SlideShare a Scribd company logo

Packed Objects: Fast Talking Java Meets Native Code - Steve Poole (IBM)

J
jaxLondonConference

The presentation by Steve Poole focuses on IBM's development of 'packed objects' or 'compact off-heap structures in the Java language', aimed at enhancing interoperability between Java and non-Java applications. Packed objects allow for efficient data sharing with minimal overhead, improving performance over JNI while maintaining a simpler programming model. The talk includes demonstrations of this technology's benefits, particularly in graphics processing using SDL and discusses the implications for future Java development.

TechnologyEducation
1 of 54
Downloaded 16 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
Steve Poole IBM Fast Talking Java Meets Native Code
Important Disclaimers THE INFORMATION CONTAINED IN THIS PRESENTATION IS PROVIDED FOR INFORMATIONAL PURPOSES ONLY. WHILST EFFORTS WERE MADE TO VERIFY THE COMPLETENESS AND ACCURACY OF THE INFORMATION CONTAINED IN THIS PRESENTATION, IT IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED. ALL PERFORMANCE DATA INCLUDED IN THIS PRESENTATION HAVE BEEN GATHERED IN A CONTROLLED ENVIRONMENT. YOUR OWN TEST RESULTS MAY VARY BASED ON HARDWARE, SOFTWARE OR INFRASTRUCTURE DIFFERENCES. ALL DATA INCLUDED IN THIS PRESENTATION ARE MEANT TO BE USED ONLY AS A GUIDE. IN ADDITION, THE INFORMATION CONTAINED IN THIS PRESENTATION IS BASED ON IBM’S CURRENT PRODUCT PLANS AND STRATEGY, WHICH ARE SUBJECT TO CHANGE BY IBM, WITHOUT NOTICE. IBM AND ITS AFFILIATED COMPANIES SHALL NOT BE RESPONSIBLE FOR ANY DAMAGES ARISING OUT OF THE USE OF, OR OTHERWISE RELATED TO, THIS PRESENTATION OR ANY OTHER DOCUMENTATION. NOTHING CONTAINED IN THIS PRESENTATION IS INTENDED TO, OR SHALL HAVE THE EFFECT OF: - CREATING ANY WARRANT OR REPRESENTATION FROM IBM, ITS AFFILIATED COMPANIES OR ITS OR THEIR SUPPLIERS AND/OR LICENSORS
Steve Poole ! Works at IBM’s Hursley Laboratory in the UK Involved in IBM Java VM development since before Java was 1 Currently leading IBM’s OpenJDK technical engagement
What this talk is about “Packed Objects” aka “Compact Off-Heap Structures in the Java Language” ! This technology is being developed to help address important pressures on Java ! This talk focuses on how to improve Java interop to Non Java applications
You’ll learn “Compact Off-Heap Structures in the Java Language” 1. An overview of this technology 2. Why we need your input and support 3. How to get started with using this technology yourself
Part 1 Native interoperability Why it’s so important C
Interoperability - why is it important? Java is built on talking native ! Operating Systems Processor Architectures File Systems User interfaces Network Interfaces ! Java keeps the WORA monsters away ! !
WORA - the components JDK Java code JVM JIT native Class Library Native code Your JNI code Your Java application
2013 - Not so Java-centric JVM JIT Class Library native Your Native code JDK Java code Your JNI code Your Java application
2013 - clusters and new processors non Java Runtime High speed memory channels (RDMA etc) Specialised Processors (GPUs etc)
2013 - Can we survive with JNI 1.1? JIT ? Other runtime JVM Class Library Your Native code JDK Java code Your JNI code Your Java application native High speed memory channels (RDMA etc) Specialised Processors (GPUs etc) ? ?
2013 - we can’t survive with JNI 1.1 • 1997 - JNI 1.1 • Designed for independence • Version independence • Platform Independence • VM Independence • Java at the centre of the world • Not performance focused
Packed Objects to the rescue • What we need is • To be able to share data between Java and other runtimes with minimal overhead • To not require a Java centric data view • and share data off-heap too! • Can we do that? • Time for a demo.
A simple demo • • Screensaver style Rectangles moved around the screen, bounce off the sides…
A simple demo Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! RGB colour elements ! typedef struct { int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! } RECT; !15 position adjustment vectors SDL structure contains location , height and width float version of location - makes it easier to handle different h/w speeds
A simple demo Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !16
More than one demo Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !17 Frames per second number of JNI calls per frame
Demo 1 - all C Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !18
Results Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !19 Frames per second 20 number of JNI calls per frame 0
Demo 2- Mixed JNI and C Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !20
Results Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !21 Frames per second 20 number of JNI calls per frame 0
er frame) Results Demo Frames per second Standard - all C code calling the SDL graphics routines 20 Mixed C and JNI. Java updates the location of each element 11 number of JNI calls per frame 0 200000 (~4MB per frame) Mixed C with PackedObjects Java updates the location of each element 200 000 JNI calls per frame. Copying data, validating data everytime. AND ensuring there is no chance of optimization by the JIT. !22
Demo 2- Details Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !23
Demo 2- Details Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! for every rectangle… C function calls Java method via JNI passes in x,y, vx and vy (by copy) Java method updates the rectangles new position Java native calls native method passing in new x,y, vx and vy values (by copy) C native method returns Java method returns C function returns !24
Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !25
er frame) Results Demo Frames per second Standard - all C code calling the SDL graphics routines 20 Mixed C and JNI. Java updates the location of each element 11 Mixed C with PackedObjects Java updates the location of each element !26 number of JNI calls per frame 0 200000 (~4MB per frame)
er frame) Results Demo Frames per second number of JNI calls per frame Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element 11 200000 (~4MB per frame) Mixed C with PackedObjects Java updates the location of each element !27 20 18 1 (4 bytes per frame) 0
Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !28
Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! once per frame.. Java method updates each rectangle’s position directly ! } !29 for(Rectangle r: rectangles) { r.x=r.x+r.vx; r.y=r.y+r.vy; …
Today Java only speaks Java • Getting data into and out of Java has always required some form of marshaling or serialization process ! • Interaction with native data structures in memory is particularly problematic • JNI is the slowest but safest – But you need good C / C++ knowledge • Unsafe and NIO are faster but more challenging to use – They both have their own programming ‘model’ • If the Java side is in control of storage layouts it’s easier • When mapping existing native structures it’s much, much, much more difficult ! • Ironically the JVM is completely aware of native data formats
Tomorrow, with Packed Objects.. Java will let you reference data stored like this ! typedef struct { int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! } RECT; C structure
Tomorrow, with Packed Objects.. as if it looked like this ! ! typedef struct { public class Rectangle { ! int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } RECT; C structure } Java structure
Tomorrow, with Packed Objects.. ! ! typedef struct { public class Rectangle { ! int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; without copying ! } RECT; C structure } Java structure
Tomorrow, with Packed Objects.. ! !! ! typedef struct { ! int red,green,blue; ! float vx,vy; ! SDL_Rect rec; ! float x,y; !! ! } RECT; ! ! ‘C’ runtime And off ! ! ! public class Rectangle { ! ! ! int red,green,blue; ! float vx,xy; ! heap SDLRect rec; ! float x,y; ! ! ! } ! ! Java heap
Part 2 Under the covers http://www.flickr.com/photos/mwichary/
Under the covers ! public class Rectangle { Let’s look at these classes from a JVM point of view.. ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } public class SDLRect { short x,y int width,height }
Under the covers ! ! ! ! ! ! ! ! ! ! ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap ! public class Rectangle { ! short x short y int width int height int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } public class SDLRect { short x,y int width,height } Java structure !37 J
! ! ! ! ! ! ! ! ! ! ! Questions ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !38 1 - How does a JVM ‘know’ the class of these objects? ! ! short x short y int width int height 2 - What are the data format rules for these types? ! ! 3 - When synchonizing on a object - where is the monitor stored?
! ! ! ! ! ! ! ! ! ! ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !39 Answers class ptr monitor ! class ptr monitor short x short y int width int height 1 - How does a JVM ‘know’ the class of these objects? Additional internal ptr ! 2 - What are the data format rules for these types? JVM impl specific (including ordering) 3 - When synchonizing on a object - where is the monitor stored? Additional internal field
How to add the metadata ! ! ! ! ! ! ! ! ! ! ! class ptr monitor int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !40 class ptr monitor short x short y int width int height ! ! ! ! ! ! ! ! ! ! ! [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime No space No space for the extra for the extra Java fields -Java fields what do we what do we do? do?
Under the covers ! ! ! ! ! ! ! ! ! ! ! Java heap !41 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Under the covers ! ! ! ! ! ! ! ! ! ! ! Java heap ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height } Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); !42
Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” Java heap !43 ! ! ! ! ! ! Constructor ! was not ! called! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” Java heap SDLRect s=r.rect; ! !44 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr Java heap SDLRect s=r.rect; ! !45 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr Java heap SDLRect s=r.rect; SDLRect t=r.rect; !46 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr “t” class ptr monitor data ptr Java heap SDLRect s=r.rect; SDLRect t=r.rect; !47 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
Part 3 Truth and Consequences
Design considerations • Off heap Packed Objects are considered to already exist so constructors are not called. • Since there is limited opportunity to store the vital metadata it is constructed when needed. • The assumption is that creating new objects when accessing embedded data is confusing and dangerous. • How does ‘==‘ work? • What about synchronization? • The idea of ‘identityless’ objects is being discussed. • Something that PackedObjects has in common with ValueTypes • Could you handle Objects of this style?
The way forward? • This approach offers Java a way forward in the multi environment world • Almost zero cost data interop will benefit Java when exploiting new processors and memory systems. • There are additional use cases that ‘fall out’ from this design. • Low cost r/w of files • Instant application startup • Basic unmanaged memory capabilities • Under the covers heap usage benefits through better control for ‘locality of reference’
Is this the right or only way? • IBM and Oracle have been discussing these sorts of concepts for some time • We need your input. How important are these use cases to you? What’s missing, not needed? What do you hate? • We are trying to keep clear of syntax discussions at this point - it’s too early and just distracting.
You too can use Packed Objects • Google for “IBM Java 8 Beta”
Hard hats must be worn… The current prototype implementation surfaces a lot of the internals that will eventually just disappear This approach makes it easier for us to quickly revise the design but means users of the prototype have to work harder. ! AND ! it avoids syntax wars :-) picture courtesy http://www.flickr.com/people/karen_od/
Wrap up -hopefully you now: “Compact Off-Heap Structures in the Java Language” 1. Have a basic understanding of this technology 2. Know why we need your input and support 3. Know how to get started with using this technology Thank You

