introduction to x10
play

Introduction to X10 Olivier Tardieu IBM Research This material is - PowerPoint PPT Presentation

Feb 01, 2023 •510 likes •835 views

Introduction to X10 Olivier Tardieu IBM Research This material is based upon work supported by the Defense Advanced Research Projects Agency, by the Air Force Office of Scientific Research, and by the Department of Energy. Take Away X10

  1. Introduction to X10 Olivier Tardieu IBM Research This material is based upon work supported by the Defense Advanced Research Projects Agency, by the Air Force Office of Scientific Research, and by the Department of Energy.

  2. Take Away § X10 is § a programming language § derived from and interoperable with Java § an open source tool chain § compilers, runtime, IDE § developed at IBM Research since 2004 with support from DARPA, DoE, and AFOSR § >100 contributors (IBM and Academia) § a growing community § >100 papers § workshops, tutorials, courses § X10 tackles the challenge of programming at scale § first HPC, then clusters, now cloud § scale out: run across many distributed nodes § scale up: exploit multi-core and accelerators § elasticity and resilience § double goal: productivity and performance 2

  3. Links § Main X10 website http://x10-lang.org § X10 Language Specification http://x10.sourceforge.net/documentation/languagespec/x10-latest.pdf § A Brief Introduction to X10 (for the HPC Programmer) http://x10.sourceforge.net/documentation/intro/intro-223.pdf § X10 2.5.3 release (command line tools only) http://sourceforge.net/projects/x10/files/x10/2.5.3/ § X10DT 2.5.3 release (Eclipse-based IDE) http://sourceforge.net/projects/x10/files/x10dt/2.5.3/ 3

  4. Current IBM X10 Team 4

  5. Agenda § X10 overview § APGAS programming model § X10 programming language § Tool chain § Implementation § Applications § 2014/2015 Highlights § Grid X10 5

  6. X10 Overview 6

  7. Asynchronous Partitioned Global Address Space (APGAS) Memory abstraction § Message passing § each task lives in its own address space; example: MPI § Shared memory § shared address space for all the tasks; example: OpenMP § PGAS § global address space: single address space across all tasks § partitioned address space: each partition must fit within a shared-memory node § examples: UPC, Co-array Fortran, X10, Chapel Execution model § SPMD § symmetric tasks progressing in lockstep; examples: MPI, OpenMP 3, UPC, CUDA § APGAS § asynchronous tasks; examples: Cilk, X10, OpenMP 4 tasks 7

  8. Places and Tasks Global ¡Reference ¡ Local ¡ ¡ Local ¡ ¡ Heap ¡ Heap ¡ … ¡ … ¡ … … ¡ … … … … Tasks ¡ Tasks ¡ Place ¡0 ¡ Place ¡N ¡ Task parallelism Concurrency control § async § when async S when(c) S § finish § atomic finish S atomic S Place-shifting operations Distributed heap § at § GlobalRef at(p) S GlobalRef[T] § at § PlaceLocalHandle at(p) e PlaceLocalHandle[T] 8

  9. Idioms § Remote procedure call § SPMD finish for(p in Place.places()) { finish v = at at(p) evalThere(arg1, arg2); at at(p) async async runEverywhere(); } § Active message at(p) async at async runThere(arg1, arg2); § Atomic remote update at at(ref) async async atomic atomic ref() += v; § Divide-and-conquer parallelism def fib(n:Long):Long { § Computation/communication overlap if(n < 2) return n; val acc = new Accumulator(); val f1:Long; while(cond) { val f2:Long; finish finish { finish finish { val v = acc.currentValue(); async async f1 = fib(n-1); at at(ref) async async ref() = v; f2 = fib(n-2); acc.updateValue(); } } return f1 + f2; } } 9

  10. BlockDistRail.x10 public class BlockDistRail[T] { protected val sz:Long; // block size protected val raw:PlaceLocalHandle[Rail[T]]; public def this(sz:Long, places:Long){T haszero} { this.sz = sz; raw = PlaceLocalHandle.make[Rail[T]](PlaceGroup.make(places), ()=>new Rail[T](sz)); } public operator this(i:Long) = (v:T) { at(Place(i/sz)) raw()(i%sz) = v; } public operator this(i:Long) = at(Place(i/sz)) raw()(i%sz); public static def main(Rail[String]) { val rail = new BlockDistRail[Long](5, 4); rail(7) = 8; Console.OUT.println(rail(7)); } } 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 0 0 0 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 0 Place 0 Place 1 Place 2 Place 3 10

