Full HTML for

Basic foilset Integration of Scientific and Technical Computing with Enterprise Information Systems

Given by Geoffrey C. Fox at RCI North American Annual Member Executive Conference October 13-15 98 on October 15 98. Foils prepared October 16 98
Outside Index Summary of Material

We discuss a multi tier (Gateway) approach to HPCC and how this allows appropriate use of ideas from both high performance and enterprise Intranet communities.
We introduce the Pragmatic Object Web (POW)
We give examples of using WebFlow and JWORB with Globus and show this gives natural integration of CORBA into HPCC
The use of HLA/RTI as control of a bunch of workstations is recommended
We describe Java Grande as an effort to bring better software tools to HPCC

Table of Contents for full HTML of Integration of Scientific and Technical Computing with Enterprise Information Systems

 Denote Foils where Image Critical
Denote Foils where Image has important information
Denote Foils where HTML is sufficient
1 Integration of Scientific and Technical Computing with Enterprise Information Systems
2 Abstract of RCI HPCC/Enterprise Systems Integration Presentation
3 Optimistic Scenario
4 The HPCC Dilemma and its Solution
5 What is Commodity Software
6 Multi-Tier Client Server Service
7 Abstraction of Multi Tier Architectures
8 Multi-Server Middle Tier Model
9 Exploiting Multi-Tier Commodity Software Model
10 Multi-Server Gateway Tier
11 More Features of the Three Tiers
12 Implications of Multi Tier Architectures
13 Synergy of Parallel Computing and Web Internetics as Unifying Principle
14 Hardware Architecture for multi Tier Systems
15 What are General Capabilities in Gateway Tier?
16 What Particular Programs could run in Gateway Tier?
17 HPcc Prototype WebFlow and JWORB over GLOBUS
18 Gateway Building Blocks: JWORB WebFlow on GLOBUS
19 WebFlow WaveFilter Module
20 WebFlow + Globus Functional Architecture
21 WebFlow as front end for Globus in Alliance Quantum Chemistry Simulations
22 WebFlow over Globus for NCSA Alliance Quantum Chemistry Application View
23 Next Steps for HPcc using JavaBeans
24 Summary of NPAC's JWORB natural Building Block of the Gateway
25 IIOP Performance for Java ORB's
26 General HPcc Software Strategy
27 Java and multi-tier Grande Systems
28 What is Java Grande?
29 Java Grande Process: Approach and Activities
30 Why is Java Grande Worth Looking at?
31 What is the Competition?
32 Why could Java succeed where Fortran and C++ failed?
33 What is Goal of Java Grande Forum?
34 Two types of Things we are doing
35 Activities of the Java Grande Forum I
36 Java Rules for Floating Point
37 JGF Proposed FP Execution Modes
38 Activities of the Java Grande Forum II
39 Java and Parallelism?
40 Where are we now?
41 What should you do as a Java Grande believer?

Outside Index Summary of Material

HTML version of Basic Foils prepared October 16 98

Foil 1 Integration of Scientific and Technical Computing with Enterprise Information Systems

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Invited Talk at RCI Executive Conference October 15, 98
Geoffrey Fox
Northeast Parallel Architectures Center
Syracuse University
111 College Place
Syracuse NY
gcf@npac.syr.edu
http://www.javagrande.org
http://www.npac.syr.edu/users/gcf/rcihpccoct98
HTML version of Basic Foils prepared October 16 98

Foil 2 Abstract of RCI HPCC/Enterprise Systems Integration Presentation

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
We discuss a multi tier (Gateway) approach to HPCC and how this allows appropriate use of ideas from both high performance and enterprise Intranet communities.
We introduce the Pragmatic Object Web (POW)
We give examples of using WebFlow and JWORB with Globus and show this gives natural integration of CORBA into HPCC
The use of HLA/RTI as control of a bunch of workstations is recommended
We describe Java Grande as an effort to bring better software tools to HPCC
HTML version of Basic Foils prepared October 16 98