More Related Content

What's hot (20)

KEY
Everything I Ever Learned About JVM Performance Tuning @Twitter
Attila Szegedi
 
PDF
Spark as the Gateway Drug to Typed Functional Programming: Spark Summit East ...
Spark Summit
 
PDF
High Performance Python on Apache Spark
Wes McKinney
 
PPTX
Deep Learning and Recurrent Neural Networks in the Enterprise
Josh Patterson
 
PPTX
Inferno Scalable Deep Learning on Spark
DataWorks Summit/Hadoop Summit
 
PDF
Deploying Machine Learning Models to Production
Anass Bensrhir - Senior Data Scientist
 
PPT
An Introduction to JVM Internals and Garbage Collection in Java
Abhishek Asthana
 
PPTX
From I/O To RAM
Baruch Sadogursky
 
PDF
The Future of Real-Time in Spark
Reynold Xin
 
PPT
Persisting Your Objects In The Database World @ AlphaCSP Professional OSS Con...
Baruch Sadogursky
 
ODP
Quick introduction to Java Garbage Collector (JVM GC)
Marcos García
 
PPTX
Deep Learning Intro - Georgia Tech - CSE6242 - March 2015
Josh Patterson
 
PDF
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
NETFest
 
PDF
JavaOne 2013: Memory Efficient Java
Chris Bailey
 
PDF
Making Sense of Spark Performance-(Kay Ousterhout, UC Berkeley)
Spark Summit
 
PDF
Deep Dive: Memory Management in Apache Spark
Databricks
 
PDF
Super Sizing Youtube with Python
didip
 
PPTX
Georgia Tech cse6242 - Intro to Deep Learning and DL4J
Josh Patterson
 
PPT
Java Performance Monitoring & Tuning
Muhammed Shakir
 
PDF
Remote Log Analytics Using DDS, ELK, and RxJS
Sumant Tambe
 
Everything I Ever Learned About JVM Performance Tuning @Twitter
Attila Szegedi
 
Spark as the Gateway Drug to Typed Functional Programming: Spark Summit East ...
Spark Summit
 
High Performance Python on Apache Spark
Wes McKinney
 
Deep Learning and Recurrent Neural Networks in the Enterprise
Josh Patterson
 
Inferno Scalable Deep Learning on Spark
DataWorks Summit/Hadoop Summit
 
Deploying Machine Learning Models to Production
Anass Bensrhir - Senior Data Scientist
 
An Introduction to JVM Internals and Garbage Collection in Java
Abhishek Asthana
 
From I/O To RAM
Baruch Sadogursky
 
The Future of Real-Time in Spark
Reynold Xin
 
Persisting Your Objects In The Database World @ AlphaCSP Professional OSS Con...
Baruch Sadogursky
 
Quick introduction to Java Garbage Collector (JVM GC)
Marcos García
 
Deep Learning Intro - Georgia Tech - CSE6242 - March 2015
Josh Patterson
 
.NET Fest 2018. Maarten Balliauw. Let’s refresh our memory! Memory management...
NETFest
 
JavaOne 2013: Memory Efficient Java
Chris Bailey
 
Making Sense of Spark Performance-(Kay Ousterhout, UC Berkeley)
Spark Summit
 
Deep Dive: Memory Management in Apache Spark
Databricks
 
Super Sizing Youtube with Python
didip
 
Georgia Tech cse6242 - Intro to Deep Learning and DL4J
Josh Patterson
 
Java Performance Monitoring & Tuning
Muhammed Shakir
 
Remote Log Analytics Using DDS, ELK, and RxJS
Sumant Tambe
 

Viewers also liked (20)

PPTX
Legal and ethical considerations redone
Nicole174
 
PDF
Garbage Collection: the Useful Parts - Martijn Verburg & Dr John Oliver (jCla...
jaxLondonConference
 
PDF
Scaling Scala to the database - Stefan Zeiger (Typesafe)
jaxLondonConference
 
PDF
Databases and agile development - Dwight Merriman (MongoDB)
jaxLondonConference
 
PDF
How Hailo fuels its growth using NoSQL storage and analytics - Dave Gardner (...
jaxLondonConference
 
PPTX
Bringing your app to the web with Dart - Chris Buckett (Entity Group)
jaxLondonConference
 
PPT
How Java got its Mojo Back - James Governor (Redmonk)
jaxLondonConference
 
PDF
Introducing Vert.x 2.0 - Taking polyglot application development to the next ...
jaxLondonConference
 
PDF
Designing and Building a Graph Database Application - Ian Robinson (Neo Techn...
jaxLondonConference
 
PPTX
Interactive media applications
Nicole174
 
PDF
Big Events, Mob Scale - Darach Ennis (Push Technology)
jaxLondonConference
 
PDF
What makes Groovy Groovy - Guillaume Laforge (Pivotal)
jaxLondonConference
 
PDF
Are Hypermedia APIs Just Hype? - Aaron Phethean (Temenos) & Daniel Feist (Mul...
jaxLondonConference
 
