing is a novel paradig
m
for the provision of computing infrastructure, whic
h aims to shift the location of the computing
infrastructure to the network in order to reduce th
e maintenance costs of hardware and software resour
ces.
Cloud computing systems vitally provide access to l
arge pools of resources. Resources provided by clou
d
computing systems hide a great deal of services fro
m the user through virtualization. In this paper, t
he
cloud data center is modelled as
queuing system with a single task arrivals
and a task request buffer of infinite capacity.
A Defect Prediction Model for Software Product based on ANFISIJSRD
Artificial intelligence techniques are day by day getting involvement in all the classification and prediction based process like environmental monitoring, stock exchange conditions, biomedical diagnosis, software engineering etc. However still there are yet to be simplify the challenges of selecting training criteria for design of artificial intelligence models used for prediction of results. This work focus on the defect prediction mechanism development using software metric data of KC1.We have taken subtractive clustering approach for generation of fuzzy inference system (FIS).The FIS rules are generated at different radius of influence of input attribute vectors and the developed rules are further modified by ANFIS technique to obtain the prediction of number of defects in software project using fuzzy logic system.
A Defect Prediction Model for Software Product based on ANFISIJSRD
Artificial intelligence techniques are day by day getting involvement in all the classification and prediction based process like environmental monitoring, stock exchange conditions, biomedical diagnosis, software engineering etc. However still there are yet to be simplify the challenges of selecting training criteria for design of artificial intelligence models used for prediction of results. This work focus on the defect prediction mechanism development using software metric data of KC1.We have taken subtractive clustering approach for generation of fuzzy inference system (FIS).The FIS rules are generated at different radius of influence of input attribute vectors and the developed rules are further modified by ANFIS technique to obtain the prediction of number of defects in software project using fuzzy logic system.
CONFIGURABLE TASK MAPPING FOR MULTIPLE OBJECTIVES IN MACRO-PROGRAMMING OF WIR...ijassn
Macro-programming is the new generation advanced method of using Wireless Sensor Network (WSNs), where application developers can extract data from sensor nodes through a high level abstraction of the system. Instead of developing the entire application, task graph representation of the WSN model presents simplified approach of data collection. However, mapping of tasks onto sensor nodes highlights several problems in energy consumption and routing delay. In this paper, we present an efficient hybrid approach of task mapping for WSN – Hybrid Genetic Algorithm, considering multiple objectives of optimization – energy consumption, routing delay and soft real time requirement. We also present a method to configure the algorithm as per user's need by changing the heuristics used for optimization. The trade-off analysis between energy consumption and delivery delay was performed and simulation results are presented. The algorithm is applicable during macro-programming enabling developers to choose a better mapping according to their application requirements.
CONFIGURABLE TASK MAPPING FOR MULTIPLE OBJECTIVES IN MACRO-PROGRAMMING OF WIR...ijassn
Macro-programming is the new generation advanced method of using Wireless Sensor Network (WSNs),
where application developers can extract data from sensor nodes through a high level abstraction of the
system. Instead of developing the entire application, task graph representation of the WSN model presents
simplified approach of data collection.
THE REMOVAL OF NUMERICAL DRIFT FROM SCIENTIFIC MODELSIJSEA
Computer programs often behave differently under different compilers or in different computing
environments. Relative debugging is a collection of techniques by which these differences are analysed.
Differences may arise because of different interpretations of errors in the code, because of bugs in the
compilers or because of numerical drift, and all of these were observed in the present study. Numerical
drift arises when small and acceptable differences in values computed by different systems are
integrated, so that the results drift apart. This is well understood and need not degrade the validity of the
program results. Coding errors and compiler bugs may degrade the results and should be removed. This
paper describes a technique for the comparison of two program runs which removes numerical drift and
therefore exposes coding and compiler errors. The procedure is highly automated and requires very little
intervention by the user. The technique is applied to the Weather Research and Forecasting model, the
most widely used weather and climate modelling code.
Student information management system project report ii.pdfKamal Acharya
Our project explains about the student management. This project mainly explains the various actions related to student details. This project shows some ease in adding, editing and deleting the student details. It also provides a less time consuming process for viewing, adding, editing and deleting the marks of the students.
About
Indigenized remote control interface card suitable for MAFI system CCR equipment. Compatible for IDM8000 CCR. Backplane mounted serial and TCP/Ethernet communication module for CCR remote access. IDM 8000 CCR remote control on serial and TCP protocol.
• Remote control: Parallel or serial interface.
• Compatible with MAFI CCR system.
• Compatible with IDM8000 CCR.
• Compatible with Backplane mount serial communication.
• Compatible with commercial and Defence aviation CCR system.
• Remote control system for accessing CCR and allied system over serial or TCP.
• Indigenized local Support/presence in India.
• Easy in configuration using DIP switches.
Technical Specifications
Indigenized remote control interface card suitable for MAFI system CCR equipment. Compatible for IDM8000 CCR. Backplane mounted serial and TCP/Ethernet communication module for CCR remote access. IDM 8000 CCR remote control on serial and TCP protocol.
Key Features
Indigenized remote control interface card suitable for MAFI system CCR equipment. Compatible for IDM8000 CCR. Backplane mounted serial and TCP/Ethernet communication module for CCR remote access. IDM 8000 CCR remote control on serial and TCP protocol.
