My research broadly falls into the general areas of parallel and distributed computing with a focus on green High Performance Computing (HPC) and extreme-scale computing. Specifically, my main research addresses key issues to enable the scalability and energy efficiency of next generation computing systems, including:
- Autonomic application-aware cross-layer power management
- Energy efficiency of scalable scientific data analysis pipelines
- In-situ data analytics and co-processing at extreme scales
Complementing the previous key research issues, my research also considers cloud computing mechanisms and abstractions for large-scale computing systems and includes the federation of various types of advanced cyber-infrastructure for running complex Computational and Data-Enabled Scientific and Engineering (CDS&E) problems.
These are important challenges and in its Strategy for American Innovation, the Obama administration listed extreme-scale computing among the Grand Challenges of the 21st century in science, technology and innovation.
Project Title: II-NEW: An Experimental Platform for Investigating Energy-Performance Tradeoffs for Systems with Deep Memory Hierarchies
PI: Manish Parashar, Co-PIs: Ivan Rodero, Dario Pompili
National Science Foundation (NSF CRI)
Amount and Dates: $300,000 (10/01/2013 - 09/30/2016)
Project Title: Scalable Data Coupling Abstraction for Data-Intensive Simulation Workflows
PI: Manish Parashar, Co-PI: Ivan Rodero
National Science Foundation (NSF CDSE)
Amount and Dates: $547,283 (05/01/2013 - 04/30/2016)
Project Title: Exploring Cloud Paradigm and Practices for Science and Engineering
PI: Manish Parashar, Co-PIs: Ivan Rodero, Javier Diaz
National Science Foundation (NSF EAGER)
Amount and Dates: $299,984 (06/01/2013 - 05/31/2015)
Project Title: Cross-Layer Application-Aware Power/Energy Management on Systems with Deep Memory Hierarchies and Accelerators
PI: Ivan Rodero
Beacon Project, National Science Foundation (May'13 - Present)
Amount: 7,500 node hours (i.e., 120,000 CPU-core hours + 1.8M MIC-core hours)
Project Title: Adaptive Parametric Space Exploration and Data Assimilation on Federated Heterogeneous Resources
PI: Javier Diaz, Co-PIs: Ivan Rodero, Manish Parashar
Source of Support: National Science Foundation XSEDE startup grant CCR130025 (Mar'13 - Present)
Amount: 100,000 SUs
Project Title: Towards Extreme Scale Computing in Federated HPC Cyber-infrastructure
PI: Ivan Rodero
Source of Support: US Department of Energy NERSC startup awards 85357/86428 (Oct'11 - Present)
Amount: 100,000 SUs (startup + production)
Project Title: Evaluation of MPI Collectives for HPC Applications on Distributed Virtualized Environments
PI: Ivan Rodero
Source of Support: National Science Foundation Future Grid project number 159 (Oct'11 - Present)
Project Title: Towards Layered Cloud Federation Model in HPC Cyber-infrastructure
PI: Ivan Rodero
Source of Support: National Science Foundation XSEDE startup grant CCR110035 (Oct'11 - Apr'13)
Amount: 175,000 SUs
Current Research Projects
Green High Performance Computing (http://greenhpc.org) (Sep'09 - Present)
This is my main area of research that I have developed and lead at Rutgers since I joined. It includes several efforts in the intersection of energy efficiency, scalable computing and high performance computing. This effort is one of the first to explore key related problems, including:
- Application-aware cross-layer power management for High Performance Computing systems
The goals of this effort are investigating aggressive power management strategies and their impact on the overall energy consumption and developing autonomic advanced runtimes to improve the energy efficiency of High Performance Computing systems, datacenters and platforms based on many-core architectures such as the Intel Single-chip Cloud Computer (SCC). It investigates proactive component-based power management and cross-layer interactions using PGAS language extensions and runtime mechanisms that can be used to achieve a wide range of energy and performance behaviors and manipulate power/performance tradeoffs
Collaborators: Oak Ridge National Laboratory and Intel.
- Energy efficiency for scientific data analysis pipelines at scale
The goals of this effort are understanding power/performance behaviors and tradeoffs associated with data placement, data movement and data processing associated with data analytics pipelines on systems with emerging architectures and deep memory hierarchies, and to develop strategies that can fundamentally enable data-intensive workflows on current and future large-scale systems. Furthermore, it addresses energy/power-efficiency tradeoffs in a holistic manner in combination with performance, resilience, quality of solution, and other objectives.
