scholarly journals Strategies to Improve the Performance and Energy Efficiency of Stencil Computations for NVIDIA GPUs

2018 ◽  
Author(s):  
Pablo José Pavan ◽  
Matheus da Silva Serpa ◽  
Víctor Martínez ◽  
Edson Luiz Padoin ◽  
Jairo Panetta ◽  
...  
Keyword(s):  
Energy Efficiency ◽  
Energy Efficient ◽  
Large Scale ◽  
Data Locality ◽  
Large Scale Systems ◽  
Systems Research ◽  
Gpu Algorithms ◽  
Read Only Memory ◽  
And Performance ◽  
Gpu Architecture

Energy and performance of parallel systems are an increasing concern for new large-scale systems. Research has been developed in response to this challenge aiming the manufacture of more energy efficient systems. In this context, we improved the performance and achieved energy efficiency by the development of three different strategies which use the GPU memory subsystem (global-, shared-, and read-only- memory). We also develop two optimizations to use data locality and use of registers of GPU architecture. Our developed optimizations were applied to GPU algorithms for stencil applications achieve a performance improvement of up to 201:5% in K80 and 264:6% in P 100 when used shared memory and read-only cache respectively over the naive version. The computational results have shown that the combination of use read-only memory, the Z-axis internalization of stencil application and reuse of specific architecture registers allow increasing the energy efficiency of up to 255:6% in K80 and 314:8% in P 100.

scholarly journals Simba

2021 ◽  
Vol 64 (6) ◽  
pp. 107-116
Author(s):  
Yakun Sophia Shao ◽  
Jason Cemons ◽  
Rangharajan Venkatesan ◽  
Brian Zimmer ◽  
Matthew Fojtik ◽  
...  
Keyword(s):  
Deep Learning ◽  
Large Scale ◽  
Data Locality ◽  
Coarse Grained ◽  
Batch Size ◽  
Peak Performance ◽  
Large Scale Systems ◽  
High Area ◽  
On Chip ◽  
And Storage

Package-level integration using multi-chip-modules (MCMs) is a promising approach for building large-scale systems. Compared to a large monolithic die, an MCM combines many smaller chiplets into a larger system, substantially reducing fabrication and design costs. Current MCMs typically only contain a handful of coarse-grained large chiplets due to the high area, performance, and energy overheads associated with inter-chiplet communication. This work investigates and quantifies the costs and benefits of using MCMs with finegrained chiplets for deep learning inference, an application domain with large compute and on-chip storage requirements. To evaluate the approach, we architected, implemented, fabricated, and tested Simba, a 36-chiplet prototype MCM system for deep-learning inference. Each chiplet achieves 4 TOPS peak performance, and the 36-chiplet MCM package achieves up to 128 TOPS and up to 6.1 TOPS/W. The MCM is configurable to support a flexible mapping of DNN layers to the distributed compute and storage units. To mitigate inter-chiplet communication overheads, we introduce three tiling optimizations that improve data locality. These optimizations achieve up to 16% speedup compared to the baseline layer mapping. Our evaluation shows that Simba can process 1988 images/s running ResNet-50 with a batch size of one, delivering an inference latency of 0.50 ms.

scholarly journals Building the Table of Energy and Power Leverages for Energy Efficient Large Scale Systems

2018 ◽  
Author(s):  
Issam Rais ◽  
Mathilde Boutigny ◽  
Laurent Lefevre ◽  
Anne-Cecile Orgerie ◽  
Anne Benoit
Keyword(s):  
Energy Efficient ◽  
Large Scale ◽  
Large Scale Systems

scholarly journals BUILDING ENERGY AND ARCHITECTURAL FORM RELATIONSHIPS / PASTATO ENERGIJOS IR ARCHITEKTŪRINĖS FORMOS RYŠIAI

2011 ◽  
Vol 3 (3) ◽  
pp. 67-71 ◽  
Author(s):  
Roberts Riekstiņš
Keyword(s):  
Energy Efficiency ◽  
Environmental Sustainability ◽  
Energy Efficient ◽  
Large Scale ◽  
Building Energy ◽  
Substantial Part ◽  
Building Energy Efficiency ◽  
Total Energy Consumption ◽  
Architectural Form ◽  
Energy Efficient Building