• Remote control: Parallel or serial interface
• Compatible with MAFI CCR system
• Copatiable with IDM8000 CCR
• Compatible with Backplane mount serial communication.
• Compatible with commercial and Defence aviation CCR system.
• Remote control system for accessing CCR and allied system over serial or TCP.
• Indigenized local Support/presence in India.
Application
• Remote control: Parallel or serial interface.
• Compatible with MAFI CCR system.
• Compatible with IDM8000 CCR.
• Compatible with Backplane mount serial communication.
• Compatible with commercial and Defence aviation CCR system.
• Remote control system for accessing CCR and allied system over serial or TCP.
• Indigenized local Support/presence in India.
• Easy in configuration using DIP switches.
Explore the innovative world of trenchless pipe repair with our comprehensive guide, "The Benefits and Techniques of Trenchless Pipe Repair." This document delves into the modern methods of repairing underground pipes without the need for extensive excavation, highlighting the numerous advantages and the latest techniques used in the industry.
Learn about the cost savings, reduced environmental impact, and minimal disruption associated with trenchless technology. Discover detailed explanations of popular techniques such as pipe bursting, cured-in-place pipe (CIPP) lining, and directional drilling. Understand how these methods can be applied to various types of infrastructure, from residential plumbing to large-scale municipal systems.
Ideal for homeowners, contractors, engineers, and anyone interested in modern plumbing solutions, this guide provides valuable insights into why trenchless pipe repair is becoming the preferred choice for pipe rehabilitation. Stay informed about the latest advancements and best practices in the field.
Saudi Arabia stands as a titan in the global energy landscape, renowned for its abundant oil and gas resources. It's the largest exporter of petroleum and holds some of the world's most significant reserves. Let's delve into the top 10 oil and gas projects shaping Saudi Arabia's energy future in 2024.
CFD Simulation of By-pass Flow in a HRSG module by R&R Consult.pptxR&R Consult
CFD analysis is incredibly effective at solving mysteries and improving the performance of complex systems!
Here's a great example: At a large natural gas-fired power plant, where they use waste heat to generate steam and energy, they were puzzled that their boiler wasn't producing as much steam as expected.
R&R and Tetra Engineering Group Inc. were asked to solve the issue with reduced steam production.
An inspection had shown that a significant amount of hot flue gas was bypassing the boiler tubes, where the heat was supposed to be transferred.
R&R Consult conducted a CFD analysis, which revealed that 6.3% of the flue gas was bypassing the boiler tubes without transferring heat. The analysis also showed that the flue gas was instead being directed along the sides of the boiler and between the modules that were supposed to capture the heat. This was the cause of the reduced performance.
Based on our results, Tetra Engineering installed covering plates to reduce the bypass flow. This improved the boiler's performance and increased electricity production.
It is always satisfying when we can help solve complex challenges like this. Do your systems also need a check-up or optimization? Give us a call!
Work done in cooperation with James Malloy and David Moelling from Tetra Engineering.
More examples of our work https://www.r-r-consult.dk/en/cases-en/
Sachpazis:Terzaghi Bearing Capacity Estimation in simple terms with Calculati...Dr.Costas Sachpazis
Terzaghi's soil bearing capacity theory, developed by Karl Terzaghi, is a fundamental principle in geotechnical engineering used to determine the bearing capacity of shallow foundations. This theory provides a method to calculate the ultimate bearing capacity of soil, which is the maximum load per unit area that the soil can support without undergoing shear failure. The Calculation HTML Code included.
Pile Foundation by Venkatesh Taduvai (Sub Geotechnical Engineering II)-conver...
APPROXIMATE ARITHMETIC CIRCUIT DESIGN FOR ERROR RESILIENT APPLICATIONS
1. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
DOI: 10.5121/vlsic.2022.13601 1
APPROXIMATE ARITHMETIC CIRCUIT DESIGN FOR
ERROR RESILIENT APPLICATIONS
Viraj Joshi and Pravin Mane
Department of EEE, Bits Pilani, K.K. Birla Goa Campus, Goa, India
ABSTRACT
When the application context is ready to accept different levels of exactness in solutions and is supported
by human perception quality, then the term ‘Approximate Computing’ tossed before one decade will
become the first priority . Even though computer hardware and software are working to generate exact
results, approximate results are preferred whenever an error is in predefined bound and adaptive. It will
reduce power demand and critical path delay and improve other circuit metrics. When it comes to
traditional arithmetic circuits, those generating correct results with limitations on performance are rapidly
getting replaced by approximate arithmetic circuits which are the need of the hour, and so on about their
design.
KEYWORDS
Approximate Computing (AC), Error Rate, Acceptance Probability, Mean Error Distance, Error Metrics,
Circuit Metrics
1. INTRODUCTION
The need for approximate computing arose due to changing face of computing workloads as well
as the requirement for fully efficient resources [1],[2]. Approximate computing is supporting
various applications ranging from filtering, and convolutional neural networks to big data
applications because of a few of the reasons. Due to psycho-visual limitations of a human being,
a small error in the visual output is unnoticed by the user. The second one is the resilience to
input noise so that the user can easily accept the noisy output (approximate output). Next is
instead of one golden output for a given input, multiple solutions will go correct. One more is that
during the approximation process if an error occurs at one stage will be getting nullified with
equal and opposite error at other stages. Also, EDU (Error Detection unit) and ECU (Error
Correction Unit) are actively working on error reduction. Other than that as silicon device scaling
reached saturation level, to improve circuit efficiency approximate computing is the best choice
[1].