Foil 3 Optimistic Scenario

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Java becomes a better Grande (HPCC) programming environment than Fortran and C++
High speed networks become pervasively deployed
Microsoft does not win -- some combination of IBM Sun Netscape (...) stops the advance of NT, COM, Internet Explorer, ChromeEffects ...
The Web remains a creative open environment
Supercomputers remain a small market but do not entirely disappear
Technical computing (on supercomputers and below) continues to be important but problems become complex with linked databases and linked multiple disciplines
HTML version of Basic Foils prepared October 16 98

Foil 4 The HPCC Dilemma and its Solution

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
HPCC has developed good research ideas but cannot implement them as solving computing's hardest problem with 1 percent of the funding
  • HPCC applications are very complex and use essentially all computer capabilities and also have synchronization and performance constraints from HPCC
We have learnt to use commodity hardware either
  • partially as in Origin 2000/SP2 with consumer CPU's but custom network or
  • fully as in PC cluster with fast ethernet/ATM
Let us do the same with software and design systems with maximum possible commodity software basis
HTML version of Basic Foils prepared October 16 98

Foil 5 What is Commodity Software

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
The world is building a wonderful distributed computing (information processing) environment using Web (dissemination) and distributed object (CORBA COM, Java, XML) technologies
This includes Java, Web-linked databases and the essential standards such as HTML(documents), VRML(3D objects), JDBC (Java database connectivity).
  • The standard interfaces are essential in that they allow modular (component based) software
We will "just" add high performance to this commodity distributed infrastructure
  • Respecting architecture of the object web, should allow us to naturally use improved software as it produced
The alternative strategy starts with HPCC technologies (such as MPI,HPF) and adds links to commodity world. This approach does not easily track evolution of commodity systems and so has large maintenance costs
HTML version of Basic Foils prepared October 16 98

Foil 6 Multi-Tier Client Server Service

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
1: Client Tier
JavaBean Enterprise JavaBean
Old and New Useful Backend Systems including MPP
3: Back-end Tier
Services
2: Middle Tier
Servers
HTML version of Basic Foils prepared October 16 98

Foil 7 Abstraction of Multi Tier Architectures

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
1: Client Tier running on PC or Workstations
2: Middle Tier -- a growing sea of (application) servers where most Intranet software activity is concentrated
  • Contains business logic
  • Modest Performance High Functionality
  • Software architecture reflects Macroscopic loosely coupled structure of problems
3: Back end Capabilities such as MPP's
  • Good use requires understanding microscopic tightly coupled parallel structure of applications
HTML version of Basic Foils prepared October 16 98

Foil 8 Multi-Server Middle Tier Model

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
HTML version of Basic Foils prepared October 16 98

Foil 9 Exploiting Multi-Tier Commodity Software Model

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Essential HPcc idea is consider a three tier model
  • Top tier is the client (in "Network Computer" based 4 tier architectures this becomes 2 tiers)
  • Second tier are servers coordinated by commodity technologies such as the Web and CORBA and communicating via HTTP(Web), IIOP(CORBA), RMI or custom Java sockets.
  • Use middle tier component/container model -- Enterprise Javabeans
  • Third tier are services such as databases, MPP's
Preserve the first two tiers as a high functionality commodity information processing system and confine HPCC to the third (lowest) tier.
  • MPI becomes the high performance "machine code" for message passing which you use if HTTP, IIOP or RMI have insufficient performance
HTML version of Basic Foils prepared October 16 98

Foil 10 Multi-Server Gateway Tier

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Database
Matrix Solver
Optimization Service
MPP
MPP
Parallel DB Proxy
NEOS Control Optimization
Origin 2000 Proxy
NetSolve Linear Alg. Server
IBM SP2 Proxy
Gateway Control
Agent-based Choice of Compute Engine
Multidisciplinary Control (WebFlow)
Data Analysis Server
HTML version of Basic Foils prepared October 16 98

Foil 11 More Features of the Three Tiers

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Middle Tier
Distributed (Pragmatic) Object Web
High Functionality Modest Performance
Java RMI/Jini
XML COM CORBA
Distributed Computing
The "Operating System" level of Computational Grids, Enterprise Information Systems etc.
Client Tier
Java GUI Seamless Computing Interfaces
CollaborationVisualization
Microsoft Netscape Browser battle makes unattractive investment
Backend Tier
Special Services such as HPCC
Instruments Traditional Databases
Can use Java as a Scientific Computing Language?
Parallel Computing
HTML version of Basic Foils prepared October 16 98