Collaborators: Oak Ridge National Laboratory and Fusion-io.
- In-situ data analytics and co-processing at extreme scale
This research effort focuses on developing new formulations and analysis strategies to support the increasing volumes and rates at which scientific simulations running at extreme scale generate data, which needs to be transported and analyzed before scientific discovery can be realized. Its overarching goal is exploring data-related energy/performance trade-offs at extreme scales. Specifically, it aims at analyzing the behavior of large-scale simulation workflows with an in-situ and other data analytics pipelines, running on a current high-end computing platform and beyond to develop performance and power models, which can be validated using an instrumented platform. Models can be used then to explore energy/performance tradeoffs on current systems, to help answer system design questions, and to analyze the power requirements and usage modes for emerging architectures such as the Intel Many Integrated Core (MIC) architecture.
Collaborators: Sandia National Laboratories, Lawrence Livermore National Laboratory, Los Alamos National Laboratory, Pacific Northwest National Laboratory, Oak Ridge National Laboratory and University of Utah.
Cloud Abstractions for CDS&E (Sep'11 - Present)
The goal of this project is to understand science and engineering application formulations that are meaningful in a hybrid federated cyber- infrastructure that includes clouds, and to explore programming and middleware support that can enable these applications, including application formulations, programming models, abstractions and systems, and middleware stacks and services. Key efforts include:
- CometCloud: An autonomic computing cloud engine (http://cometcloud.org)
- The Uber-Cloud Experiment (Round 1) - Team 23 (Sep'12 - Nov'12)
- The Uber-Cloud Experiment (Round 2) - Team 53 (Nov'12 - Apr'13) - http://nsfcac.rutgers.edu/CometCloud/uff/
"CometCloud Adaptive Sparse Grid Collocation over Federated Heterogeneous Computing Architectures"
Collaborators: Iowa State University.
Content-based Medical Image Retrieval (Jan'12 - Present)
The overarching goal of this project is to develop mechanisms and techniques based on novel software and technologies to accelerate medical image processing algorithms and enable their execution at scale on high performance and distributed computing systems. In collaboration with Center for Biomedical Imaging & Informatics at UMDNJ-Robert Wood Johnson, Medical School at The Cancer Institute of New Jersey.
Cross-layer Autonomic Research (NSF I/UCRC under award 0758596) (Sep'09 - Present)
The I/UCRC focuses on multi university research on improving the design and engineering systems that are capable of funning themselves, adapting their resources and operations to current workloads and anticipating the needs of their users. The project aims at improving hardware, networks and storage, middleware, service and information layers used by modern industry.
Technology Transfer Projects
- Cloud Computing Cyber-security [with Avirtek and University of Arizona] (Feb'12 - Dec'12)
- Experiments with the PGAS Model on SCC [with Intel Corp.] (May'11 - May'12)
- Online Analytics using Cluster-based Feature Tracking [with Xerox Corp.] (Sep'10 - Sep'12)
- Layered Cloud Federation model [with IBM Corp. and Florida International University] (Sep'09 - Aug'12)
Previous Projects (2003-2009)
- Programming models and parallel execution environments (Jul'09 - Sep'09)
Government of Catalonia, 2009-SGR-980
- High Performance Computing V (Oct'07 - Sep'09) Spanish Ministry of Science and Technology, TIN2007-60625
- MareIncognito (Dec'06 - Sep'09)
Barcelona Supercomputing Center - IBM. WP Load Balancing
- XtreemOS (EU Project, IST-FP6-033576) (Sep'06 - Sep'09)
WP3.3: Application Execution Management
- Latin American Grid (Partnership between IBM and Universities) (Apr'06 - Sep'09)
Meta-scheduling and workflow project
- CoreGRID Network of Excellence (EU Project, FP6-004265) (Sep'04 - Sep'08)
Institute on Resource Management and Scheduling
- Thematic Network for the Coordination of the Middleware Activities in Grid (Jul'04 - Jul'07)
Spanish Ministry of Science and Technology, TIN2002-12422-E and TIN2005-25849-E
- High Performance Computing IV (Jul'04 - Jul'07)
Spanish Ministry of Science and Technology, TIN2004-07739-C02-01
- HPC-Europa (EU Project, FP6-506079) (Jul'04 - Dec'06)
JRA2: Single Point of Access
- eNANOS (Sep'03 - Sep'09)
Barcelona Supercomputing Center