A growing number of applications are designed to tolerate "noisy" real-world inputs and use
statistical or probabilistic types of computations. These computations are trying to generate the
results best matching the standard and are not traditional ones where an exact answer is required.
These computations provide acceptable quality results and not to be perfect [1], [2]. And, here
approximate computing is preferred. Actually, approximate computing is a concept in which
inaccurate results are generated that are slightly deviating from exact results but not wrong or
random. These results are good enough to produce output not fully correct but partially. This
conflict in the case of AC gave birth to a trade-off between its basic parameters like quality and
performance as shown in fig. 1[3].
2. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
2
Figure 1. A trade-off in Approximate Computing
2. FROM ACCURATE TO APPROXIMATE
For any arithmetic circuit adder, subtractor and multiplier are the basic building blocks. Usually,
an approximation can be added at one of these stages or at all. Fig.2 describes the basic concept
of adding approximation in accurate system
An approximation is usually introduced either in construction or at the logic level or at the
hardware description of arbitrary circuits. In other words, it is stated that approximation is
achieved at four different levels [1]. First is the algorithm level where the actual algorithm is kept
intact either by altering the inputs or the hyperparameters preferred in machine learning and
called meta-learning. Another approach is the application level where algorithms are not intact
and modified to achieve a level of approximation. Loop perforation is the one and in this, the
loop iterations are managed by the user. The third approach is working at an architectural level.
An approximation is achieved by doing modifications in an instruction set with error resilience
bounded for each instruction. The last one works at the circuit level and is related to hardware
circuits. Based on this a lot of work is available w.r.t. adders and multipliers categorized into
deterministic and non-deterministic types.
3. WHY AC?
Figure 2. From Accurate to Approximate
Though approximate computing is tossed and practiced for the last decade, it is still in trend for
researchers due to two key parameters: power and reliability. Fig. 3 describes the concept of
approximate computing. When approximation is introduced at a software level, better energy or
power-efficient system is realized.
3. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
3
Figure 3. The idea of Approximate Computing
Approximate Computing is used in error-tolerant applications ranging from multimedia
processing, machine learning, signal processing, scientific computing, and many more. There are
several reasons that AC is preferred in these applications.
1. Need for energy-efficient systems: Calculations needed for approximate computing are
less, accordingly less hardware requirement and so it is cost-effective.
2. In the case of big data applications, the database is carrying huge data. Out of that AC is
working on the data which is used to generate actual output. This is time-efficient. Also,
real-time computations are done so that resultant data values are used to generate
approximate statistical values.
3. Need for an hour is a performance as compared to accuracy. In those situations, AC is
preferred.
4. To reduce common dependencies like ‘carry chain propagation’ in the case of arithmetic
circuits. It minimizes dependency chains to reduce calculations as well as limit the bit
length of operands.
5. Parallelization: When calculations are lengthy, then the serialization of data or
calculations is a big problem. This concept manages time constraints in a powerful way.
6. Ease of use: By putting in very small efforts approximate computing is showing
maximum adoption for a wide data range.
7. Suitable for inexact model: When models are naturally inexact then approximate
computing perfectly suits this application domain.
8. Limitations on supply voltage usage are strict or voltage scaling techniques reach to
saturation, AC is having full scope in such situations .
9. Deteriorating Moore’s law in practise pushes towards AC.
10. Using Multiple Inexact Program Versions or Inexact or Faulty Hardware will indirectly
supports AC.
4. REQUIREMENTS OF AC
It must be a result-oriented application-specific system.
While applying approximations safety and quality properties must be distinguished. For
a successful approximate computing system, hardware and software co-design is an
important factor.
The measure of the approximate system decides the trade-off between generality and potential
efficiency. Every approximate model must have probabilistic and statistical error analysis
characteristics.
4. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
4
5. BACKGROUND
5.1. Review of Approximate Adders
Vaibhav Gupta et.al. [4] proposed various inexact full adders with the reduction in a number of
transistors in basic full adder cells and utilized them in the construction of multibit approximate
adders used in signal processing applications showing up to 69 % power saving without much
degradation in the quality of output.
Arnab Raha et.al. [5] designed Reconfigurable Adder/subtractor Block (RAB) for video encoding
which controls the degree of approximation (DA) across different videos maintaining PSNR
degradation within 1% to 10% while achieving up to 38% energy saving compared to exact
encoders.
Bharath Srinivas Prabakaran et.al. [6] proposed different FPGA-based multi-bit architectures for
performing approximate addition. The outcome of the paper is in the achievement of gains in the
area of 50%, latency of 38%, and power-delay product of 53% as compared to 16-bit exact
adders. The Register Transfer Level and behavioral codes of these approximate modules are
open-source and can be used for further fueling the research in designing approximate adders.