PDF
Streams and Things - Darach Ennis (Ubiquiti Networks)
jaxLondonConference
 
PPTX
Why other ppl_dont_get_it
jaxLondonConference
 
PDF
Big data from the LHC commissioning: practical lessons from big science - Sim...
jaxLondonConference
 
PDF
How Windows 10 will change the way we use devices
Commelius Solutions
 
PPTX
45 second video proposal
Nicole174
 
PDF
The state of the art biorepository at ILRI
Absolomon Kihara
 
PDF
Lambda Expressions: Myths and Mistakes - Richard Warburton (jClarity)
jaxLondonConference
 
Legal and ethical considerations redone
Nicole174
 
Garbage Collection: the Useful Parts - Martijn Verburg & Dr John Oliver (jCla...
jaxLondonConference
 
Scaling Scala to the database - Stefan Zeiger (Typesafe)
jaxLondonConference
 
Databases and agile development - Dwight Merriman (MongoDB)
jaxLondonConference
 
How Hailo fuels its growth using NoSQL storage and analytics - Dave Gardner (...
jaxLondonConference
 
Bringing your app to the web with Dart - Chris Buckett (Entity Group)
jaxLondonConference
 
How Java got its Mojo Back - James Governor (Redmonk)
jaxLondonConference
 
Introducing Vert.x 2.0 - Taking polyglot application development to the next ...
jaxLondonConference
 
Designing and Building a Graph Database Application - Ian Robinson (Neo Techn...
jaxLondonConference
 
Interactive media applications
Nicole174
 
Big Events, Mob Scale - Darach Ennis (Push Technology)
jaxLondonConference
 
What makes Groovy Groovy - Guillaume Laforge (Pivotal)
jaxLondonConference
 
Are Hypermedia APIs Just Hype? - Aaron Phethean (Temenos) & Daniel Feist (Mul...
jaxLondonConference
 
Streams and Things - Darach Ennis (Ubiquiti Networks)
jaxLondonConference
 
Why other ppl_dont_get_it
jaxLondonConference
 
Big data from the LHC commissioning: practical lessons from big science - Sim...
jaxLondonConference
 
How Windows 10 will change the way we use devices
Commelius Solutions
 
45 second video proposal
Nicole174
 
The state of the art biorepository at ILRI
Absolomon Kihara
 
Lambda Expressions: Myths and Mistakes - Richard Warburton (jClarity)
jaxLondonConference
 
Ad

Similar to Packed Objects: Fast Talking Java Meets Native Code - Steve Poole (IBM) (20)

PDF
NDK Primer (Wearable DevCon 2014)
Ron Munitz
 
PDF
NASAfinalPaper
Danish Vaid
 
PDF
NDK Primer (AnDevCon Boston 2014)
Ron Munitz
 
PDF
Jvm internals
Luiz Fernando Teston
 
PDF
IBM Java PackedObjects
Marcel Mitran
 
PDF
JNA - Let's C what it's worth
Idan Sheinberg
 
PDF
Runtime Innovation - Nextgen Ninja Hacking of the JVM, by Ryan Sciampacone
ZeroTurnaround
 
PPTX
Java on arm theory, applications, and workloads [dev5048]
Aleksei Voitylov
 
PPTX
GOTO Night with Charles Nutter Slides
Alexandra Masterson
 
PPT
Android JNI
Siva Ramakrishna kv
 
PDF
JavaOne 2013: Introduction to PackedObjects
Ryan Sciampacone
 
PDF
Ola Bini Evolving The Java Platform
deimos
 
PPTX
Simple insites into JVM
Ramakanth Tarimala
 
PDF
Understanding the Dalvik bytecode with the Dedexer tool
Gabor Paller
 
PPTX
Java 7 - New Features - by Mihail Stoynov and Svetlin Nakov
Svetlin Nakov
 
PDF
Android and cpp
Joan Puig Sanz
 
PPTX
Intel JIT Talk
iamdvander
 
PDF
JVM Under the Hood
Serkan Özal
 
PPT
What is Java Technology (An introduction with comparision of .net coding)
Shaharyar khan
 
PPTX
Millions quotes per second in pure java
Roman Elizarov
 
NDK Primer (Wearable DevCon 2014)
Ron Munitz
 
NASAfinalPaper
Danish Vaid
 
NDK Primer (AnDevCon Boston 2014)
Ron Munitz
 
Jvm internals
Luiz Fernando Teston
 
IBM Java PackedObjects
Marcel Mitran
 
JNA - Let's C what it's worth
Idan Sheinberg
 
Runtime Innovation - Nextgen Ninja Hacking of the JVM, by Ryan Sciampacone
ZeroTurnaround
 
