Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National.

Slides:



Advertisements
Similar presentations
Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Connecting Canadians to Earth Observation David.
Advertisements

SLA-Oriented Resource Provisioning for Cloud Computing
CERN LCG Overview & Scaling challenges David Smith For LCG Deployment Group CERN HEPiX 2003, Vancouver.
Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Natural Research Council Canada Conseil national.
Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Natural Research Council Canada Conseil national.
Evaluating landscape flammability through simulation modeling Marc Parisien 1, Victor Kafka 2, Bernie Todd 1, Kelvin Hirsch 1, and Suzanne Lavoie 1 1 Canadian.
S.Chechelnitskiy / SFU Simon Fraser Running CE and SE in a XEN virtualized environment S.Chechelnitskiy Simon Fraser University CHEP 2007 September 6 th.
Analysis of the fire environment: MW FMA area March 1, 2001 Analysis of the fire environment: MW FMA area March 1, 2001 Natural Resources Canada Ressources.
Spatial Fire Management System
A Grid Parallel Application Framework Jeremy Villalobos PhD student Department of Computer Science University of North Carolina Charlotte.
David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL March 25, 2003 CHEP 2003 Data Analysis Environment and Visualization.
Fault-tolerant Adaptive Divisible Load Scheduling Xuan Lin, Sumanth J. V. Acknowledge: a few slides of DLT are from Thomas Robertazzi ’ s presentation.
CUDA Programming Lei Zhou, Yafeng Yin, Yanzhi Ren, Hong Man, Yingying Chen.
Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Natural Research Council Canada Conseil national.
Ian Gable University of Victoria/HEPnet Canada 1 GridX1: A Canadian Computational Grid for HEP Applications A. Agarwal, P. Armstrong, M. Ahmed, B.L. Caron,
CompuCell Software Current capabilities and Research Plan Rajiv Chaturvedi Jesús A. Izaguirre With Patrick M. Virtue.
Ashok Agarwal 1 BaBar MC Production on the Canadian Grid using a Web Services Approach Ashok Agarwal, Ron Desmarais, Ian Gable, Sergey Popov, Sydney Schaffer,
Test Of Distributed Data Quality Monitoring Of CMS Tracker Dataset H->ZZ->2e2mu with PileUp - 10,000 events ( ~ 50,000 hits for events) The monitoring.
Chao “Bill” Xie, Victor Bolet, Art Vandenberg Georgia State University, Atlanta, GA 30303, USA February 22/23, 2006 SURA, Washington DC Memory Efficient.
Network Aware Resource Allocation in Distributed Clouds.
November , 2009SERVICE COMPUTATION 2009 Analysis of Energy Efficiency in Clouds H. AbdelSalamK. Maly R. MukkamalaM. Zubair Department.
Daniel Vanderster University of Victoria National Research Council and the University of Victoria 1 GridX1 Services Project A. Agarwal, A. Berman, A. Charbonneau,
March 3rd, 2006 Chen Peng, Lilly System Biology1 Cluster and SGE.
Scheduling Many-Body Short Range MD Simulations on a Cluster of Workstations and Custom VLSI Hardware Sumanth J.V, David R. Swanson and Hong Jiang University.
GRAM5 - A sustainable, scalable, reliable GRAM service Stuart Martin - UC/ANL.
Wenjing Wu Andrej Filipčič David Cameron Eric Lancon Claire Adam Bourdarios & others.
Sep 21, 20101/14 LSST Simulations on OSG Sep 21, 2010 Gabriele Garzoglio for the OSG Task Force on LSST Computing Division, Fermilab Overview OSG Engagement.
CSF4 Meta-Scheduler Name: Zhaohui Ding, Xiaohui Wei
High Performance Computing on Virtualized Environments Ganesh Thiagarajan Fall 2014 Instructor: Yuzhe(Richard) Tang Syracuse University.
BOF: Megajobs Gracie: Grid Resource Virtualization and Customization Infrastructure How to execute hundreds of thousands tasks concurrently on distributed.