Foil 12 Implications of Multi Tier Architectures

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Hardware Architecture
Software Architecture and Philosophy
  • Build loosely coupled systems
Challenges and Opportunities Opened Up
What do we adopt of Commodity Tier
  • Software Infrastructure (Pragmatic Object Web) and architecture
  • Java Grande
What do we retain from HPCC
  • Some component tools such as MPI; most ideas; existing applications
Implications for Education
Note opposite progression from some other approaches:
HTML version of Basic Foils prepared October 16 98

Foil 13 Synergy of Parallel Computing and Web Internetics as Unifying Principle

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
The two forms of Large Scale Computing Scale Computer for Scale Users in Proportion Power User to number of computers
Parallel Commodity Distributed Computers Information Systems Technology <--------------- Internetics Technologies --------------->
Send your Staff to Online Course: Certificate in "Internet Applications"
aka Internetics -- Offered over Internet at 6 hours per week
1% market
99% of market driving
student interest and (Java) technologies
HTML version of Basic Foils prepared October 16 98

Foil 14 Hardware Architecture for multi Tier Systems

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Geographically
Distributed
Grandecomputer
Resources
Gateway System
hosting Seamless Access
Database, Collaboration
Visualization
and other Services
Geographically Distributed users
1
2
3
HTML version of Basic Foils prepared October 16 98

Foil 15 What are General Capabilities in Gateway Tier?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Seamless Interface -- an Enterprise JavaBean which processes input from user's Java Applet interface and maps user generic commands to those on specific machine
  • Uses agents to determine optimal execution platform
  • Accounting, Security, Compiling Interface, Seamless Tools Interface, global data and file system interface
Resource management of heterogeneous MPP backend (linked to seamless interface)
Database and Object Brokers; Network servers like Netsolve
Collaboration Servers including Habanero, Tango, Lotus Notes ...
Visualization Servers
"Business Logic" to map user data view (e.g. objects) to persistent store (e.g. Oracle database) and simulation engine (MPP) preferred format
HTML version of Basic Foils prepared October 16 98

Foil 16 What Particular Programs could run in Gateway Tier?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
The 90% of users who only need HPCC occasionally
Most of a Command and Control Application
Several FMS and IMT Applications
Some I/O Intensive applications
High value services with modest computational needs e.g. grid generation and other pre-processing, data manipulation and other post-processing
Video Servers for Training
Design and Planning Tools
"Glue" for Multidisciplinary Interactions
Control of Metacomputing applications
JINI Java Resource Registration and Discovery Service
HTML version of Basic Foils prepared October 16 98

Foil 17 HPcc Prototype WebFlow and JWORB over GLOBUS

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
NPAC Project illustrating multi tier Gateway incorporating Java Applet Program Composition at Client
DMSO HLA/RTI Compliance (FMS,IMT)
CORBA COM Java Web Interoperable Gateway GLOBUS High Performance Backend inherits Gateway Services: Collaboration, Distributed Objects, Databases, Visualization .....
HTML version of Basic Foils prepared October 16 98

Foil 18 Gateway Building Blocks: JWORB WebFlow on GLOBUS

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Client (Tier 1): Java Graph Editor for Webflow linked to interpreted debugger (DARP), Java Visualizer SciVis
  • In four tier model, these Java tools would run on Java Web Server with pure HTML client
Middle Tier 2: Network of Pragmatic Object Web JWORB or Java Servers running on Gateway Tier linking to CORBA COM and HTTP(Internet)
  • Java Servlet wraps non Java Application (Apache Java Server) or CORBA handles any language
Back-end Tier 3: Simple MPI or Globus (or similar) where available and where necessary
Next foils show
  • Pure Gateway Tier: WebFlow system with simple Java Image filters
  • Customized "database" solution for NCSA Alliance Grand Challenge in Quantum Monte Carlo
HTML version of Basic Foils prepared October 16 98

