The HDF Group Parallel HDF5 Developments 1 Copyright © 2010 The HDF Group. All Rights Reserved Quincey Koziol The HDF Group

Slides:



Advertisements
Similar presentations
TWO STEP EQUATIONS 1. SOLVE FOR X 2. DO THE ADDITION STEP FIRST
Advertisements

Advanced Piloting Cruise Plot.
Programming with Android: SDK install and initial setup Luca Bedogni Marco Di Felice Dipartimento di Scienze dellInformazione Università di Bologna.
Copyright © 2003 Pearson Education, Inc. Slide 1 Computer Systems Organization & Architecture Chapters 8-12 John D. Carpinelli.
1 Copyright © 2013 Elsevier Inc. All rights reserved. Chapter 4 Computing Platforms.
Processes and Operating Systems
1 Copyright © 2013 Elsevier Inc. All rights reserved. Chapter 1 Embedded Computing.
1 Copyright © 2013 Elsevier Inc. All rights reserved. Appendix 01.
1 Copyright © 2010, Elsevier Inc. All rights Reserved Fig 2.1 Chapter 2.
1 Copyright © 2013 Elsevier Inc. All rights reserved. Chapter 38.
1 Chapter 40 - Physiology and Pathophysiology of Diuretic Action Copyright © 2013 Elsevier Inc. All rights reserved.
By D. Fisher Geometric Transformations. Reflection, Rotation, or Translation 1.
Towards Automating the Configuration of a Distributed Storage System Lauro B. Costa Matei Ripeanu {lauroc, NetSysLab University of British.
and 6.855J Cycle Canceling Algorithm. 2 A minimum cost flow problem , $4 20, $1 20, $2 25, $2 25, $5 20, $6 30, $
© 1998, Progress Software Corporation 1 Migration of a 4GL and Relational Database to Unicode Tex Texin International Product Manager.
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
0 - 0.
DIVIDING INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
MULTIPLYING MONOMIALS TIMES POLYNOMIALS (DISTRIBUTIVE PROPERTY)
ADDING INTEGERS 1. POS. + POS. = POS. 2. NEG. + NEG. = NEG. 3. POS. + NEG. OR NEG. + POS. SUBTRACT TAKE SIGN OF BIGGER ABSOLUTE VALUE.
SUBTRACTING INTEGERS 1. CHANGE THE SUBTRACTION SIGN TO ADDITION
MULT. INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
FACTORING Think Distributive property backwards Work down, Show all steps ax + ay = a(x + y)
Addition Facts
Making the System Operational
NGS computation services: API's,
INTERNET PROTOCOLS Class 9 CSCI 6433 David C. Roberts Entire contents copyright 2011, David C. Roberts, all rights reserved.
Real Time Versions of Linux Operating System Present by Tr n Duy Th nh Quách Phát Tài 1.
Addison Wesley is an imprint of © 2010 Pearson Addison-Wesley. All rights reserved. Chapter 10 Arrays and Tile Mapping Starting Out with Games & Graphics.
Copyright © 2007 Ramez Elmasri and Shamkant B. Navathe Slide
13 Copyright © 2005, Oracle. All rights reserved. Monitoring and Improving Performance.
Database Performance Tuning and Query Optimization
© 2006 Cisco Systems, Inc. All rights reserved. MPLS v MPLS TE Overview Configuring MPLS TE on Cisco IOS Platforms.
The Platform as a Service Model for Networking Eric Keller, Jennifer Rexford Princeton University INM/WREN 2010.
INTRODUCTION TO SIMULATION WITH OMNET++ José Daniel García Sánchez ARCOS Group – University Carlos III of Madrid.
Trap Diagnostic Facility Todays Software Diagnostic Tool with innovative features for the z/OS software developer Arney Computer Systems.
© S Haughton more than 3?
Making Time-stepped Applications Tick in the Cloud Tao Zou, Guozhang Wang, Marcos Vaz Salles*, David Bindel, Alan Demers, Johannes Gehrke, Walker White.
Lets play bingo!!. Calculate: MEAN Calculate: MEDIAN
Past Tense Probe. Past Tense Probe Past Tense Probe – Practice 1.
User Query Control An Enhancement For AS/400 Query On The IBM iSeries from  Copyright I/O International, 2005 Skip Intro.
This, that, these, those Number your paper from 1-10.
2  Industry trends and challenges  Windows Server 2012: Modern workstyle, enabled  Access from virtually anywhere, any device  Full Windows experience.
Addition 1’s to 20.
25 seconds left…...
Test B, 100 Subtraction Facts
11 = This is the fact family. You say: 8+3=11 and 3+8=11
Week 1.
We will resume in: 25 Minutes.
Figure Essential Cell Biology (© Garland Science 2010)
1 Unit 1 Kinematics Chapter 1 Day
The HDF Group ESIP Summer Meeting Easy access HDF files via Hyrax Kent Yang The HDF Group 1 July 8 – 11, 2014.
Choosing an Order for Joins
Chapter 13 The Data Warehouse
How Cells Obtain Energy from Food
CpSc 3220 Designing a Database
Non-Blocking Collective MPI I/O Routines Ticket #273.
A PLFS Plugin for HDF5 for Improved I/O Performance and Analysis Kshitij Mehta 1, John Bent 2, Aaron Torres 3, Gary Grider 3, Edgar Gabriel 1 1 University.
Parallel HDF5 Introductory Tutorial May 19, 2008 Kent Yang The HDF Group 5/19/20081SCICOMP 14 Tutorial.
1 High level view of HDF5 Data structures and library HDF Summit Boeing Seattle September 19, 2006.
HDF5 A new file format & software for high performance scientific data management.
The HDF Group Multi-threading in HDF5: Paths Forward Current implementation - Future directions May 30-31, 2012HDF5 Workshop at PSI 1.
SciDAC All Hands Meeting, March 2-3, 2005 Northwestern University PIs:Alok Choudhary, Wei-keng Liao Graduate Students:Avery Ching, Kenin Coloma, Jianwei.
Project 4 : SciDAC All Hands Meeting, September 11-13, 2002 A. Choudhary, W. LiaoW. Gropp, R. Ross, R. Thakur Northwestern UniversityArgonne National Lab.
SDM Center High-Performance Parallel I/O Libraries (PI) Alok Choudhary, (Co-I) Wei-Keng Liao Northwestern University In Collaboration with the SEA Group.
Current status and future work
Overview of big data tools
Presentation transcript:

