View: session overviewtalk overview
Abstract:
The total energy consumption and carbon footprint of data centers will substantially increase in the next years after more than a decade of relative slow growths. This trend has been indicated by a new study modelling the carbon footprint of ICT in Germany within the framework of the GreenICT@FMD framework project. The keynote has two objectives. The first objective of the keynote is to provide an insight into the methodological approach to the lifecycle environmental assessment of computer systems. In this context, the factors that lead to increasing environmental impacts in the production and use of enterprise computers are explained in detail. This includes, among other things, topics such as the environmental impact of semiconductor production in the context of Moore's Law and the topic of the use of renewable energies. The second objective of the keynote addresses the immediate factors that are currently contributing to a rapid increase in computer power consumption in data centers. Using the example of the current technical development of high-end CPUs, the declining of server-related energy efficiency is analyzed. Other topics include trends concerning chip cooling and waste heat utilization in data centers. The keynote intends to provide a holistic perspective on the environmental aspects of higher performing computer systems.
Bio:
Lutz Stobbe is a senior scientist at the Fraunhofer Institute for Reliability and Microintegration (IZM) with 25 years of work experience in Green information and communication technology (ICT). The research of his group, “Sustainable Networks and Computing'', is focused on the methodical issues of lifecycle assessment (LCA) and applied eco-design for data center and telecommunication equipment. He developed the 5C methodology, which supports a structured modelling of complex lifecycle inventories. As a project manager, he has been involved in dozens of national and international research projects. Most notably are six preparatory studies developing measures for the implementation of the EU Ecodesign Directive. This includes the ENTR Lot 9 study on enterprise servers and data storage equipment. He also led expert teams (Begleitforschung) accompanying large publicly funded research programs such as IT2Green, 5G Industrial Internet, Green HPC, and Green ICT. His work for industry includes trainings and consultations, focusing on applied LCA and eco-design.
16:30 | An Efficient Multi Quantile Regression Network with Ad Hoc Prevention of Quantile Crossing ABSTRACT. This article presents the Sorting Composite Quantile Regression Neural Network (SCQRNN), an advanced quantile regression model designed to prevent quantile crossing and enhance computational efficiency. Integrating ad hoc sorting in training, the SCQRNN ensures non-intersecting quantiles, boosting model reliability and interpretability. We demonstrate that the SCQRNN not only prevents quantile crossing and reduces computational complexity but also achieves faster convergence than traditional models. This advancement meets the requirements of high-performance computing for sustainable, accurate computation. In organic computing, the SCQRNN enhances self-aware systems with predictive uncertainties, enriching applications across finance, meteorology, climate science, and engineering. |
16:55 | Modifiable Artificial DNA - Change your System’s ADNA at any Time PRESENTER: Aleksey Koschowoj ABSTRACT. The Artificial DNA (ADNA) and Artificial Hormone System (AHS) together form a middleware that uses Organic Computing tech- niques to improve the robustness and adaptability of distributed embed- ded systems. These systems then have the properties of self-organization, self-healing, self-configuration and self-improvement. However, the adapt- ability of the system is limited by the rigidity of the ADNA, since it cannot be modified at runtime. Recent research approaches already as- sume the existence of a modifiable ADNA in their applications without actually implementing it or evaluating its behavior. In this paper, we present two crucial steps to extend the ADNA with the ability to allow run-time modifications. First, we describe the possible modifications that can be made at runtime, and how they can be implemented without making significant changes to the ADNA implementation. Second, we provide an experimental evaluation of this new feature and contextualize its behavior within the framework of traditional ADNA. |
17:20 | From Structured to Unstructured: A Comparative Analysis of CV and Graph Models in solving Mesh-based PDEs ABSTRACT. This article investigates the application of computer vision and graph-based models in solving mesh-based partial differential equations within high-performance computing environments. Focusing on structured, graded structured, and unstructured meshes, the study compares the performance and computational efficiency of three computer vision-based models against three graph-based models across three datasets. The research aims to identify the most suitable models for different mesh topographies, particularly highlighting the exploration of graded meshes, a less studied area. Results demonstrate that computer vision-based models, notably U-Net, outperform the graph models in prediction performance and efficiency in two (structured and graded) out of three mesh topographies. The study also reveals the unexpected effectiveness of computer vision-based models in handling unstructured meshes, suggesting a potential shift in methodological approaches for data-driven partial differential equation learning. The article underscores deep learning as a viable and potentially sustainable way to enhance traditional high-performance computing methods, advocating for informed model selection based on the topography of the mesh. |