Analysis and optimization of in-situ error detection techniques in ultra-low-voltage pipeline

2014 ◽  
Author(s):  
Seongjong Kim ◽  
Mingoo Seok
Keyword(s):  
Error Detection ◽  
Low Voltage ◽  
Detection Techniques

scholarly journals In Situ Error Detection Techniques in Ultralow Voltage Pipelines: Analysis and Optimizations

2017 ◽  
Vol 25 (3) ◽  
pp. 1032-1043 ◽  
Author(s):  
Wei Jin ◽  
Seongjong Kim ◽  
Weifeng He ◽  
Zhigang Mao ◽  
Mingoo Seok
Keyword(s):  
Error Detection ◽  
Detection Techniques

scholarly journals Low-Voltage Energy Efficient Neural Inference by Leveraging Fault Detection Techniques

2021 ◽  
Author(s):  
Mehdi Safarpour ◽  
Tommy Z. Deng ◽  
John Massingham ◽  
Lei Xun ◽  
Mohammad Sabokrou ◽  
...  
Keyword(s):  
Fault Detection ◽  
Error Detection ◽  
Energy Savings ◽  
Low Voltage ◽  
Building Blocks ◽  
Feedback Mechanism ◽  
Total System ◽  
Data Set ◽  
Detection Techniques ◽  
Energy Efficiency Improvement

This paper presents simple techniques to significantly reduced energy consumption of DNNs: Operating at reduced voltages offers substantial energy efficiency improvement but at the expense of increasing the probability of computational errors due to hardware faults. In this context, we targeted Deep Neural Networks (DNN) as emerging energy hungry building blocks in embedded applications. Without an error feedback mechanism, blind voltage down-scaling will result in degraded accuracy or total system failure. To enable safe voltage down-scaling, in this paper two solutions based on Self-Supervised Learning (SSL) and Algorithm Based Fault Tolerance (ABFT) were developed. A DNN model trained on MNIST data-set was deployed on a Field Programmable Gate Array (FPGA) that operated at reduced voltages and employed the proposed schemes. The SSL approach provides extremely low-overhead (≈0.2%) fault detection at the cost of lower error coverage and extra training, while ABFT incurs less than 8%overheads at run-time with close to 100% error detection rate. By using the solutions, substantial energy savings, i.e., up to 40.3%,without compromising the accuracy of the model was achieved

scholarly journals Low-Voltage Energy Efficient Neural Inference by Leveraging Fault Detection Techniques

2021 ◽  
Author(s):  
Mehdi Safarpour ◽  
Tommy Z. Deng ◽  
John Massingham ◽  
Lei Xun ◽  
Mohammad Sabokrou ◽  
...  
Keyword(s):  
Fault Detection ◽  
Error Detection ◽  
Energy Savings ◽  
Low Voltage ◽  
Building Blocks ◽  
Feedback Mechanism ◽  
Total System ◽  
Data Set ◽  
Detection Techniques ◽  
Energy Efficiency Improvement

Operating at reduced voltages offers substantial energy efficiency improvement but at the expense of increasing the probability of computational errors due to hardware faults. In this context, we targeted Deep Neural Networks (DNN) as emerging energy hungry building blocks in embedded applications. Without an error feedback mechanism, blind voltage down-scaling will result in degraded accuracy or total system failure. To enable safe voltage down-scaling, in this paper two solutions based on Self-Supervised Learning (SSL) and Algorithm Based Fault Tolerance (ABFT) were developed. A DNN model trained on MNIST data-set was deployed on a Field Programmable Gate Array (FPGA) that operated at reduced voltages and employed the proposed schemes. The SSL approach provides extremely low-overhead (≈0.2%) fault detection at the cost of lower error coverage and extra training, while ABFT incurs less than 8%overheads at run-time with close to 100% error detection rate. By using the solutions, substantial energy savings, i.e., up to 40.3%,without compromising the accuracy of the model was achieved

scholarly journals Low-Voltage Energy Efficient Neural Inference by Leveraging Fault Detection Techniques