The HDF Group Parallel HDF5 Developments 1 Copyright © 2010 The HDF Group. All Rights Reserved Quincey Koziol The HDF Group

2 Goal is to be invisible: get same performance with HDF5 as with MPI I/O Project with LBNL/NERSC to improve HDF5 performance on parallel applications: 6-12x performance improvements on various applications (so far) Parallel I/O in HDF5 Copyright © 2010 The HDF Group. All Rights Reserved

3 Up to 12GB/s to shared file (out of 15GB/s) on NERSCs franklin system: Parallel I/O In HDF5 Copyright © 2010 The HDF Group. All Rights Reserved

The HDF Group Recent Improvements to Parallel HDF5 4 Copyright © 2010 The HDF Group. All Rights Reserved

5 Reduce number of file truncation operations Distribute metadata I/O over all processes Detect same shape of selection in more cases, allowing optimized I/O path to be taken more often Many other, smaller, improvements to library algorithms for faster/better use of MPI Recent Parallel I/O Improvements Copyright © 2010 The HDF Group. All Rights Reserved

Reduced File Truncations HDF5 library was very conservative about truncating file when H5Fflush called. However, file truncation very expensive in parallel. Library modified to defer truncation until file closed. Copyright © 2010 The HDF Group. All Rights Reserved6

Distributed Metadata Writes HDF5 caches metadata internally, to improve both read and write performance Historically, process 0 writes all dirtied metadata to HDF5 file, while other processes wait Changed to distribute ranges of metadata within the file across all processes Results in ~10x improvement in I/O for Vorpal (see next slide) Copyright © 2010 The HDF Group. All Rights Reserved7

Dsitributed Metadata Writes I/O Trace Before Changes Note long sequence of I/O from process 0 I/O Trace After Changes Note distribution of I/O across all processes, taking much less time Copyright © 2010 The HDF Group. All Rights Reserved8

