h1. Laurent : non présent car meeting ELCI chez Safran/Chateaufort
- Equipe associée SUSTAM (Sustainable Ultra Scale compuTing, dAta and energy Management) entre Avalon et Rutgers University (Manish Parashar) acceptée et financée par Inria pour 3 ans. * Article soumis a Networking avec Radu/MArcos/Olivier, * Article presque soumis à IJHPCA avec Issam * Article CCGrid accepté avec Bull et Grenoble * Article Journal CCPE avec Daniel accepté pour publicaitonen 2017 (sur projet Nu@ge) Continue reading
* Papier IPDPS → PDCO
** reject @ IPDPS
** rework + resubmission @ PDCO (deadline extended to 2017-01-31)
* Following work
** Now moving toward instance choice (homogeneous → heterogeneous)
** Thinking about cluster growth in this context
* Work w\ Ctrl-A
** Autonomic deployment of the meta-tasks produced by our clustering algorithm
* Results of the Byosil (m)eating day
** Discussion w\ Biologist on workflow deployment (description of what we do, how we can solve some of there issues, how they can provide us good usecases …) → meeting soon
* Mid-PhD (D2) evaluation
* Also, moving forward with the PhD students website (project)
** Hope to get some PhD students motivated and actually do something
2016-06-07 – Hayri Acar
Title: Software development methodology in a Green IT environment
Speaker: Hayri Acar
Abstract: Writing sustainable, power efficient and green software necessitates understanding the power consumption behavior of a computer program. One of the benefits is the fact that developers, by improving their source code implementations, can optimize power consumption of a software. Existing power consumption models need to be improved by taking into account more components susceptible to consume energy during runtime of an application. In this paper, we first present a detailed classification of previous works on power consumption modelization. Then, we introduce TEEC (Tool to Estimate Energy Consumption) model in order to estimate the power consumed by CPU, memory and disk due to the execution of an application at runtime. The main goal is to guide developers to improve their source code for optimizing energy consumption. TEEC enables determining the part of the code consuming the highest power. This will help to obtain a less energy consuming software with the same functionalities.
2016-05-17 – Philippe Virouleau
Title: Improving OpenMP compilers and runtimes for task-based applications on NUMA architectures
Speaker: Philippe Virouleau
Abstract: The most popular architecture for building large-scale shared memory machines nowodays is the NUMA architecture (Non-Uniform Memory Access). In such architecture, the shared memory and cores are split in nodes, physically separated from each others. The memory access time depends on which core wants to access which data, and the distance between the core and the data’s NUMA node. A popular application design to efficiently exploit the parallelism offered by large multi processors architectures is to use fine-grain dependent tasks. In order to successfully use this approach on NUMA architectures, the application’s programmer should take great care of the locality between the task being executed, and the data manipulated by the task. OpenMP is the de-facto standard for shared-memory parallel programming, and the revision 4.0 introduced the tasks with dependencies model, in which the programmer can specify which data are read and/or written by a given task. Having the runtime use these informations is a first step to dynamically improve the application’s performances, however more flexibility could be given to the programmer, e.g. by giving him the possibility to specify which data are important for a given task. This presentation will describe my PhD works, which focus on proposing and evaluating compilers and runtimes extensions to help reduce the impact of NUMA architectures on the application’s performances and scalability.
2016-05-03 – Issam Raïs
Title: Towards Green Exascale Computing Challenges
Speaker: Issam Raïs
Abstract: Exascale is coming. Massively heterogeneous machines with hundreds of thousands of computing nodes and each of these nodes possessing hundreds of cores, bounded to each other by a dedicated and efficient network. On every component composing such a machine, we can spot many techniques to reduce energy consumption while maintaining good computing power. In such a context, this presentation aims at presenting the problematics being tackled in the current thesis.
2016-04-19 Hadrien Croubois
Title: Toward an autonomic approach of workflows distribution on cloud ;
Speaker: Hadrien Croubois
Abstract: Advances in distributed systems technologies require a constant rethinking of previous deployments methods. Development of the Cloud paradigm is symptomatic of a broader tendency toward more dynamicity in the management of tasks and resources. However, the scientific computing still mostly uses old paradigms when deploying complex workflows. Our focus is therefore to propose a solution that will make the link between the needs of user in terms of scientific computation and the features offered by cloud providers. After having modelled both those needs and features, we describe the different mechanisms which are part of what should allow for an autonomous platform dedicated to collaborative scientific computing.
2016-04-16 Issam Raïs
Title: An analysis of the feasibility of energy harvesting with thermoelectric generators on petascale and exascale systems;
Speaker: Issam Raïs
Abstract: The heat induced by computing resources is generally a waste of energy in supercomputers. This is especially true in very large scale supercomputers, where the produced heat has to be compensated with expensive and energy consuming cooling systems. An analysis of the feasibility of energy harvesting with thermoelectric generators on petascale and exascale systems; Energy is a critical point for future supercomputing trends that currently try to achieve exascale, without having its energy consumption reaching an important fraction of a nuclear power plant. Thus, new ways of generating or recovering energy have to be explored. Energy harvesting consists in recovering wasted energy. ThermoElectric Generators (TEGs) aim to recover energy by converting wasted dissipated energy into usable electricity. By combining computing units (CU) and TEGs at very large scale, we spotted a potential way to recover energy from wasted heat generated by computations on supercomputers. In this paper, we study the potential gains in combining TEGs with computational units at petascale and exascale. We present the technology behind TEGs, the study of a typical supercomputer environment, and finally our results concerning binding TEGs and computational units in a petascale and exascale system. With the available technology, we demonstrate that the use of TEGs in a supercomputer environment could be realistic and quickly profitable, and hence have a positive environmental impact.
2016-03-29 Romaric Guillier
Speaker: Romaric Guillier.
Abstract: an experimental platform for BioInformatics; Abstract: The E-Biothon platform is an experimental Cloud platform to help speed up and advance research in biology, health and environment. It is based on a Blue Gene/P system and a web portal that allow members of the bioinformatics community to easily launch their scientific applications. This presentation describes the technical capacities of the platform, the different applications supported and finally a set of user experiences on the platform.
Title: Detecting Silent Data Corruption Using an Auxiliary Method and External Observer
Speaker: Hadrien Croubois
Abstract: HPC platforms and application are becoming increasingly complex. Consequently, protecting results against all forms of corruption and ensuring trustworthiness are becoming more important. While previous work focuses on application-specific detectors, the dataflow manager in our current work in the Decaf project aims to have an efficient generic mechanism. We address those issues using new replication patterns that rely on the use of an auxiliary method and an external learning observer. In this talk, we present both the theoretical validation mechanisms and different use cases where our mechanism can be applied to detect silent data corruption.