Sunil Dutt et.al. [7] proposed energy-efficient, high performance Approximate Full Adders
(AFAs), by using the idea of breaking the carry chain subject to low error rate in adder design
and constructed N-bit approximate adder which has shown 46.31% and 28.57% improvement in
power and area respectively with respect to ripple carry adder. In order to improve Error Distance
(ED) and Error Rate (ER), the concept of carry lifetime, error detection, and correction which
provide bit-width aware constant delay is used.
Mahmoud Masadeh et.al. [8] proposed a method to consider the type and position of approximate
full adders used in an approximate multiplier as crucial to designing the target design module.
The algorithm suggested in the article is Pareto-space exploration technique. The significance of
this technique is to mark most optimal approximate designs.
5.2. Review of Approximate Multipliers
Weiqiang Liu et.al. [9] designed two approximate Radix-4 Booth Encoders and two 4:2
compressors which are used to design 4 approximate redundant binary multipliers. The proposed
Radix-4 Booth encoders help in generating approximate partial products and the 4:2 compressors
replicate a reduction in the number of partial products. The work can be extended by designing
new approximate Booth Encoders and 4:2 compressors and applying them to the different
multiplier algorithms.
Muhammad Shafique et.al. [10] presented a survey of different techniques used for designing
power-/energy efficient approximate arithmetic components like adders and multipliers. The
author proposed a method to bridge the gap between the logic layer and architecture layer for
enabling cross-layer approximate computing. The design space exploration technique developed
by them can be used for building multiple approximate components with improved performance
metrics.
K. Bhardwaj et.al. [11] designed a new power-area efficient Approximate Wallace Tree
Multiplier (AWTM) with a bit-width aware approximate multiplication algorithm. The carry-in
prediction method is employed augmented with hardware efficient carry-in precomputation.
5. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
5
Their result has shown up to 99.85% accuracy, up to 24% reduction in latency, up to 41.96%
reduction in power, and up to 34.49% reduction in area. In complex data path designs, errors in
anapproximation of individual circuits accumulate and are not accepted beyond the error limit.
Parag Kulkarni et. al. [12] proposes a design of approximate multiplier showing power savings of
31.78% − 45.4% when compared with an exact multiplier design at the cost of average error of
1.39% -3.32%. Also, it shows improvement in SNR 2X- 8X times than the voltage-over scaling-
based power error trade-off method for the multiplier for the same power savings. Chia-Hao Lin
et. al. [13] proposes a counter design to effectively reduce partial product stages in Wallace tree
multiplier. As compared to the normal Wallace tree multiplier this design is 10.74% power
efficient. Also showing 9.8% low latency w.r.t. normal Wallace tree. But error rates for proposed
designs are in the range of 0.2% to 13.76% in comparison with Wallace tree multiplier. It is
approaching but not exactly 100% with an accuracy of about 99%. With the error detection and
correction unit, it is showing more promising results like for 32-bit length, power savings are
10% over Wallace tree multiplier with 6% faster response time. Burdon of adding this error
correction and detection unit is less than 6% on the overall system.
Amir Momeni et.al.[14] discusses the design of two proposed 4-2 compressors used in the
approximate multiplier and their analysis. Four different schemes have been proposed for these
compressors and it has been utilized in Dadda multiplier. These two compressors and four
multipliers are simulated using HSPICE for different CMOS feature sizes viz. 32, 22, 16nm.
Finally, eight different modes of approximate multipliers are applied to image processing
applications and results are compared for error metrics using visual studio.
Sana Mazahir et.al. [15] developed an adaptive error detection and correction scheme in order to
restrict the size of errors in data path design. The guidelines were developed for designing
flexible approximate adders and multipliers taking into account their integration in data path
designs.
6. RESEARCH GAPS AND SCOPE FOR ADDER AND MULTIPLIER DESIGN
After going through the literature survey for approximate adder and multiplier following gaps are
noticed where there is a scope for research.
Analysis of error metrics linked with circuit metrics. With an upper bound on error parameters,
approximate circuits will be designed with less degradation in area and power consumption and
considerable speed improvement.
Operand size is a major issue in arithmetic circuits from hardware and software points of view.
Error metric and circuit metric analysis are strongly dependent on operand sizes. Their
dependence on operand size can be considerably reduced using approximate circuits.
Approximate tree adders and multipliers can be designed for larger operand lengths by
introducing approximation techniques.
Signed arithmetic circuits have their own challenges in interpreting results because of their
representation in a binary world. After a literature survey, limited work has been reported on
signed approximate arithmetic circuits. There seems a scope for approximation in arithmetic
circuits with signed operands.
When approximation is introduced, by default error will be present. Error detection (EDU) and
correction (ECU) units will be designed for adaptive bound-on on error parameters so that
circuits can be used in a multitude of applications. Depending on circuit metric and error metric
analysis, the mode of operation is decided for a particular application. Although there will be
6. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
6
degradation in terms of accuracy; as per requirement accurate or approximate mode is selected
based on EDU and ECU reports.
Multiplier is the heart of almost all arithmetic and logical circuits and so of many electronic
systems. It is the most complex part for design because of the repeated addition of partial
products with carry. The scope is to introduce a new technique for partial product generation and
addition.