Energy efficiency of buildings, of course, is now a major issue in the construction industry. It is being widely examined both among construction professionals and amateurs. There is no doubt that energy efficiency as a key factor in ensuring environmental sustainability will become the main driving force of the construction in the future. Buildings have to become more energy-efficient. This opinion is supported by the existing energy-use balance in Europe, indicating that the housing sector spends almost half of total energy consumption and building sector forms more than a third of total CO2 emissions (Bradley 2010). While discussing the subject of building energy efficiency, mostly different technical characteristics of buildings and engineering solutions are talked over. However, it has been relatively little examined how energy-efficient design affects the building’s architecturally-aesthetic side, styles of expression and trends in the architect’s profession. We learn that the essence for an energy-efficient building lies in smart modesty (Bokalders, Block 2010) and the rational utilization of materials (aim high – go low). And still – can energy efficient building be expressive, extravagant, and perhaps – even ambitious? There are many ideas implemented in projects which show that energy efficiency is not an obstacle to large scale architectural ideas. However, in order to combine architectural and artistic ambitions with the principles of sustainability, architects should search for an entirely new approach to architectural expression based on a detailed assessment of solutions applied from environmental point of view. It requires a complex understanding of building shape, applied technologies, energetic benefits and cost parameters. This article identifies the realised and experimental projects of the world and presents an analysis of classification of buildings according to typology. This publication gives general impression of the amplitude and topicality of the study issue, as well as the diversity applied to the architectural techniques. The article concludes that even creating a building’s shape in a smart way makes it possible to use substantial part of the renewable energy offered by nature.

Energy-efficient distributed control of large-scale systems: A switched system approach

2015 ◽  
Vol 26 (14) ◽  
pp. 3101-3117 ◽  
Author(s):  
Dan Zhang ◽  
Peng Shi ◽  
Qing-Guo Wang
Keyword(s):  
Distributed Control ◽  
Energy Efficient ◽  
System Approach ◽  
Large Scale ◽  
Switched System ◽  
Large Scale Systems

ERIDIS: ENERGY-EFFICIENT RESERVATION INFRASTRUCTURE FOR LARGE-SCALE DISTRIBUTED SYSTEMS

2011 ◽  
Vol 21 (02) ◽  
pp. 133-154 ◽  
Author(s):  
ANNE-CECILE ORGERIE ◽  
LAURENT LEFEVRE
Keyword(s):  
Energy Efficiency ◽  
Cloud Computing ◽  
Distributed Systems ◽  
Energy Efficient ◽  
Large Scale ◽  
Generic Framework ◽  
Peer To Peer Systems ◽  
Dedicated Networks ◽  
Electrical Consumption ◽  
Hardware Designs

At the age of petascale machines, cloud computing and peer-to-peer systems, large-scale distributed systems need an ever-increasing amount of energy. These systems urgently require effective and scalable solutions to manage and limit their electrical consumption. As of now, most efforts are focused on energy-efficient hardware designs. Thus, the challenge is to coordinate all these low-level improvements at the middleware level to improve the energy efficiency of the overall systems. Resource-management solutions can indeed benefit from a broader view to pool the resources and to share them according to the needs of each user. In this paper, we propose ERIDIS, an Energy-efficient Reservation Infrastructure for large-scale DIstributed Systems. It provides a unified and generic framework to manage resources from Grids, Clouds and dedicated networks in an energy-efficient way.

scholarly journals A survey on software methods to improve the energy efficiency of parallel computing

2016 ◽  
Vol 31 (6) ◽  
pp. 517-549 ◽  
Author(s):  
Chao Jin ◽  
Bronis R de Supinski ◽  
David Abramson ◽  
Heidi Poxon ◽  
Luiz DeRose ◽  
...  
Keyword(s):  
Energy Efficiency ◽  
Parallel Computing ◽  
Energy Efficient ◽  
Large Scale ◽  
Energy Use ◽  
Power Saving ◽  
Computing Methods ◽  
Power Budget ◽  
Load Imbalance ◽  
Mixed Precision

Energy consumption is one of the top challenges for achieving the next generation of supercomputing. Codesign of hardware and software is critical for improving energy efficiency (EE) for future large-scale systems. Many architectural power-saving techniques have been developed, and most hardware components are approaching physical limits. Accordingly, parallel computing software, including both applications and systems, should exploit power-saving hardware innovations and manage efficient energy use. In addition, new power-aware parallel computing methods are essential to decrease energy usage further. This article surveys software-based methods that aim to improve EE for parallel computing. It reviews the methods that exploit the characteristics of parallel scientific applications, including load imbalance and mixed precision of floating-point (FP) calculations, to improve EE. In addition, this article summarizes widely used methods to improve power usage at different granularities, such as the whole system and per application. In particular, it describes the most important techniques to measure and to achieve energy-efficient usage of various parallel computing facilities, including processors, memories, and networks. Overall, this article reviews the state-of-the-art of energy-efficient methods for parallel computing to motivate researchers to achieve optimal parallel computing under a power budget constraint.