Foil 19 WebFlow WaveFilter Module

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Original Image
Output Image
Some of
Available Image Filters
CLIENT Visual DataFlow
Interface
SERVER Processes Filters
HTML version of Basic Foils prepared October 16 98

Foil 20 WebFlow + Globus Functional Architecture

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
HTML version of Basic Foils prepared October 16 98

Foil 21 WebFlow as front end for Globus in Alliance Quantum Chemistry Simulations

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Client Tier
IIOP High Functionality
Middle Tier
Future Globus
Globus
Future Parallel I/O
HTML version of Basic Foils prepared October 16 98

Foil 22 WebFlow over Globus for NCSA Alliance Quantum Chemistry Application View

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
HTML version of Basic Foils prepared October 16 98

Foil 23 Next Steps for HPcc using JavaBeans

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Currently WebFlow uses a Java Server and manipulates Java applications which can be front ends with native methods to Fortran C or C++ routines
Change Java Server to JWORB -- server integrating HTTP and IIOP (Web and CORBA)
Change Java Applications to JavaBeans and non-Java apps to CORBA objects
Change linkage in WebFlow to respect JavaBean event mechanism
Then we get HPComponentware
And using our multi-tier model high performance CORBA
WebFlow is HPCC version of a
Typical Visual Interface for JavaBeans
HTML version of Basic Foils prepared October 16 98

Foil 24 Summary of NPAC's JWORB natural Building Block of the Gateway

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
JWORB - Java Web Object Request Broker - multi-protocol middleware network server (HTTP + IIOP + DCE RPC + RMI transport)
Current prototype integrates HTTP and IIOP i.e. acts as Web Server and CORBA Broker
  • HTTP Services built in terms of CORBA services
  • Gives you immediately web interfaces to CORBA
  • CORBA supports applications in any language
Currently testing support of Microsoft COM
JWORB - our trial implementation of Pragmatic Object Web
First non DMSO implementation of RTI -- HLA (distributed event driven simulation) Runtime at 5% cost(!)
HTML version of Basic Foils prepared October 16 98

Foil 25 IIOP Performance for Java ORB's

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
JacORB
JWORB
ORBIX
RMI
Variable Size Integer Arrays
HTML version of Basic Foils prepared October 16 98

Foil 26 General HPcc Software Strategy

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
What do we keep from HPCC
  • Nearly all the ideas which were typically good
  • All technologies associated with decomposition and support of tightly coupled applications such as MPI
  • Components from toolkits like Globus
  • Existing applications
What do we take from Enterprise Systems
  • All coarse grain distributed computing technologies
  • Core software and distributed object infrastructure
  • Java as a potentially better language
HTML version of Basic Foils prepared October 16 98

Foil 27 Java and multi-tier Grande Systems

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Geographically
Distributed
Grandecomputer
Resources
Gateway System
hosting Seamless Access
Database, Collaboration
Visualization
and other Services
Geographically Distributed users
Java Applet Clients
Parallel or Sequential Java Grande Codes
Java Servers
Java wins today?
Major Commercial Activity
Java Grande needs to look at scaling etc.
Only Java Grande will promote?
HTML version of Basic Foils prepared October 16 98

Foil 28 What is Java Grande?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Use of Java for:
High Performance Network Computing
Scientific and Engineering Computation
(Distributed) Modeling and Simulation
Parallel and Distributed Computing
Data Intensive Computing
Communication and Computing Intensive Commercial and Academic Applications
HPCC Computational Grids ........
Very difficult to find a "conventional name" that doesn't get misunderstood by some community!
HTML version of Basic Foils prepared October 16 98

Foil 29 Java Grande Process: Approach and Activities

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
We have had several conferences with 50---- attendees
  • Syracuse December 96
  • Las Vegas June 97
  • Palo Alto February 98
  • Southampton (Europe) September 98
  • Next one just before JavaOne next year (March 99)
Topics of conference papers:
  • Applications; algorithms; benchmarking; compilers; Java-based programming tools; parallel computing (tightly coupled) and high performance distributed computing
"Spun off" Java Grande forum to promote needed community standards and activities.
Must be proactive because Grande computer market is perhaps 1% of total computing market -- not Sun's highest priority
HTML version of Basic Foils prepared October 16 98