Java on arm theory, applications, and workloads [dev5048]
Aleksei Voitylov
 
GOTO Night with Charles Nutter Slides
Alexandra Masterson
 
Android JNI
Siva Ramakrishna kv
 
JavaOne 2013: Introduction to PackedObjects
Ryan Sciampacone
 
Ola Bini Evolving The Java Platform
deimos
 
Simple insites into JVM
Ramakanth Tarimala
 
Understanding the Dalvik bytecode with the Dedexer tool
Gabor Paller
 
Java 7 - New Features - by Mihail Stoynov and Svetlin Nakov
Svetlin Nakov
 
Android and cpp
Joan Puig Sanz
 
Intel JIT Talk
iamdvander
 
JVM Under the Hood
Serkan Özal
 
What is Java Technology (An introduction with comparision of .net coding)
Shaharyar khan
 
Millions quotes per second in pure java
Roman Elizarov
 
Ad

More from jaxLondonConference (17)

PDF
Conflict Free Replicated Data-types in Eventually Consistent Systems - Joel J...
jaxLondonConference
 
PDF
JVM Support for Multitenant Applications - Steve Poole (IBM)
jaxLondonConference
 
PDF
What You Need to Know About Lambdas - Jamie Allen (Typesafe)
jaxLondonConference
 
PDF
Java Testing With Spock - Ken Sipe (Trexin Consulting)
jaxLondonConference
 
PDF
The Java Virtual Machine is Over - The Polyglot VM is here - Marcus Lagergren...
jaxLondonConference
 
PDF
Java EE 7 Platform: Boosting Productivity and Embracing HTML5 - Arun Gupta (R...
jaxLondonConference
 
PPT
Exploring the Talend unified Big Data toolset for sentiment analysis - Ben Br...
jaxLondonConference
 
PDF
The Curious Clojurist - Neal Ford (Thoughtworks)
jaxLondonConference
 
PPTX
TDD at scale - Mash Badar (UBS)
jaxLondonConference
 
PDF
Run Your Java Code on Cloud Foundry - Andy Piper (Pivotal)
jaxLondonConference
 
PDF
Put your Java apps to sleep? Find out how - John Matthew Holt (Waratek)
jaxLondonConference
 
PPTX
Project Lambda: Functional Programming Constructs in Java - Simon Ritter (Ora...
jaxLondonConference
 
PPTX
Do You Like Coffee with Your dessert? Java and the Raspberry Pi - Simon Ritte...
jaxLondonConference
 
PDF
Little words of wisdom for the developer - Guillaume Laforge (Pivotal)
jaxLondonConference
 
PPTX
Large scale, interactive ad-hoc queries over different datastores with Apache...
jaxLondonConference
 
PDF
Designing Resilient Application Platforms with Apache Cassandra - Hayato Shim...
jaxLondonConference
 
PPTX
Practical Performance: Understand the Performance of Your Application - Chris...
jaxLondonConference
 
Conflict Free Replicated Data-types in Eventually Consistent Systems - Joel J...
jaxLondonConference
 
JVM Support for Multitenant Applications - Steve Poole (IBM)
jaxLondonConference
 
What You Need to Know About Lambdas - Jamie Allen (Typesafe)
jaxLondonConference
 
Java Testing With Spock - Ken Sipe (Trexin Consulting)
jaxLondonConference
 
The Java Virtual Machine is Over - The Polyglot VM is here - Marcus Lagergren...
jaxLondonConference
 
Java EE 7 Platform: Boosting Productivity and Embracing HTML5 - Arun Gupta (R...
jaxLondonConference
 
Exploring the Talend unified Big Data toolset for sentiment analysis - Ben Br...
jaxLondonConference
 
The Curious Clojurist - Neal Ford (Thoughtworks)
jaxLondonConference
 
TDD at scale - Mash Badar (UBS)
jaxLondonConference
 
Run Your Java Code on Cloud Foundry - Andy Piper (Pivotal)
jaxLondonConference
 
Put your Java apps to sleep? Find out how - John Matthew Holt (Waratek)
jaxLondonConference
 
Project Lambda: Functional Programming Constructs in Java - Simon Ritter (Ora...
jaxLondonConference
 
Do You Like Coffee with Your dessert? Java and the Raspberry Pi - Simon Ritte...
jaxLondonConference
 
Little words of wisdom for the developer - Guillaume Laforge (Pivotal)
jaxLondonConference
 
Large scale, interactive ad-hoc queries over different datastores with Apache...
jaxLondonConference
 
Designing Resilient Application Platforms with Apache Cassandra - Hayato Shim...
jaxLondonConference
 
Practical Performance: Understand the Performance of Your Application - Chris...
jaxLondonConference
 

Recently uploaded (20)

PDF
What’s my job again? Slides from Mark Simos talk at 2025 Tampa BSides
Mark Simos
 
PDF
Reverse Engineering of Security Products: Developing an Advanced Microsoft De...
nwbxhhcyjv
 
PPTX
Agentforce World Tour Toronto '25 - MCP with MuleSoft
Alexandra N. Martinez
 
PDF
Transforming Utility Networks: Large-scale Data Migrations with FME
Safe Software
 
PDF
Newgen 2022-Forrester Newgen TEI_13 05 2022-The-Total-Economic-Impact-Newgen-...
darshakparmar
 