7. PROPOSED DESIGN
This is the proposed idea for adding two numbers without carry propagation. Throughout addition
from LSB to MSB, no carry is considered and propagated.
Fig. 4 indicates the schematic of this technique for the addition of two numbers (of any bit
length). Let us start with the important features of this technique.
No carry is considered while adding numbers, in short, carry independent technique.
As carry is not waving from LSB to MSB, the carry chain propagation problem does not exist.
Non-existence of carry chain propagation problem fastens the arithmetic operations by relaxing
the limit on the power requirement.
Which in turn is the best-suited method for battery-operated appliances.
The method itself is self-corrective that suppress the separate need of error detection and
correction unit.
Attainable accuracy with the help of circuit parameters makes this method flexible for exact as
well as approximate applications.
Figure 4. Schematic of the proposed method for addition
The proposed method is slightly different because of the absence of carry propagation, but it also
gives accurate results too. Though the number of steps required to generate correct results is more
than conventional then also it is energy efficient as well as faster because carry chain propagation
(conventional method) consumes more energy and lowers the speed. Evaluation metrics are
focusing more on it.
7. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
7
7.1. Quality or Results Metrics
In energy-efficient applications, approximate circuits are playing a vital role. Basically, each
approximate circuit is analyzed on the basis of circuit metrics and error metrics. After getting
combined results, application level is decided.
The basic circuit metrics include critical path delay (latency), power, and area [LPA]for regular
circuits, and one additional constraint called ‘error’ is added for approximate circuits [16], [18],
[23].
It is already proved that a large approximation is tolerated in multipliers rather than adders
because complex computations are more prone to errors in addition as compared to the
multiplication process with the same operand length [23].
7.1.1. Error Metrics
Various approaches are used to perform error analysis of approximate circuits using MATLAB or
PYTHON. Few papers suggested data-based error analysis with random inputs to cover
maximum inputs for that particular bit length and analysis [19]. This is the most widely used
method for error analysis as the number of iterations/ trials is increased then analysis results
approach towards accuracy.
Another approach is probabilistic error mode for predicting the possibility of error in the output
[15]. This method works on bit position error probability so that the prediction is expected to be
closer to reality.
The important error parameters considered for comparative analysis are Error Rate (ER), Mean
Error Distance (MED), Average Error, Average Hamming Distance, Mean Relative Error
Distance, Normalized Mean Error Distance, Acceptance Probability, etc.
Error rate and Acceptance Probability play a vital role to decide the level of approximation for a
particular application.
In this article, the first approach is used for error analysis which considers random inputs.
The number of steps is deciding the mode of operation of a circuit. Circuit switches from
approximate to exact mode of operation when a number of steps vary from 5 to 128
progressively.
The number of steps is nothing but the repeated process of logical EX-ORing and ANDing of the
operands till getting the correct result.
Table I describes the behavior of a 128-bit length adder for a various number of steps generating
different error rates and acceptance probabilities.
8. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
8
Table 1. Error Analysis for 128 Bit Length.
Size Trial
Number
of Steps
Error Rate Acceptance
Probability
128 1000
5 88.4 85
128 1000
9 10.6 98.6
128 1000
13 0.9 100
128 1000
17 0 100
128 1000 21 0 100
128 1000 41 0 100
128 1000 53 0 100
128 1000
65
0 100
128 1000
97
0 100
128 1000
129
0 100
Fig.5 explains about the acceptance probability variation w.r.t. bit length for fixed number of
trials, say 1000. As the number of steps varied from 5,9,13,17 to 21, acceptance probability
increased by 15.0%. It is because of an increase in accuracy which leads the results towards
exactness.
Figure 5. Acceptance Probability versus Bit Length
While in the case of response in fig.6, when the number of steps is varied from 9 to 17 through
13, the error rate is reduced by 98.98% because of an increase in accuracy. Improvement in error
rate (accuracy) is at the cost of increased delay and power demand.
9. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
9
Figure 6. Error Rate versus Bit Length
7.1.2. Circuit Metrics
Any design is synthesized for basic circuit metrics like area, power dissipation, and critical path
delay (latency). Along with that compound metrics like power-delay product (PDP), area-delay
product (ADP), and energy-delay product (EDP) can also be studied.
Electronic Design Automation (EDA) tool is used for circuit design and performance evaluation.
For evaluation of resource parameters, different process technologies and software compilers are
available. Few of them are Cadence RTL compiler, predictive technology model (PMT) those are
used for different values of foundries. The Monte Carlo algorithm is used as an assessment
method for PVT analysis of a circuit.
Table 2 is showing circuit metrics analysis using the cadence–genus tool for synthesis.
Resource parameters chosen are area, power, and delay (latency). The circuit is analyzed under
three different operating modes viz. full approximate mode (0.5 * bit-length), half approximate
mode (0.75* bit-length), and accurate mode (1 * bit–length).
While in the case of response in fig.7, when the number of steps is varied from 9 to 17 through
13, the error rate is reduced by 98.98% because of an increase in accuracy. Improvement in error
rate (accuracy) is at the cost of increased delay and power demand.
10. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
10
Table 2. Circuit Metrics Analysis
Bit Length Resource
Parameter
Number of steps
(0.5)*bit
length
(0.75)*bit length Equal to
bit length
8
Area (unit)2
215.1 233.2 263.3
Power (mW) 2.466 6.368 6.356
Delay(nS) 5.429 5.710 5.922
16
Area (unit)2
650.1 732.906 776.340
Power (mW) 4.2 12.079 12.081
Delay(nS) 5.96 6.5 7.004
32
Area (unit)2
2170.3 2558.844 2639.556
Power (mW) 8.808 26.143 26.159
Delay(nS) 7.046 8.147 9.172
64
Area (unit)2
7867.3 9373.536 9522.99
Power (mW) 17.139 51.246 51.299
Delay(nS) 9.202 11.372 13.502
128
Area (unit)2
29264.598 35249.94 36140.166
Power (mW) 104.035 104.566 104.773
Delay(nS) 13.602 17.941 22.231
256
Area (unit)2
111244.392 136385.838 144509.706
Power (mW) 210.029 212.010 212.937
Delay(nS) 22.497 31.201 39.702
Fig.7 explaining about area demand w.r.t. bit length variation. When a circuit operates in an
approximate mode, the area occupied is less because of the less number of hardware involved in
generating results. The circuit occupies 29.90% more area when operated in full exact mode from
full approximate mode in case of 256 bit-length.
Figure 7. Area versus Bit Length
As mapped in fig.8, power demand is plotted versus bit length. When the number of steps is
reduced (approximate mode), power (energy) demand is reduced. This circuit demands 1.38%
more power when switching from approximate mode to exact mode for 256 bit-length. This
means when we are moving from exact to approximate mode circuit becomes more energy
efficient.
11. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
11
Figure 8. Power (Energy) versus Bit Length
Fig. 9 is all about Delay offered w.r.t. bit length. Delay increases by 9.08% when the circuit
switches from approximate to the accurate mode for 256 bit-length.
Figure 9. Delay Vs. Bit Length
Overall it is concluded that the circuit becomes more energy-efficient as well as faster at the cost
of accuracy when switching from an accurate to an approximate mode of operation.
After complete analysis, it is concluded that this kind of approximate adder is more energy
efficient and faster in its approximate mode of operation with a bounded compromise in
accuracy.
8. APPLICATION DOMAIN OF AC
AC itself is covering a broad range of applications from data processing, filtering, and neural
networks to big data applications and image processing. Each application domain is having a
specific reason for incorporating approximation in either input or in methodology or in logic.
Here are a few examples of these application domains which work on the principle of error
resilience [16],[21].
Image Processing: In the case of image or signal processing applications, the inputs are accepted
from discrete sensors or cameras that are already noisy in nature as compared to the analog sensor
outputs. Still users unnoticed noise in input as well as output because of perceptual limitations.
12. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
12
AC makes use of this particular property for generating imperfect outputs with good acceptance
probability [16], [17].
Computer Graphics: In case of Computer Graphics applications, the most important parameters
are speed (frame rate) as well as computational resources especially needed for gaming. If these
requirements are not meeting, then compromise in the accuracy of output for available frame rate
so that erroneous (approximate) output can be accepted [16], [17].
Cloud Computing or Big Data Applications: These are falling under machine learning area. This
category of applications is handling a huge data that is collected as well as processed from/to
various resources or users and because of it, there is a huge possibility of having errors in the
output for which the system and users are trained. This quality strongly supports approximate
computing [16], [17].
Computer Vision: Mostly gaming and dynamic image processing applications are related to
computer vision which covers motion tracking, pose estimation, and object detection by writing
algorithms for it. Not all algorithms are optimal to achieve the required accuracy and generate
exact digital images. Users are also not able to read exact images allowing them to add
approximation in the output [17], [18].
Biometric Security: Biometric applications are a deadly combination of image processing,
computer vision, and machine learning. These are data-rich applications including face
recognition, fingerprinting, iris scanning, biometric signatures, etc. Biometric details are
dedicated and unique w.r.t. individual. If the details vary minutely (or in a controlled manner)
then also it will maintain its uniqueness w.r.t. that individual and where exactly it shows the
scope for approximation [16],[17].
9. APPLICATIONS
Although AC is very promising still it will not be the first choice every time because of a few
limitations like a proper selection of mode as well as approximation technique, also approximate
code or data portion because all these parameters are application-oriented and cause quality loss
with the same choice every time [21].
Usually, error rates less than 10% and PSNR values greater than 30dB are acceptable respectively
in error-resilient and image processing applications respectively [21], [22]. When image
processing applications are discussed, approximate multipliers with a lower error rate and higher
acceptance probability outperform over the same circuit with a higher error rate for applications
viz. image sharpening, image smoothening, brightness and contrast control, image compression
[23]. But same thumb rule is not applicable for approximate adders as for complex operations,
these adders do not give expected results.
The operand length and an approximation technique are the most important factors on which
evaluation reports are dependent. After proper evaluation of the arithmetic circuit, its application
domain is decided.
In case of image processing applications, the combination of approximate adder and multiplier is
preferred to reconstruct a good quality processed image. Using either approximate adder or
multiplier is not a good option to get the same result.
Applications like image sharpening are tolerating more approximation in addition process as
compared to JPEG compression and face detection and alignment.
13. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
13
Approximate adders are mainly designed for reducing critical path delay by lowering error rates
and improving circuit performance. But in the case of approximate adders as the carries are
ignored or reduced throughout addition will be prone to generate single-sided errors that will
become prominent in repeated or iterative addition. When approximate full adders are used at
LSBs produce high ER but low power dissipation. So while selecting an approximate adder for
any application the best trade-off should be achieved between error metrics and circuit metrics.
Unsigned multipliers show improvement in circuit area by truncating part of PPs or part of LSBs
in input operands at the cost of degraded accuracy. The tradeoff is between area and accuracy [4].
Few are showing a trade-off between energy and accuracy [4]. For signed multipliers, the booth
multiplier shows promising results due to good error compensation as compared to TBMs. Other
than that MRED[Mean Relative Error Distance] is a measure of accuracy in the case of
multiplication applications like face detection.
10. CASE STUDY
To illustrate the applicability of the proposed technique to real-time use, it is applied to
benchmark images as shown in fig.10. It is tested by implementing image processing applications
like brightness and contrast control. Further same adder technique is used in approximate
multipliers for partial product addition and showcasing the corresponding effect on chosen
standard images.
As mentioned earlier, the application-level approximation technique is used where the
approximation is introduced by managing the loop iterations.
We can apply the proposed adder in any of the machine learning algorithms where addition and
subtraction are heavily performed. Output generated by an accurate adder is considered a golden
reference for application.
Image b is the original image with brightness and contrast setting limits where it is showing the
minimum and maximum range for brightness and contrast variation.
In this case study, the value of image parameters says brightness and contrast changes
automatically after the application of proposed adder.
Image c is the resultant image generated by an accurate adder (when the number of
steps=1*bitlength). It is exactly matching with the original one as brightness and contrast are
maintained the same as the original image.
14. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
14
Figure 10. Original image (a) and output images using brightness and contrast control with accurate and
approximate adder
Image d (when the number of steps=0.75*bit-length),e (when the number of
steps=0.50*bitlength),f(when the number of steps=0.35*bit-length)are the outputs generated by
an approximate adder when approximation increased progressively through the loop perforation
technique. Brightness and contrast value changes automatically which deviates the output image
from the original one. As per the required quality of the output, the level of approximation is
used. Along with this practical application approach, through circuit metric analysis it is already
proved that how proposed adder significantly reduced hardware resource consumption.
11. CONCLUSION, CHALLENGES, AND PROSPECTS
This article is an overall review of approximate circuits also proposing an energy-efficient
approximate adder. It covers basics, related work, approaches for approximation and challenges,
evaluation, applications, and future scope.
The purpose of designing Approximate Arithmetic adders is basically for improvement in power
efficiency as well as critical path delay. But sometimes due to approximation techniques used the
errors become prominent in complex applications, even though the error rates are low. Therefore,
adders are showing less scope for approximation than multipliers. Approximate adders are always
showing conflicts between circuit and error metrics. This conflict is cross verified with the help
of a proposed technique.
On the other hand, approximate multipliers with lower error rates are showing promising results
in complex applications. While selecting any multiplier tradeoff between accuracy and circuit
metrics viz. area and energy is taken to be into an account.
15. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
15
In case of hardware platform, approximate computing has been proven by fabricating a silicon
chip with approximate processors for recognition and data mining applications.
Nowadays approximate computing is quite settled and the major goal in it is automation. Due to
this researchers are designing approximate computing systems for partial or total cognition to
decide the implication of approximation. A literature survey for image processing applications
shows that almost all applications are implemented for static systems and not for dynamic and
there it seems a big scope for research work. Because run time error quality monitoring system
(dynamic) is more complex, costly, and maintenance prone than compared to a static system. But
these systems are more beneficial for working.
According to ITRS roadmap, to chase Moore’s law, the semiconductor industry is progressing
with 10nm microchips for production and looking for 7nm in the upcoming 5 years. But it
increases the heavy Burdon on design constraints like supply voltage, frequency, yield along with
operational faults. These hardware faults might propagate in software that can deflect the actual
output. While applying approximate computing, we have to take care of these issues. Otherwise,
approximate computing is really a good hope for green energy.
REFERENCES
[1] M. Ammar Ben Khadra, “An introduction to approximate computing”, 10 Dec. 2017
[2] Source: Anand Raghunathan, Writer: Emil Venere, ‘Approximate Computing’ improves
efficiency, saves energy, December 17, 2013
[3] Approximate Computing: Wikipedia.
[4] V. Gupta, D. Mohapatra, A. Raghunathan, and K. Roy, “Low-power digital signal processing using
approximate adders,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and
Systems, vol. 32, no. 1, pp. 124–137, Jan 2013.
[5] A. Raha, H. Jayakumar, and V. Raghunathan, “Input-based dynamic reconfiguration of approximate
arithmetic units for video encoding,” IEEE Transactions on Very Large Scale Integration (VLSI)
Systems, vol. 24, no. 3, pp. 846–857, March 2016.
[6] B. S. Prabakaran, S. Rehman, M. A. Hanif, S. Ullah, G. Mazaheri, A. Kumar, and M. Shafique,
“Demas: An efficient design methodology for building approximate adders for fpga-based systems,”
in 2018 Design, Automation Test in Europe Conference Exhibition (DATE), March 2018, pp. 917–
920.