Ashok Agarwal University of Victoria 1 GridX1 : A Canadian Particle Physics Grid A. Agarwal, M. Ahmed, B.L. Caron, A. Dimopoulos, L.S. Groer, R. Haria,
Parallel Programming on the SGI Origin2000 With thanks to Igor Zacharov / Benoit Marchand, SGI Taub Computer Center Technion Moshe Goldberg,
© July 2011 Linear and Nonlinear Imaging Spectrometer Denoising Algorithms Assessed Through Chemistry Estimation David G. Goodenough 1,2, Geoffrey S. Quinn.
Tool Integration with Data and Computation Grid GWE - “Grid Wizard Enterprise”
E-science grid facility for Europe and Latin America E2GRIS1 Gustavo Miranda Teixeira Ricardo Silva Campos Laboratório de Fisiologia Computacional.
Operating a distributed IaaS Cloud for BaBar Randall Sobie, Ashok Agarwal, Patrick Armstrong, Andre Charbonneau, Kyle Fransham, Roger Impey, Colin Leavett-
Improving I/O with Compiler-Supported Parallelism Why Should We Care About I/O? Disk access speeds are much slower than processor and memory access speeds.
U N I V E R S I T Y O F S O U T H F L O R I D A Hadoop Alternative The Hadoop Alternative Larry Moore 1, Zach Fadika 2, Dr. Madhusudhan Govindaraju 2 1.
Ian Gable University of Victoria 1 Deploying HEP Applications Using Xen and Globus Virtual Workspaces A. Agarwal, A. Charbonneau, R. Desmarais, R. Enge,
June 30 - July 2, 2009AIMS 2009 Towards Energy Efficient Change Management in A Cloud Computing Environment: A Pro-Active Approach H. AbdelSalamK. Maly.
Towards large-scale parallel simulated packings of ellipsoids with OpenMP and HyperFlow Monika Bargieł 1, Łukasz Szczygłowski 1, Radosław Trzcionkowski.
Economic and On Demand Brain Activity Analysis on Global Grids A case study.
Tool Integration with Data and Computation Grid “Grid Wizard 2”
Performance guided scheduling in GENIE through ICENI
Condor + Cloud Scheduler Ashok Agarwal, Patrick Armstrong, Andre Charbonneau, Ryan Enge, Kyle Fransham, Colin Leavett-Brown, Michael Paterson, Randall.
Miron Livny Computer Sciences Department University of Wisconsin-Madison Condor and (the) Grid (one of.
Alien and GSI Marian Ivanov. Outlook GSI experience Alien experience Proposals for further improvement.
Holding slide prior to starting show. Processing Scientific Applications in the JINI-Based OGSA-Compliant Grid Yan Huang.
Written By: Presented By: Swarup Acharya,Amr Elkhatib Phillip B. Gibbons, Viswanath Poosala, Sridhar Ramaswamy Join Synopses for Approximate Query Answering.
PARALLEL AND DISTRIBUTED PROGRAMMING MODELS U. Jhashuva 1 Asst. Prof Dept. of CSE om.
1 This Changes Everything: Accelerating Scientific Discovery through High Performance Digital Infrastructure CANARIE’s Research Software.
INTRODUCTION TO GRID & CLOUD COMPUTING U. Jhashuva 1 Asst. Professor Dept. of CSE.
Computer Photography -Scene Fixed 陳立奇.
Enabling Grids for E-sciencE LRMN ThIS on the Grid Sorina CAMARASU.
INTRODUCTION TO XSEDE. INTRODUCTION  Extreme Science and Engineering Discovery Environment (XSEDE)  “most advanced, powerful, and robust collection.
Generalized and Hybrid Fast-ICA Implementation using GPU
Dag Toppe Larsen UiB/CERN CERN,
Credits: 3 CIE: 50 Marks SEE:100 Marks Lab: Embedded and IOT Lab
Dag Toppe Larsen UiB/CERN CERN,
GWE Core Grid Wizard Enterprise (
Ruslan Fomkin and Tore Risch Uppsala DataBase Laboratory
Design of a Multi-Agent System for Distributed Voltage Regulation
Mean Value Analysis of a Database Grid Application
An easier path? Customizing a “Global Solution”
IMAGE MOSAICING MALNAD COLLEGE OF ENGINEERING
CIS16 Application Development – Programming with Visual Basic
Multi-hop Coflow Routing and Scheduling in Data Centers
Multithreaded Programming
Parallel Programming in C with MPI and OpenMP
Approximate Mean Value Analysis of a Database Grid Application
Presentation transcript:

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada NEP-02: Service Oriented Scientific Computing for SAFORAH – Non-linear Denoising Parallelization Demo Ashok Agarwal, Patrick Armstrong, Andre Charbonneau, Hao Chen, Ronald J. Desmarais, Ian Gable, David G. Goodenough, Piper Gordon, Aimin Guan, Roger Impey, Kelsey Lang, Belaid Moa, Susan Perkins, Wayne Podaima, Randall Sobie  February 15, 2011

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Road Map  Introduction Introduction to non-linear denoising Problem – algorithm is expensive Solution – parallelize  Non-linear denoising application How we parallelize using the grid?  Application work done during extension Non-linear Denoising application SAFORAH: CUDOS web interface  Future work

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Introduction  Hyperspectral noise has been assumed to be linear Linear Stochastic Processes were used to model the noise  Recently, Han and Goodenough (2009) have proved that this is not the case Non-linear analysis is used to reduce the noise Results show significant signal to noise ratio SNR boost:

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Introduction Original Image Difference Image Denoised Image Aviris Image (600x1000x179) RGB: 1503, 750, 645 nm Linear 2% trimming Non-linear Denoising Results:

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Introduction  Typical hyperspectral image cube 600 samples, 1000 lines, 179 bands Over 200 MB  Problem: The algorithm is computationally very expensive. This 200MB image will take approximately four days to process on a single cpu.  Solution: Give the algorithm more cpu’s!

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada   Solution: Give the algorithm more cpu’s: Parallelize using the computational grid.   Embarrassingly parallel problem: No spatial data interdependence Data can be divided spatially for processing Non-linear Denoising

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Non-linear Denoising   The image is split into N pieces using the gdal library:   Number of pieces based on image size.   For example, this sample image is broken into 250 pieces.

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Non-linear Denoising   Each piece is sent to the grid with the application source code and a script to compile and run it: Denoise.cpp Makefile Denoise.sh Computational grid Job # n

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Non-linear Denoise Computational grid Non-linear Denoising Job # 1 Job # 2 Job # 3 Job # … Job # n Metascheduler Resource 1 Resource 2 Resource m   The pieces are sent to the metascheduler The metascheduler allocates resources for the jobs:

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Non-linear Denoising Computational grid   The nodes copy the completed images back to the local machine The Non-linear Denoising application stitches the pieces together:

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada   Advantages:   Two orders of magnitude faster!   Completed in as little as 1 hour on a grid with 74 nodes, instead of 4 days, or about 100 hours.   The time to process dependent on: Size of image file Number of nodes available on the grid How many other jobs are running on the grid Non-linear Denoising

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada  The activities in this quarter dealt with making the parallel Non-linear Denoising (NLD) application available in the CUDOS GUI, and preparing the application and software infrastructure for longevity. Developed the parallel NLD application Improved stability Improved response time Made NLD available from CUDOS GUI Ran end-user testing for NLD  The infrastructure activities included: Improvements to source code version management Extension

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada 13 Changes to the Architecture Extension

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada System Flow Chart Results Pool of Computational Resources Splitter Scatterer Gatherer Job Scheduler Product Request Merger Gridway MetaschedulerGridway Metascheduler Portable Batch System (PBS)Portable Batch System (PBS) Extension

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Demo of Automated Parallelization of Denoising SAFORAH CUDOS

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Future Work   Update the other existing EO applications according to the user requirements on parallelization (Radarsat 2 decomposition, Aboveground carbon mapping)   Update Grid Web Integration Service to use supported APIs Globus 4.0 no longer supported Other possibilities: Condor Globus 5.x Cloud computing   Add more nodes to grid to improve performance. Make more applications available on OGC and on CUDOS

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Conclusion  Accomplished what we set out to do for extension: Parallelized Non-linear Denoising. Using grid-computing for speed and efficiency. Obtained significant performance improvements from parallelization: ~100 x faster!! Made the application available on SAFORAH Improved the system infrastructure  Thank you CANARIE!

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada 18 CUDOS Portal - Create New Product

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada 19 CUDOS Portal - Product Parameters

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada 20 CUDOS Portal - Product Order

Natural Resources Canada Ressources naturelles Canada Canadian Forest Service Service canadien des forêts Conseil national de recherches Canada National Research Council of Canada Results on noise reduction (1) The effectiveness of noise reduction by the LGP was evaluated by comparing the noise intensity and signal-to-noise ratio (SNR) between the original and the de-noised images. A covariance method proposed by Roger and Arnold (1996) was employed here for noise estimation. Without relying on any additional information, this method is capable of producing reasonable noise estimates based on the covariance matrix of the hyperspectral image alone. The average noise intensity of the original AVIRIS image was estimated at 1.2, which varies between 0.3 (minimum) and 11.3 (maximum) across all bands. The LGP removed an average of 34.4% of the noise in the AVIRIS image, which left the noise intensity in the post-denoised image varying between 0.2 and 9.0 with an average of 0.9.