  11. Like Java § Objects § classes and interfaces § single-class inheritance, multiple interfaces § fields, methods, constructors § virtual dispatch, overriding, overloading, static methods § Packages and files § Garbage collector § Variables and values (final variables, but final is the default) § definite assignment (extended to tasks) § Expressions and statements § control statements: if, switch, for, while, do-while, break, continue, return § Exceptions § try-catch-finally, throw § Comprehension loops and iterators 11

  12. Beyond Java § Syntax § types “ x:Long ” rather than “ Long x ” § declarations val , var , def § function literals (a:Long, b:Long) => a < b ? a : b § ranges 0..(size-1) § operators user-defined behavior for standard operators § Types § local type inference val b = false; § function types (Long, Long) => Long § typedefs type BinOp[T] = (T, T) => T; § structs headerless inline objects; extensible primitive types § arrays multi-dimensional, distributed; implemented in X10 § properties and constraints extended static checking gradual typing § reified generics templates; constrained kinds 12

  13. Tool Chain 13

  14. Tool Chain § Eclipse Public License § “Native” X10 implementation § C++ based; CUDA support § distributed multi-process (one place per process + one place per GPU) § C/POSIX network abstraction layer (X10RT) § x86, x86_64, Power; Linux, AIX, OS X, Windows/Cygwin, BG/Q; TCP/IP, PAMI, MPI § “Managed” X10 implementation § Java 6/7 based; no CUDA support § distributed multi-JVM (one place per JVM) § pure Java implementation over TCP/IP or using X10RT via JNI (Linux & OS X) § X10DT (Eclipse-based IDE) available for Windows, Linux, OS X § supports many core development tasks including remote build & execute facilities 14

  15. Compilation and Execution X10 Compiler Front-End AST Optimizations X10 Parsing / X10 AST AST Lowering Source Type Check Managed X10 Native X10 X10 AST Java C++ Back-End Back-End Java Interop Java Code C++ Code Generation Generation Java Source C++ Source Cuda Source XRJ Java Compiler XRX Platform Compilers XRC Existing Native (C/C++/ Java Byteode Native executable Existing Java Application etc) Application JNI Native Environment X10RT Java VMs (CPU, GPU, etc) 15

  16. X10DT Building Source navigation, syntax Browsing highlighting, parsing errors, folding, hyperlinking, outline and quick outline, hover help, content assist, type - Java/C++ support hierarchy, format, search, - Local and remote call graph, quick fixes Launching Editing Debug 16

  17. Implementation 17

  18. Runtime § X10RT (X10 runtime transport) § core API: active messages § extended API: collectives & RDMAs X10 Application § emulation layer § two versions: C (+JNI bindings) or pure Java X10 Core Class Libraries § Native runtime XRX § processes, threads, atomic ops § object model (layout, RTTI, serialization) § two versions: C++ and Java Native Runtime § XRX (X10 runtime in X10) X10RT § async, finish, at, when, atomic § X10 code compiled to C++ or Java PAMI TCP/IP MPI SHM CUDA § Core X10 libraries § x10.array, io, util, util.concurrent 18

  19. APGAS Constructs § One process per place § Local tasks: async & finish § thread pool; cooperative work-stealing scheduler § Remote tasks: at(p) async § source side: synthetize active message § async id + serialized heap + control state (finish, clocks) § compiler identifies captured variables (roots); runtime serializes heap reachable from roots § destination side: decode active message § polling (when idle + on runtime entry) § Distributed finish § complex and potentially costly due to message reordering § pattern-based specialization; program analysis 19

  20. Applications 20

  21. HPC Challenge 2012 – X10 at Petascale – Power 775 G-FFT G-HPL EP Stream (Triad) 26958 500000 30000 1.20 800000 24.00 15 396614 22.4 589231 25000 1.00 Gflops/place Gflops/place 400000 GB/s/place 600000 22.00 0.88 Gflops 20000 0.80 Gflops 10 GB/s 300000 7.23 15000 0.60 400000 20.00 0.82 18.0 200000 10000 0.40 5 7.12 200000 18.00 100000 5000 0.20 0 0.00 0 16.00 0 0 0 16384 32768 0 16384 32768 0 27840 55680 Places Places Places G-RandomAccess UTS 844 900 0.9 700000 14.00 596451 800 0.8 10.93 10.87 600000 12.00 0.82 Million nodes/s/place 0.82 700 0.7 Million nodes/s 500000 10.00 600 0.6 Gups/place 10.71 400000 8.00 Gups 500 0.5 400 0.4 300000 6.00 356344 300 0.3 200000 4.00 200 0.2 100000 2.00 100 0.1 0 0 0 0.00 0 8192 16384 24576 32768 0 13920 27840 41760 55680 Places Places 21

