A5044146242- Parallel Computing and Optimization Techniques
- Scientific Computing and Data Management
- Cloud Computing and Resource Management
- Radiation Effects in Electronics
- Embedded Systems Design Techniques
- Distributed and Parallel Computing Systems
- Advanced Data Storage Technologies
- Low-power high-performance VLSI design
- Advancements in Semiconductor Devices and Circuit Design
- Simulation Techniques and Applications
- Semiconductor materials and devices
- Scheduling and Optimization Algorithms
- Semantic Web and Ontologies
- Manufacturing Process and Optimization
- Advanced Database Systems and Queries
- Computer Graphics and Visualization Techniques
- Topological and Geometric Data Analysis
- IoT and Edge/Fog Computing
- Green IT and Sustainability
- Real-Time Systems Scheduling
- Advanced Memory and Neural Computing
- Machine Learning in Materials Science
- Numerical Methods and Algorithms
- Logic, programming, and type systems
- Data Stream Mining Techniques
IBM Research - Ireland
2018-2024
University of Thessaly
2015-2021
Centre for Research and Technology Hellas
2015-2017
Several applications may trade-off output quality for energy efficiency by computing only an approximation of their output. Current approaches to software-based approximate often require the programmer specify parts code or data structures that can be approximated. A largely unaddressed challenge is how automate analysis significance quality. To this end, we propose a methodology and toolset automatic analysis. We use interval arithmetic algorithmic differentiation in our profile-driven yet...
We introduce a task-based programming model and runtime system that exploit the observation not all parts of program are equally significant for accuracy end-result, in order to trade off quality outputs increased energy-efficiency. This is done structured flexible way, allowing easy exploitation different points quality/energy space, without adversely affecting application performance. The can apply number policies decide whether it will execute less-significant tasks accurately or...
We introduce a task-based programming model and runtime system that exploit the observation not all parts of program are equally significant for accuracy end-result, in order to trade off quality outputs increased energy-efficiency. This is done structured flexible way, allowing easy exploitation different points quality/energy space, without adversely affecting application performance. The can apply number policies decide whether it will execute less-significant tasks accurately or...
To improve power efficiency, researchers are experimenting with dynamically adjusting the voltage and frequency margins of systems to just above minimum required for reliable operation. Traditionally, manufacturers did not allow reducing these margins. Consequently, existing studies use system simulators, or software fault-injection methodologies, which slow, inaccurate cannot be applied on realistic workloads. However recent CPUs operation outside nominal voltage/frequency envelope. We...
This article introduces a significance-centric programming model and runtime support that sets the supply voltage in multicore CPU to sub-nominal values reduce energy footprint provide mechanisms control output quality. The developers specify significance of application tasks respecting their contribution quality check repair functions for handling faults. On system, we evaluate five benchmarks using an quantifies reduction. When executing least-significant unreliably, our approach leads 20%...
Computational workflows are important methods for automating complex data-generation and analysis pipelines. Workflows composed of sub-graphs that perform specific tasks. Certain may appear in multiple workflows. This implies the same task, with input, execute times thereby wasting computational resources. Additionally, hybrid cloud environments spanning across on-premises public deployments increasingly popular. Identical tasks not only but also on heterogeneous environments. Memoization, a...
Approximate execution is a viable technique for energy-constrained environments, provided that applications have the mechanisms to produce outputs of highest possible quality within given energy budget.
Cloud data centers require enormous amounts of energy to run their clusters computers. There are huge financial and environmental incentives for cloud service providers increase efficiency without causing significant negative impacts on customers' qualities experience. Increasing resource utilization reduces consumption by consolidating workloads fewer machines allows turn off inactive devices. While traditional architectures only allow virtual (VMs) use the memory CPU resources a single...
In this paper we introduce a framework which automates the task of data management for OpenCL programs across multiple devices heterogeneous system. Our approach uses compile-time analysis, based on polyhedral model, to associate computations with they consume / produce. The results analysis are then used by runtime system management. Beyond alleviating programmer from burden management, our enables partitioning all computational systems according power and memory capacity each device, thus...
For micelles, "shape" is prominent in rheological computations of fluid flow, but this often expressed too informally to be useful for rigorous analyses. We formalize topological "shape equivalence" both globally and locally, enable visualization computational dynamics. Although methods provide significant insights into flows, opportunity has been limited by the known difficulties creating representative geometry. present an agile geometric algorithm represent micellar shape input flow...
Hardware reliability is adversely affected by the downscaling of semiconductor devices and scale-out systems necessitated modern applications. Apart from crashes, this unreliability often manifests as silent data corruptions (SDCs), affecting application output. Therefore, we need low-cost low-human-effort solutions to reduce incidence rate effects SDCs on quality outputs. We propose Artificial Neural Networks (ANNs) an effective mechanism for online error detection. train ANNs using...
Reducing energy consumption is one of the key challenges in computing technology. One factor that contributes to high all parts program are considered equally significant for accuracy end-result. However, many cases, computations can be performed an approximate way, or even dropped, without affecting quality final output a degree. In this paper, we introduce task-based programming model and runtime system exploit observation trade off outputs increased energy-efficiency. This done structured...
The use of approximation is fundamental in computational science. Almost all methods adopt approximations some form order to obtain a favourable cost/accuracy trade-off and there are usually many that could be used. As result, when researcher wishes measure property system with technique, they faced an array options. Current workflow frameworks focus on helping researchers automate sequence steps particular platform. aim often measurement property. However these unaware may large number ways...