scholarly journals Energy efficiency of a passenger ship in Turkey

2020 ◽  
Vol XXIII (1) ◽  
pp. 15-21
Author(s):  
Aydɪn Tokuslu
Keyword(s):  
Energy Efficiency ◽  
Energy Efficient ◽  
Performance Standards ◽  
Design Stage ◽  
Exhaust Gas ◽  
Passenger Ships ◽  
Exhaust Gas Emission ◽  
And Performance ◽  
Energy Efficiency Design Index ◽  
Harmful Effects

The International Maritime Organization (IMO) introduced a new measure in 2011 with a set of technical innovation and performance standards to increase the energy efficiency of new ships at the design stage and aim to reduce CO2 emissions. This new measure is called Energy Efficiency Design Index (EEDI) and this became a fast key instrument for the ships to be energy efficient. In this paper, one of the passenger ships in the Istanbul Strait was investigated and its emissions were estimated. The Energy Efficiency Design Index (EEDI) of the ship was calculated. EEDI formula equations based on the study of passenger ships have been compared with the IMO other ship equations and some useful proposals have been presented to reduce the harmful effects of CO2 exhaust gas emission.

A new case-deletion strategy for case-base maintenance based on K-means Clustering Algorithm applied to medical data

2021 ◽  
pp. 1-19
Author(s):  
Akila Djebbar ◽  
Hayet Farida Merouani ◽  
Hayet Djellali
Keyword(s):  
Large Scale ◽  
Clustering Algorithm ◽  
Case Based Reasoning ◽  
Medical Systems ◽  
Medical Databases ◽  
Case Deletion ◽  
Systems Research ◽  
System Maintenance ◽  
Case Base ◽  
And Performance

Case-Based Reasoning (CBR) system maintenance is an important issue for current medical systems research. Large-scale CBR systems are becoming more omnipresent, with immense case libraries consisting of millions of cases. Case-Base Maintenance (CBM) is the implementation of the following policies allowing to revise the organization and/or the content (information content, representation field of application, or the implementation) of the Case Base (CB) to improve future thinking. Diverse case-base deletion and addition policies have been proposed which claim to preserve case-base competence. This paper presents a novel clustering-based deletion policy for CBM that exploits the K-means clustering algorithm. Thus, CBM becomes a central subject whose objective is to guarantee the quality of the CB and improve the performance of CBM. The proposed approach exploited clustering, which groups similar cases using the K-means algorithm. We rely on the characterization made of the different cases in the CB, and we find this characterization by a method based on a criterion of competence and performance. From this categorization, case deletion becomes obvious. This quality depends on the competence and performance of the CB. Test results show that the proposed deletion strategy improved the efficiency of the CB while preserving competence.Furthermore, its performance was 13% more reliable. The effectiveness of the proposed approach examined on the medical databases and its performance has been compared with the existing approaches on deletion policy. Experimental results are very encouraging.

A Network Coding-Based Energy-Efficient Multicast Routing in Multi-Hop Wireless Networks

2013 ◽  
Vol 765-767 ◽  
pp. 1775-1779
Author(s):  
Ding De Jiang ◽  
Ya Li ◽  
Wei Han Zhang ◽  
Wen Pan Li ◽  
Chun Ping Yao
Keyword(s):  
Energy Efficiency ◽  
Wireless Networks ◽  
Network Coding ◽  
Energy Efficient ◽  
Multicast Routing ◽  
Routing Algorithm ◽  
Performance Improvements ◽  
And Performance ◽  
Multicast Routing Algorithm ◽  
New Generation

In this paper, an energy-efficient multicast routing algorithm in multi-hop wireless networks is proposed aiming at new generation wireless communications. Different from the previous methods, this paper targets maximizing the energy efficiency of networks. In order to get the optimal energy efficiency to build the network multicast route, our proposed method tries to maximize the network throughput and minimize the network energy consumption by exploiting network coding and sleeping scheme. Simulation results show that the proposed algorithm has better energy efficiency and performance improvements comparing with the existing methods.

Sign in / Sign up

Export Citation Format

Share Document