[7] S. Dutt, S. Nandi, and G. Trivedi, “Accuracy enhancement of equal segment based approximate
adders,” IET Computers Digital Techniques, vol. 12, no. 5, pp. 206–215, 2018.M. Young, The
Technical Writer’s Handbook. Mill Valley, CA: University Science, 1989.
[8] M. Masadeh, O. Hasan, and S. Tahar, “Comparative study of approximate multipliers,” CoRR, vol.
abs/1803.06587, 2018. [Online]. Available: http://arxiv.org/abs/1803.06587
[9] W. Liu, T. Cao, P. Yin, Y. Zhu, C. Wang, E. E. Swartzlander, and F. Lombardi, “Design and analysis
of approximate redundant binary multipliers,” IEEE Transactions on Computers, vol. 68, no. 6, pp.
804–819, June 2019.
[10] M. Shafique, R. Hafiz, S. Rehman, W. El-Harouni, and J. Henkel, “Invited: Cross-layer approximate
computing: From logic to architectures,” in 2016 53nd ACM/EDAC/IEEE Design Automation
Conference (DAC), June 2016, pp. 1–6.
[11] K. Bhardwaj, P. S. Mane, and J. Henkel, “Power- and area-efficient approximate wallace tree
multiplier for error-resilient systems,” in Fifteenth International Symposium on Quality Electronic
Design, March 2014, pp. 263–269.
[12] Kulkarni, P., Gupta, P., & Ercegovac, M. (2011). Trading Accuracy for Power with an
Underdesigned Multiplier Architecture. In 2011 24th Internatioal Conference on VLSI Design
[13] Chia-Hou Lin, Ing-Chao Lin, “High Accuracy Approximate Multiplier with Error Correction", in
IEEE 31st International Conference on Computer Design (ICCD), Oct. 2013
[14] Amir Momeni, Jie Han, Paolo Montuschi, Fabrizio Lombardi, “Design and Analysis of Approximate
Compressors for Multiplication”, IEEE Transactions on Computers, Feb., 2014, pp. 984 – 994.
16. International Journal of VLSI design & Communication Systems (VLSICS)
Vol 13, No 1/2/3/4/5/6, December 2022
16
[15] Sana Mazahir, Osman Hasan, Rehan Hafiz, Muhammad Shafique, Jorg Henkel, “ProbabilisticError
Modeling for Approximate Adders”, IEEE Transactions on Computers, Vol. 66, No. 3, March 2017,
pp. 515-530.
[16] S. Reda, M. Shafique, “Approximate Circuits-Methodologies and CAD”, Springer Nature
Switzerland, Springer, 2019.
[17] Naman Maheshwari , Approximate Computing and its Applications, Aug. 6, 2016
[18] Yi Wu, You Li, Xiangxuan Ge, Yuan Gao, Weikang Qian, “An Efficient Method for Calculating the
Error Statistics of Block-Based Approximate Adders”, Ieee Transactions On Computers, Vol. 68, No.
1, January 2019, PP. 21-38.
[19] Omid Akbari, Mehdi Kamal, Ali Afzali-Kusha, Massoud Pedram, “RAP-CLA: A Reconfigurable
Approximate Carry Look-Ahead Adder”, Ieee Transactions On Circuits And Systems—Ii: Express
Briefs, Vol. 65, No. 8, August 2018, pp. 1089-1093.
[20] Sparsh Mittal, “A Survey Of Techniques for Approximate Computing”, Article in ACM Computing
Surveys · March 2016, pp1-34.
[21] Abbas Rahimi, Amirali Ghofrani, Kwang-Ting Cheng, Luca Benini, and Rajesh K Gupta,
“Approximate associative memristive memory for energy-efficient GPUs”, In Design, Automation &
Test in Europe,2015, pp. 1497–1502.
[22] Honglan Jiang, Cong Liu, Liebo Liu, Faling to predictbrizio Lombardi, Jie, han, “A Review,
Classification and Comparative Evaluation of Approximate Arithmetic Circuits”, ACM Journal on
Emerging Technologies in Computing Systems, Vol. 13, No. 4, Article 60, Pub. date: July 2017, pp.
60:1-60:37.
[23] Honglan Jiang, Francisco Javier Hernandez Santiago, Hai Mo, Leibo Liu, Jie Han, “Approximate
Arithmetic Circuits: A Survey, Characterization, and Recent Applications”, Proceedings of the IEEE,
Volume: 108,Issue: 12, Dec. 2020, pp. 2108 – 2135.
AUTHORS
Viraj Joshi received her B.E. degree from Shivaji University and M.E. degree from
Bharti Vidyapeeth. She is full time research schoar at Bits Pilani, K.K. Birla, Goa
campus in EEE department. Her research interest includes Approximate arithmetic
circuits.
Dr. Pravin Mane received his B.E. degree from RIT, Sakhrale , M. Tech. from IIT,
Roorkee and Ph.D. from Bits Pilani, K.K. Birla, Goa Campus. Currently he is working
as assistant professor in Bit Pilani, K.K. Birla, Goa campus. His area of interest is VLSI
design.