| 1. |
- Taheri, M., et al.
(författare)
-
DeepAxe : A Framework for Exploration of Approximation and Reliability Trade-offs in DNN Accelerators
- 2023
-
Ingår i: Proceedings - International Symposium on Quality Electronic Design, ISQED. - : IEEE Computer Society. - 9798350334753
-
Konferensbidrag (refereegranskat)abstract
- While the role of Deep Neural Networks (DNNs) in a wide range of safety-critical applications is expanding, emerging DNNs experience massive growth in terms of computation power. It raises the necessity of improving the reliability of DNN accelerators yet reducing the computational burden on the hardware platforms, i.e. reducing the energy consumption and execution time as well as increasing the efficiency of DNN accelerators. Therefore, the trade-off between hardware performance, i.e. area, power and delay, and the reliability of the DNN accelerator implementation becomes critical and requires tools for analysis.In this paper, we propose a framework DeepAxe for design space exploration for FPGA-based implementation of DNNs by considering the trilateral impact of applying functional approximation on accuracy, reliability and hardware performance. The framework enables selective approximation of reliability-critical DNNs, providing a set of Pareto-optimal DNN implementation design space points for the target resource utilization requirements. The design flow starts with a pre-trained network in Keras, uses an innovative high-level synthesis environment DeepHLS and results in a set of Pareto-optimal design space points as a guide for the designer. The framework is demonstrated on a case study of custom and state-of-the-art DNNs and datasets.
|
|
| 2. |
- Zhou, L. Y., et al.
(författare)
-
A transition of atmospheric emissions of particles and gases from on-road heavy-duty trucks
- 2020
-
Ingår i: Atmospheric Chemistry and Physics. - : Copernicus GmbH. - 1680-7316 .- 1680-7324. ; 20:3, s. 1701-1722
-
Tidskriftsartikel (refereegranskat)abstract
- The transition, in extent and characteristics, of atmospheric emissions caused by the modernization of the heavy-duty on-road fleet was studied utilizing roadside measurements. Emissions of particle number (PN), particle mass (PM), black carbon (BC), nitrogen oxides (NOx), carbon monoxide (CO), hydrocarbon (HC), particle size distributions, and particle volatility were measured from 556 individual heavy-duty trucks (HDTs). Substantial reductions in PM, BC, NOx, CO, and to a lesser extent PN were observed from Euro III to Euro VI HDTs by 99 %, 98 %, 93 %, and 57% for the average emission factors of PM, BC, NOx, and CO, respectively. Despite significant total reductions in NOx emissions, the fraction of NO2 in the NOx emissions increased continuously from Euro IV to Euro VI HDTs. Larger data scattering was evident for PN emissions in comparison to solid particle number (SPN) for Euro VI HDTs, indicating a highly variable fraction of volatile particle components. Particle size distributions of Euro III to enhanced environmentally friendly vehicle (EEV) HDTs were bimodal, whereas those of Euro VI HDTs were nucleation mode dominated. High emitters disproportionately contributed to a large fraction of the total emissions with the highest-emitting 10% of HDTs in each pollutant category being responsible for 65% of total PM, 70% of total PN, and 44% of total NOx emissions. Euro VI HDTs, which accounted for 53% of total kilometres driven by Swedish HDTs, were estimated to only contribute to 2 %, 6 %, 12 %, and 47% of PM, BC, NOx, and PN emissions, respectively. A shift to a fleet dominated by Euro VI HDTs would promote a transition of atmospheric emissions towards low PM, BC, NOx, and CO levels. Nonetheless, reducing PN, SPN, and NO2 emissions from Euro VI HDTs is still important to improve air quality in urban environments.
|
|
| 3. |
- Fredriksson Möller, Björn, et al.
(författare)
-
Azep gas turbine combined cycle power plants thermo-economic analysis
- 2005
-
Ingår i: Proceedings of Ecos 2005, Vols 1-3 - Shaping our future energy systems. ; , s. 819-826
-
Konferensbidrag (refereegranskat)abstract
- Conventional power plants based on fossil fuel without CO2 capture produce flue gas streams with concentrations Of CO2 between 3% and 15%, contributing to the threat of increasing global warming. Existing capture technologies such as post-combustion flue gas treatment using chemical absorption or pre-combustion carbon removal suffer from significant efficiency penalties as well as major increase in investment costs. Alternatively, combustion in O-2/CO2 atmospheres also requires expensive and energy-consuming oxygen supplies. A less energy intensive concept for oxygen production is a Mixed Conducting Membrane (MCM) reactor which produces pure oxygen from compressed air. The MCM reactor is best integrated into a conventional gas turbine combined cycle, called Advanced Zero Emissions Plant (AZEP), to provide an efficient and cost-effective power plant altogether. In this paper the economic performance of four different combined cycle alternatives in two different gas turbine sizes are evaluated; a 50 MWe size based on the Siemens SGT800 gas turbine and a 400 MWe size based on the Siemens SGT5-4000F gas turbine. ne evaluated combined cycles are one conventional combined cycle, one combined cycle with post-combustion CO2 capture and two optimised AZEP cases from a previous thermodynamic study. One AZEP alternative provides 100% CO2 capture and is thus a true zero emissions alternative, whereas the second alternative uses a sequential combustion system which enables 85% of the CO2 to be captured, making a comparison with traditional post-combustion treatment easier. The results show that the AZEP concept presents a more competitive system in terms of efficiency and economy compared to traditional capture systems.
|
|
| 4. |
- Riazati, Mohammad, et al.
(författare)
-
Adjustable self-healing methodology for accelerated functions in heterogeneous systems
- 2020
-
Ingår i: Proceedings - Euromicro Conference on Digital System Design, DSD 2020. - : Institute of Electrical and Electronics Engineers Inc.. - 9781728195353 ; , s. 638-645
-
Konferensbidrag (refereegranskat)abstract
- Self-healing is a promising approach for designing reliable digital systems. It refers to the ability of a system to detect faults and automatically fixing them to avoid total failure. With the development of digital systems, heterogeneous systems, in which some parts of the system are executed on the programmable logic, and some other parts run on the processing elements (CPU), are becoming more prevalent. In this work, we propose an adjustable self-healing method that is applicable to heterogeneous systems with accelerated functions and enables the designers to add the self-healing feature to the design. In this method, by manipulating the software codes that are being executed on the processing element, we add the ability to verify the accelerated functions on the programmable logic and heal the possible failures to the system. This is done not only in a straightforward manner but also without being forced to choose a specific reliability-overhead point. The designer will have the option to select the optimum configuration for a desired reliability level. Experimental results on a large design including several accelerated functions are provided and show 42% improvement of reliability by having 27% overhead, as an example of the reliability-overhead point.
|
|