Foil 30 Why is Java Grande Worth Looking at?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
The Java Language has several good design features
  • secure, safe (wrt bugs), object-oriented, familiar (to C C++ and even Fortran programmers)
Java has a very good set of libraries covering everything from commerce, multimedia, images to math functions (under development at http://math.nist.gov/javanumerics)
Java has best available electronic and paper training and support resources
Java is rapidly getting best integrated program development environments
Java naturally integrated with network and universal machine supports potentially powerful "write once-run anywhere (badly)" model
There is a large and growing trained labor force
Can we exploit this in Grande Computing / computational science?
HTML version of Basic Foils prepared October 16 98

Foil 31 What is the Competition?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
So existing Grande codes are written in Fortran C and C++ with a clearly unattractive and comparatively unproductive programming environment
These current languages and tools are sufficient but does not seem likely that can build much better environments around them
  • Fortran77 has excellent compilers, good user base but will not be taught broadly and clearly limited in capabilities; in particular not object oriented
  • Fortran90 and HPF will not "make it" (reasons not important)
  • Subsume discussion of C under that of C++
Five years ago, it looked as though C++ could become language of choice (perhaps with Fortran as inner core) but this appears stalled
  • The language is complex and splintered with no agreement on Grande standards -- partly because use in Grande applications is too small to motivate standards and partly due to culture
  • Java halted C++"bandwagon"
So there is no competition -- Java is currently our only hope
HTML version of Basic Foils prepared October 16 98

Foil 32 Why could Java succeed where Fortran and C++ failed?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
It has some natural advantages due its internet base with threads and distributed computing built in
It is a young language and we can take steps now to avoid unproductive proliferation of libraries and parallel constructs
  • We could be third (Fortran, C++, Java) time lucky
It could have expressivity and object oriented advantages of C++ combined with performance levels of C and Fortran
It can use its clear GUI advantages as an entrée into other aspects of Grande programming
HTML version of Basic Foils prepared October 16 98

Foil 33 What is Goal of Java Grande Forum?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Java has potential to be a better environment for "Grande application development" than any previous languages such as Fortran and C++
The Forum Goal is to develop community consensus and recommendations for either changes to Java or establishment of standards (frameworks) for "Grande" libraries and services
These Language changes or frameworks are designed to realize "best ever Grande programming environment"
First Meeting Mar 1 Palo Alto at Java 98 -- 200 Attendees set Agenda -- 30 permanent people and further meetings May 9-10, Aug 6-7
Public Discussion SC98 Orlando November 13 (3 hour panel)
http://www.npac.syr.edu/projects/javaforcse
http://www.javagrande.org
HTML version of Basic Foils prepared October 16 98

Foil 34 Two types of Things we are doing

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
1) Most important in the near term -- encourage Sun to make a few key changes in Java to allow it to be a complete efficient Grande Programming Language
  • floating point, arrays, complex etc.
2) As a community, recognize that sometimes standards are more appropriate than creativity and pool results of experiments to produce a Java Grande framework covering libraries and computer access
  • Fiscally important fields such as databases, have established such standards -- we should follow their example
1) requires us to work with the computing mainstream -- 2) is internal to community
HTML version of Basic Foils prepared October 16 98

Foil 35 Activities of the Java Grande Forum I

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Two major working groups promoting standards and community actions
Numerics: Java as a language for mathematics led by Ron Boisvert and Roldan Pozo from NIST
  • Changes in Java controversial handling of floating point which currently has goal of reproducible results but this leads to non optimal accuracy
  • Addition of Complex types or classes
  • Lightweight classes and Operator overloading -- enables implementation of complex as a class
  • "Fortran rectangular multidimensional arrays" -- Java naturally has "arrays of arrays"
  • High quality math libraries with agreed interfaces -- FFT, Matrices, Transcendental functions
HTML version of Basic Foils prepared October 16 98