PDF
Mastering Financial Management in Direct Selling
Epixel MLM Software
 
PPTX
New ThousandEyes Product Innovations: Cisco Live June 2025
ThousandEyes
 
PDF
POV_ Why Enterprises Need to Find Value in ZERO.pdf
darshakparmar
 
PDF
Agentic AI lifecycle for Enterprise Hyper-Automation
Debmalya Biswas
 
PPTX
Mastering ODC + Okta Configuration - Chennai OSUG
HathiMaryA
 
PDF
CIFDAQ Market Wrap for the week of 4th July 2025
CIFDAQ
 
PDF
Kit-Works Team Study_20250627_한달만에만든사내서비스키링(양다윗).pdf
Wonjun Hwang
 
PDF
The Rise of AI and IoT in Mobile App Tech.pdf
IMG Global Infotech
 
DOCX
Python coding for beginners !! Start now!#
Rajni Bhardwaj Grover
 
PDF
[Newgen] NewgenONE Marvin Brochure 1.pdf
darshakparmar
 
PDF
“Computer Vision at Sea: Automated Fish Tracking for Sustainable Fishing,” a ...
Edge AI and Vision Alliance
 
PDF
Go Concurrency Real-World Patterns, Pitfalls, and Playground Battles.pdf
Emily Achieng
 
PDF
Transcript: Book industry state of the nation 2025 - Tech Forum 2025
BookNet Canada
 
PDF
“NPU IP Hardware Shaped Through Software and Use-case Analysis,” a Presentati...
Edge AI and Vision Alliance
 
PPTX
COMPARISON OF RASTER ANALYSIS TOOLS OF QGIS AND ARCGIS
Sharanya Sarkar
 
What’s my job again? Slides from Mark Simos talk at 2025 Tampa BSides
Mark Simos
 
Reverse Engineering of Security Products: Developing an Advanced Microsoft De...
nwbxhhcyjv
 
Agentforce World Tour Toronto '25 - MCP with MuleSoft
Alexandra N. Martinez
 
Transforming Utility Networks: Large-scale Data Migrations with FME
Safe Software
 
Newgen 2022-Forrester Newgen TEI_13 05 2022-The-Total-Economic-Impact-Newgen-...
darshakparmar
 
Mastering Financial Management in Direct Selling
Epixel MLM Software
 
New ThousandEyes Product Innovations: Cisco Live June 2025
ThousandEyes
 
POV_ Why Enterprises Need to Find Value in ZERO.pdf
darshakparmar
 
Agentic AI lifecycle for Enterprise Hyper-Automation
Debmalya Biswas
 
Mastering ODC + Okta Configuration - Chennai OSUG
HathiMaryA
 
CIFDAQ Market Wrap for the week of 4th July 2025
CIFDAQ
 
Kit-Works Team Study_20250627_한달만에만든사내서비스키링(양다윗).pdf
Wonjun Hwang
 
The Rise of AI and IoT in Mobile App Tech.pdf
IMG Global Infotech
 
Python coding for beginners !! Start now!#
Rajni Bhardwaj Grover
 
[Newgen] NewgenONE Marvin Brochure 1.pdf
darshakparmar
 
“Computer Vision at Sea: Automated Fish Tracking for Sustainable Fishing,” a ...
Edge AI and Vision Alliance
 
Go Concurrency Real-World Patterns, Pitfalls, and Playground Battles.pdf
Emily Achieng
 
Transcript: Book industry state of the nation 2025 - Tech Forum 2025
BookNet Canada
 
“NPU IP Hardware Shaped Through Software and Use-case Analysis,” a Presentati...
Edge AI and Vision Alliance
 
COMPARISON OF RASTER ANALYSIS TOOLS OF QGIS AND ARCGIS
Sharanya Sarkar
 

Packed Objects: Fast Talking Java Meets Native Code - Steve Poole (IBM)