2021 ◽  
Author(s):  
Mehdi Safarpour ◽  
Tommy Z. Deng ◽  
John Massingham ◽  
Lei Xun ◽  
Mohammad Sabokrou ◽  
...  
Keyword(s):  
Fault Detection ◽  
Error Detection ◽  
Energy Savings ◽  
Low Voltage ◽  
Building Blocks ◽  
Feedback Mechanism ◽  
Total System ◽  
Data Set ◽  
Detection Techniques ◽  
Energy Efficiency Improvement

Operating at reduced voltages offers substantial energy efficiency improvement but at the expense of increasing the probability of computational errors due to hardware faults. In this context, we targeted Deep Neural Networks (DNN) as emerging energy hungry building blocks in embedded applications. Without an error feedback mechanism, blind voltage down-scaling will result in degraded accuracy or total system failure. To enable safe voltage down-scaling, in this paper two solutions based on Self-Supervised Learning (SSL) and Algorithm Based Fault Tolerance (ABFT) were developed. A DNN model trained on MNIST data-set was deployed on a Field Programmable Gate Array (FPGA) that operated at reduced voltages and employed the proposed schemes. The SSL approach provides extremely low-overhead (≈0.2%) fault detection at the cost of lower error coverage and extra training, while ABFT incurs less than 8%overheads at run-time with close to 100% error detection rate. By using the solutions, substantial energy savings, i.e., up to 40.3%,without compromising the accuracy of the model was achieved

Real-time error detection techniques based on FPGA

2013 ◽  
Vol 33 (5) ◽  
pp. 1459-1462
Author(s):  
Xiaoming JU ◽  
Jiehao ZHANG ◽  
Yizhong ZHANG
Keyword(s):  
Real Time ◽  
Error Detection ◽  
Time Error ◽  
Detection Techniques

Adaptive in-situ forecasting for demand-side management in low voltage power grids

2021 ◽  
Author(s):  
Christian Backe ◽  
Miguel Bande ◽  
Stefan Werner ◽  
Christian Wiezorek
Keyword(s):  
Low Voltage ◽  
Demand Side Management ◽  
Power Grids ◽  
Demand Side

scholarly journals Performance of electrokinetic treatment of fine-grained problematic soils

2021 ◽  
Vol 9 (4B) ◽  
Author(s):  
Abiola Ayopo Abiodun ◽  
◽  
Zalihe Nalbantoglu ◽  
Keyword(s):  
Low Voltage ◽  
Soil Depth ◽  
Experimental Tests ◽  
Interactive Effects ◽  
Engineering Properties ◽  
Laboratory Model ◽  
Fine Grained ◽  
Electrode Length ◽  
Problematic Soils

Electrokinetic (EK) treatment is an innovative, cost-effective in situ ground modification technology. The EK treatment uses a combination of low-voltage direct-current, electrodes, and ionic solutions across problematic soil to improve the ground conditions. This study aims to model the effect of changing electrode length (le) on the performance of the EK treatment on the engineering properties of fine-grained problematic soils. The consideration of the changing electrode lengths (le), varying soil depths (ds), and lengthwise anode to cathode distances (dA↔E), in the soil block samples, is in the form of the laboratory model test tank. The significant performance of the experimental tests was with changing electrode lengths of 0.25le (7.5 cm), 0.50le (15.0 cm), 0.75le (22.5 cm), and 1.0le (30.0 cm). The study analyzed the test data obtained from the Atterberg limit and one-dimensional swelling tests at different extraction points of the EK treated soils in the test tanks. Furthermore, the study carefully analyzed the effect of changing electrode length (le) on the performance of the EK treatment. The results of the Design of Experiment (DOE) model analysis revealed that the effect of changing electrode length (le) on the plasticity index (PI), and swelling potential (SP) of the EK treated soils, was significant. For a specific soil depth (ds), the electrode lengths (le) of 0.50le and 0.75le were significantly effective in reducing the PI, and the SP of the EK treated soils. Unlike other studies in the literature, the use of DOE analysis in the present study enabled the detection of the significant input factors and their interactive effects on the PI and the SP, thus, enabling the practicing engineers to navigate accurate design models for large in situ applications.

Sign in / Sign up

Export Citation Format

Share Document