Foil 36 Java Rules for Floating Point

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Java is particularly slow on floating point -- why?
  • Immature Compiler and library Technology
  • Does not Exploit optimally hardware on PowerPC and Intel platforms
    • Current Java rules optimized for SPARC architecture .....
  • Does not use well known compiler optimizations for blocking, unrolling, reorganizing expressions
Current Java rules imply that if careful (avoid indeterminate threads and use of JNI) Java not only runs everywhere but gets identical results everywhere
Need to study Reproducibility -- Performance Trade-offs
  • Maybe people have emphasized performance too much but you are not going to change their minds quickly -- should be tolerant of all perceived needs
  • Note current Java is both lower precision AND lower performance than other languages
HTML version of Basic Foils prepared October 16 98

Foil 37 JGF Proposed FP Execution Modes

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Language Chip Architecture
Modifier x86 PowerPC SPARC
strictfp Java 1.0 (no double Java 1.0 Java 1.0
rounding on underflow) (no fused mac)
default i.e. larger exponent range fused mac can Java 1.0
no modifier allowed be used
associativefp many optimizations allowed on all platforms
Current default is essentially strictfp although not always implemented precisely
mac is fused hardware multiply-add (a*b+c)
strictfp is a little more than 2 times slower than default on INTEL x86 for a vector dot product
HTML version of Basic Foils prepared October 16 98

Foil 38 Activities of the Java Grande Forum II

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Distributed and Parallel Computing led by Dennis Gannon and Denis Caromel (INRIA, France)
  • Performance of RMI (Attractive Java distributed object model - "remote method invocation")
  • Performance of Java runtime (the virtual machine VM) with lots of threads, I/O, memory use
  • Parallel Computing interfaces including Java MPI binding
  • Development of universal (Condor, Globus, Legion UNICORE WebSubmit ..) Java interface to computing resources -- enables seamless computing (easier than metacomputing!)
  • Very succesful seamless computing meeting at Argonne October 98
Development of Grande Application benchmarks
HTML version of Basic Foils prepared October 16 98

Foil 39 Java and Parallelism?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
The Web integration of Java gives it excellent "network" classes and support for message passing.
Thus "Java plus message passing" form of parallel computing is actually somewhat easier than in Fortran or C.
Coarse grain parallelism very natural in Java and we have illustrated this with WebFlow
"Data Parallel" languages features are NOT in Java and have to be added extending ideas from HPF and HPC++ etc
  • e.g. NPAC's HPJava translates to Java+Messaging just as HPF translates to Fortran plus message passing
Java has built in "threads" and a given Java Program can run multiple threads at a time
  • In Web use, allows one to process Image in one thread, HTML page in another etc.
  • threads can be used to do more general parallel computing but only on shared memory computers
HTML version of Basic Foils prepared October 16 98

Foil 40 Where are we now?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Both working groups have made substantial progress
  • Numerics working group has proposals essentially ready for Sun
  • Concurrency working group will propose modest RMI changes
We are initiating Community actions
  • Help us collect Java Grande benchmarks
  • Work with community on standard classes and libraries
  • Participate in seamless computing framework
  • Stress Java and Java runtime (the VM) with large applications -- where are performance problems?
Join us at SC98 November 13
Note European involvement has been excellent so far
  • Caromel, Getov, Phillipsen
  • Edinburgh and NAG
HTML version of Basic Foils prepared October 16 98

Foil 41 What should you do as a Java Grande believer?

From Integration of Scientific and Technical Computing with Enterprise Information Systems RCI North American Annual Member Executive Conference October 13-15 98 -- October 15 98. *
Full HTML Index
Don't need to rewrite existing codes in Java!
Rather use Java freely at client and "gateway" tier
Wrap existing codes as CORBA or Java distributed objects
Conduct suitable experiments in using Java in complete Grande applications
Make certain your interests are represented in Java Grande Forum
Retrain your staff in Java Web and distributed object technologies
Put "High Performance Grande Forum compliant" Java support into your RFP's for hardware and software
© Northeast Parallel Architectures Center, Syracuse University, npac@npac.syr.edu

If you have any comments about this server, send e-mail to webmaster@npac.syr.edu.

Page produced by wwwfoil on Sat Nov 28 1998