  • 1. Steve Poole IBM Fast Talking Java Meets Native Code
  • 2. Important Disclaimers THE INFORMATION CONTAINED IN THIS PRESENTATION IS PROVIDED FOR INFORMATIONAL PURPOSES ONLY. WHILST EFFORTS WERE MADE TO VERIFY THE COMPLETENESS AND ACCURACY OF THE INFORMATION CONTAINED IN THIS PRESENTATION, IT IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED. ALL PERFORMANCE DATA INCLUDED IN THIS PRESENTATION HAVE BEEN GATHERED IN A CONTROLLED ENVIRONMENT. YOUR OWN TEST RESULTS MAY VARY BASED ON HARDWARE, SOFTWARE OR INFRASTRUCTURE DIFFERENCES. ALL DATA INCLUDED IN THIS PRESENTATION ARE MEANT TO BE USED ONLY AS A GUIDE. IN ADDITION, THE INFORMATION CONTAINED IN THIS PRESENTATION IS BASED ON IBM’S CURRENT PRODUCT PLANS AND STRATEGY, WHICH ARE SUBJECT TO CHANGE BY IBM, WITHOUT NOTICE. IBM AND ITS AFFILIATED COMPANIES SHALL NOT BE RESPONSIBLE FOR ANY DAMAGES ARISING OUT OF THE USE OF, OR OTHERWISE RELATED TO, THIS PRESENTATION OR ANY OTHER DOCUMENTATION. NOTHING CONTAINED IN THIS PRESENTATION IS INTENDED TO, OR SHALL HAVE THE EFFECT OF: - CREATING ANY WARRANT OR REPRESENTATION FROM IBM, ITS AFFILIATED COMPANIES OR ITS OR THEIR SUPPLIERS AND/OR LICENSORS
  • 3. Steve Poole ! Works at IBM’s Hursley Laboratory in the UK Involved in IBM Java VM development since before Java was 1 Currently leading IBM’s OpenJDK technical engagement
  • 4. What this talk is about “Packed Objects” aka “Compact Off-Heap Structures in the Java Language” ! This technology is being developed to help address important pressures on Java ! This talk focuses on how to improve Java interop to Non Java applications
  • 5. You’ll learn “Compact Off-Heap Structures in the Java Language” 1. An overview of this technology 2. Why we need your input and support 3. How to get started with using this technology yourself
  • 6. Part 1 Native interoperability Why it’s so important C
  • 7. Interoperability - why is it important? Java is built on talking native ! Operating Systems Processor Architectures File Systems User interfaces Network Interfaces ! Java keeps the WORA monsters away ! !
  • 8. WORA - the components JDK Java code JVM JIT native Class Library Native code Your JNI code Your Java application
  • 9. 2013 - Not so Java-centric JVM JIT Class Library native Your Native code JDK Java code Your JNI code Your Java application
  • 10. 2013 - clusters and new processors non Java Runtime High speed memory channels (RDMA etc) Specialised Processors (GPUs etc)
  • 11. 2013 - Can we survive with JNI 1.1? JIT ? Other runtime JVM Class Library Your Native code JDK Java code Your JNI code Your Java application native High speed memory channels (RDMA etc) Specialised Processors (GPUs etc) ? ?
  • 12. 2013 - we can’t survive with JNI 1.1 • 1997 - JNI 1.1 • Designed for independence • Version independence • Platform Independence • VM Independence • Java at the centre of the world • Not performance focused
  • 13. Packed Objects to the rescue • What we need is • To be able to share data between Java and other runtimes with minimal overhead • To not require a Java centric data view • and share data off-heap too! • Can we do that? • Time for a demo.
  • 14. A simple demo • • Screensaver style Rectangles moved around the screen, bounce off the sides…
  • 15. A simple demo Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! RGB colour elements ! typedef struct { int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! } RECT; !15 position adjustment vectors SDL structure contains location , height and width float version of location - makes it easier to handle different h/w speeds
  • 16. A simple demo Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !16
  • 17. More than one demo Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !17 Frames per second number of JNI calls per frame
  • 18. Demo 1 - all C Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !18
  • 19. Results Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !19 Frames per second 20 number of JNI calls per frame 0
  • 20. Demo 2- Mixed JNI and C Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !20
  • 21. Results Demo Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element Mixed C with PackedObjects Java updates the location of each element !21 Frames per second 20 number of JNI calls per frame 0
  • 22. er frame) Results Demo Frames per second Standard - all C code calling the SDL graphics routines 20 Mixed C and JNI. Java updates the location of each element 11 number of JNI calls per frame 0 200000 (~4MB per frame) Mixed C with PackedObjects Java updates the location of each element 200 000 JNI calls per frame. Copying data, validating data everytime. AND ensuring there is no chance of optimization by the JIT. !22
  • 23. Demo 2- Details Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !23
  • 24. Demo 2- Details Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! for every rectangle… C function calls Java method via JNI passes in x,y, vx and vy (by copy) Java method updates the rectangles new position Java native calls native method passing in new x,y, vx and vy values (by copy) C native method returns Java method returns C function returns !24
  • 25. Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !25
  • 26. er frame) Results Demo Frames per second Standard - all C code calling the SDL graphics routines 20 Mixed C and JNI. Java updates the location of each element 11 Mixed C with PackedObjects Java updates the location of each element !26 number of JNI calls per frame 0 200000 (~4MB per frame)
  • 27. er frame) Results Demo Frames per second number of JNI calls per frame Standard - all C code calling the SDL graphics routines Mixed C and JNI. Java updates the location of each element 11 200000 (~4MB per frame) Mixed C with PackedObjects Java updates the location of each element !27 20 18 1 (4 bytes per frame) 0
  • 28. Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! Initialise Per frame Update rectangle locations Draw rectangles with SDL terminate !28
  • 29. Demo 3 - Packed Objects Draw 100 000 random sized rectangles using SDL ! Move them around the screen and get them to change direction if they hit the sides ! once per frame.. Java method updates each rectangle’s position directly ! } !29 for(Rectangle r: rectangles) { r.x=r.x+r.vx; r.y=r.y+r.vy; …