Improved Selection Matching When HDF5 performs I/O between regions in memory and the file, it compares the regions to see if the applications buffer can be directly used for I/O Historically, this algorithm couldnt detect that a region with the same shape, but embedded in arrays of different dimensionality were the same For example, a 10x10 region in a 2-D array should compare equal to the equivalent 1x10x10 region in a 3-D array Changed to detect same shaped region in arbitrary source and destination buffer array dimensions, allowing I/O from applications buffer in more circumstances. Copyright © 2010 The HDF Group. All Rights Reserved9

Improved Selection Matching Change resulted in ~20x I/O performance improvement when reading 1-D buffer from 2-D file dataset From ~5-7 seconds (or worse) to ~ seconds, on a variety of machine architectures (Linux: amani, hdfdap, jam; Solaris: linew) Copyright © 2010 The HDF Group. All Rights Reserved10

The HDF Group Upcoming Improvements to Parallel HDF5 11 Copyright © 2010 The HDF Group. All Rights Reserved

12 HPC environments typically have unusual, possibly even unique, computing, network and storage configurations. The HDF5 distribution should provide easy to use interfaces that ease scientists and developers use of these platforms: Tune and adapt to the underlying parallel file system. New high-­level API routines that wrap existing HDF5 functionality in a way that is easier for HPC application developers to use and help them move applications from one HPC environment to another. RFC: API/H5HPC_RFC pdfhttp:// API/H5HPC_RFC pdf High-Level HPC API for HDF5 Copyright © 2010 The HDF Group. All Rights Reserved

13 File System Tuning: Automatic file system tuning Pass file system tuning info to HDF5 library Convenience Routines: Macro routines Encapsulate common parallel I/O operations E.g. - create a dataset and write a different hyperslab from each process, etc. Extended routines Provide special parallel I/O operations not available in main HDF5 API Examples: Group collective I/O operations Collective raw data I/O on multiple datasets Collective multiple object manipulation Optimized collective object operations High-Level HPC API for HDF5 – API Overview Copyright © 2010 The HDF Group. All Rights Reserved

The HDF Group Parallel HDF5 in the Future 14 Copyright © 2010 The HDF Group. All Rights Reserved

15 DOE Exascale FOA w/LBNL & PNNL Proposal Funded Exascale-focused enhancements to HDF5 LLNL Support & Development Contract Performance, support and medium-term focused development DOE Exascale FOA w/ANL and ORNL Proposal Funded Research on alternate file formats for Exascale I/O LBNL Development Contract Performance and short-term focus HPC Funding in 2010 and Beyond Copyright © 2010 The HDF Group. All Rights Reserved

16 Library Enhancements Proposed: Remove collective metadata modification restriction Append-only mode, targeting restart files Embarrassingly parallel mode, for decoupled applications Overlapping compute & I/O, with asynchronous I/O Auto-tuning to underlying parallel file system Improve resiliency of changes to HDF5 files Bring FastBit indexing of HDF5 files into mainstream use for queries during data analysis and visualization Virtual file driver enhancements Improved Support: Parallel I/O performance tracking, testing and tuning Future Parallel I/O Improvements Copyright © 2010 The HDF Group. All Rights Reserved

17 Contract w/LLNL to do: Scalable I/O performance tracking, testing and tuning Virtual file driver enhancements HPC Specific fast­tracking Parallel interface enhancements Exploratory design development User support and routine maintenance Future Parallel I/O Improvements Copyright © 2010 The HDF Group. All Rights Reserved

The HDF Group Performance Hints for Using Parallel HDF5 18 Copyright © 2010 The HDF Group. All Rights Reserved

19 Pass along MPI Info hints to file open: H5Pset_fapl_mpio Use MPI-POSIX file driver to access file: H5Pset_fapl_mpiposix Align objects in HDF5 file: H5Pset_alignment Use collective mode when performing I/O on datasets: H5Pset_dxpl_mpio before H5Dwrite/H5Dread Avoid datatype conversions: make memory and file datatypes the same Advanced: explicitly manage metadata flush operations with H5Fset_mdc_config Hints for Using Parallel HDF5 Copyright © 2010 The HDF Group. All Rights Reserved