  • 30. Today Java only speaks Java • Getting data into and out of Java has always required some form of marshaling or serialization process ! • Interaction with native data structures in memory is particularly problematic • JNI is the slowest but safest – But you need good C / C++ knowledge • Unsafe and NIO are faster but more challenging to use – They both have their own programming ‘model’ • If the Java side is in control of storage layouts it’s easier • When mapping existing native structures it’s much, much, much more difficult ! • Ironically the JVM is completely aware of native data formats
  • 31. Tomorrow, with Packed Objects.. Java will let you reference data stored like this ! typedef struct { int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! } RECT; C structure
  • 32. Tomorrow, with Packed Objects.. as if it looked like this ! ! typedef struct { public class Rectangle { ! int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } RECT; C structure } Java structure
  • 33. Tomorrow, with Packed Objects.. ! ! typedef struct { public class Rectangle { ! int red,green,blue; float vx,vy; SDL_Rect rec; float x,y; ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; without copying ! } RECT; C structure } Java structure
  • 34. Tomorrow, with Packed Objects.. ! !! ! typedef struct { ! int red,green,blue; ! float vx,vy; ! SDL_Rect rec; ! float x,y; !! ! } RECT; ! ! ‘C’ runtime And off ! ! ! public class Rectangle { ! ! ! int red,green,blue; ! float vx,xy; ! heap SDLRect rec; ! float x,y; ! ! ! } ! ! Java heap
  • 35. Part 2 Under the covers http://www.flickr.com/photos/mwichary/
  • 36. Under the covers ! public class Rectangle { Let’s look at these classes from a JVM point of view.. ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } public class SDLRect { short x,y int width,height }
  • 37. Under the covers ! ! ! ! ! ! ! ! ! ! ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap ! public class Rectangle { ! short x short y int width int height int red,green,blue; float vx,xy; SDLRect rec; float x,y; ! } public class SDLRect { short x,y int width,height } Java structure !37 J
  • 38. ! ! ! ! ! ! ! ! ! ! ! Questions ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !38 1 - How does a JVM ‘know’ the class of these objects? ! ! short x short y int width int height 2 - What are the data format rules for these types? ! ! 3 - When synchonizing on a object - where is the monitor stored?
  • 39. ! ! ! ! ! ! ! ! ! ! ! int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !39 Answers class ptr monitor ! class ptr monitor short x short y int width int height 1 - How does a JVM ‘know’ the class of these objects? Additional internal ptr ! 2 - What are the data format rules for these types? JVM impl specific (including ordering) 3 - When synchonizing on a object - where is the monitor stored? Additional internal field
  • 40. How to add the metadata ! ! ! ! ! ! ! ! ! ! ! class ptr monitor int red int green int blue float vx float vy SLDRect ptr float x float y Java heap !40 class ptr monitor short x short y int width int height ! ! ! ! ! ! ! ! ! ! ! [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime No space No space for the extra for the extra Java fields -Java fields what do we what do we do? do?
  • 41. Under the covers ! ! ! ! ! ! ! ! ! ! ! Java heap !41 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 42. Under the covers ! ! ! ! ! ! ! ! ! ! ! Java heap ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height } Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); !42
  • 43. Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” Java heap !43 ! ! ! ! ! ! Constructor ! was not ! called! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 44. Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” Java heap SDLRect s=r.rect; ! !44 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 45. Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr Java heap SDLRect s=r.rect; ! !45 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 46. Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr Java heap SDLRect s=r.rect; SDLRect t=r.rect; !46 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 47. Rectangle r=PackedObject.newNativePackedObject(Rectangle.class,0x******); Under the covers ! ! ! ! ! ! ! ! ! ! ! class ptr monitor data ptr “r” “s” class ptr monitor data ptr “t” class ptr monitor data ptr Java heap SDLRect s=r.rect; SDLRect t=r.rect; !47 ! ! ! ! ! ! ! ! ! ! ! public class Rectangle { [4] red [4] green [4] blue [4] vx [4] vy [2] rect.x [2] rect.y [4] rect.width [4] rect.height [4] x [4] y ‘C’ runtime ! int red,green,blue; float vx,xy; SDLRect rec; float x,y; } public class SDLRect { short x,y int width,height }
  • 48. Part 3 Truth and Consequences
  • 49. Design considerations • Off heap Packed Objects are considered to already exist so constructors are not called. • Since there is limited opportunity to store the vital metadata it is constructed when needed. • The assumption is that creating new objects when accessing embedded data is confusing and dangerous. • How does ‘==‘ work? • What about synchronization? • The idea of ‘identityless’ objects is being discussed. • Something that PackedObjects has in common with ValueTypes • Could you handle Objects of this style?
  • 50. The way forward? • This approach offers Java a way forward in the multi environment world • Almost zero cost data interop will benefit Java when exploiting new processors and memory systems. • There are additional use cases that ‘fall out’ from this design. • Low cost r/w of files • Instant application startup • Basic unmanaged memory capabilities • Under the covers heap usage benefits through better control for ‘locality of reference’
  • 51. Is this the right or only way? • IBM and Oracle have been discussing these sorts of concepts for some time • We need your input. How important are these use cases to you? What’s missing, not needed? What do you hate? • We are trying to keep clear of syntax discussions at this point - it’s too early and just distracting.
  • 52. You too can use Packed Objects • Google for “IBM Java 8 Beta”
  • 53. Hard hats must be worn… The current prototype implementation surfaces a lot of the internals that will eventually just disappear This approach makes it easier for us to quickly revise the design but means users of the prototype have to work harder. ! AND ! it avoids syntax wars :-) picture courtesy http://www.flickr.com/people/karen_od/
  • 54. Wrap up -hopefully you now: “Compact Off-Heap Structures in the Java Language” 1. Have a basic understanding of this technology 2. Know why we need your input and support 3. Know how to get started with using this technology Thank You