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION

INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION

TRAINING COURSE ON SOCIAL PROTECTION &amp; FORMALIZATION TRINIDAD AND TOBAGO MARCH 15, 2017 INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION INTRODUCTION Design of the NIS Assistance from the

547 views • 23 slides
INTRODUCTION Introduction 2/42 INTRODUCTION Alternations I am giving bread to the

INTRODUCTION Introduction 2/42 INTRODUCTION Alternations I am giving bread to the

Patterns of variation in the expression of case and agreement Andrs Brny Leiden University Centre for Linguistics 9 November 2019, BaSIS Workshop a.barany@hum.leidenuniv.nl INTRODUCTION Introduction 2/42 INTRODUCTION

1.37k views • 42 slides
Introduction Introduction Introduction Introduction Outline Motivation Failures

Introduction Introduction Introduction Introduction Outline Motivation Failures

Introduction Introduction Introduction Introduction Outline Motivation Failures Definition and concepts Process and product properties Principles SE Approaches Motivation Software and the economy The economies of

1.03k views • 49 slides
Introduction Introduction Introduction Nationwide Cause for Concern 1

Introduction Introduction Introduction Nationwide Cause for Concern 1

Introduction Introduction Introduction Nationwide Cause for Concern 1 https://www.cdc.gov/std/stats17/infographic.htm Introduction Epidemiology in Western Colorado CPHE Surveillance Data Report, May 2018 Chlamydia Epidemiology in Western

977 views • 39 slides
DAQ introduction Purpose of this talk : (1) Introduction for those who have not been in every

DAQ introduction Purpose of this talk : (1) Introduction for those who have not been in every

DAQ introduction Purpose of this talk : (1) Introduction for those who have not been in every meeting (2) Some ideas of diagrams for section 7.1 (Introduction) of TP. Giles Barr 25 Jan 2018, Pembroke College On DUNE, the reality is there is a

474 views • 10 slides
INTRODUCTION cf. Schneider, Chapter 1 INTRODUCTION THEY ARE OUT TO GET YOU INTRODUCTION WHAT

INTRODUCTION cf. Schneider, Chapter 1 INTRODUCTION THEY ARE OUT TO GET YOU INTRODUCTION WHAT

ADVANCED COMPUTER SECURITY INTRODUCTION cf. Schneider, Chapter 1 INTRODUCTION THEY ARE OUT TO GET YOU INTRODUCTION WHAT IS SECURITY? A systems security policies describe What the system is supposed to do Store and provide access

479 views • 16 slides
Overview Introduction to SMIL Introduction to W3C and XML Introduction to SMIL

Overview Introduction to SMIL Introduction to W3C and XML Introduction to SMIL

Overview Introduction to SMIL Introduction to W3C and XML Introduction to SMIL Writing a SMIL file Adding Media Objects Martin Halvey Clipping media files January 2008 Timing and Synchronising Layout World Wide

387 views • 14 slides
14 Introduction Introduction Bad guys can put malware into Example: hosts

14 Introduction Introduction Bad guys can put malware into Example: hosts

Introduction Introduction source Encapsulation ISO/OSI reference model message M application segment H t H t M transport network datagram H n H n H t M presentation: allow applications to frame link H l H n H t M interpret

583 views • 4 slides
Previous Lecture Introduction to SMIL 2 Introduction to W3C and XML Introduction to SMIL

Previous Lecture Introduction to SMIL 2 Introduction to W3C and XML Introduction to SMIL

Previous Lecture Introduction to SMIL 2 Introduction to W3C and XML Introduction to SMIL Writing a SMIL file Adding Media Objects Martin Halvey Clipping media files January 2008 Timing and Synchronising Layout

322 views • 8 slides
Lecture 1 Andreas Habegger Introduction Zynq Introduction Zynq Introduction Zynq PS vs. PL

Lecture 1 Andreas Habegger Introduction Zynq Introduction Zynq Introduction Zynq PS vs. PL

Introduction Zynq Lecture 1 Andreas Habegger Introduction Zynq Introduction Zynq Introduction Zynq PS vs. PL Processing System Processor Peripherals Data Buses AXI Bus Conclusion BTE5380 - Embedded Systems Octobre 2014 Andreas Habegger

1.81k views • 30 slides
2018.06 01 SMILE5 Introduction S E 5 02 Alpha Cloud M I L 03 Company Introduction 04

2018.06 01 SMILE5 Introduction S E 5 02 Alpha Cloud M I L 03 Company Introduction 04

2018.06 01 SMILE5 Introduction S E 5 02 Alpha Cloud M I L 03 Company Introduction 04 Project References S E 5 PART.1 SMILE5 Introduction M I L Introduction STORION-SMILE5 Residential Energy Storage System Inverter 5kW Output Input

1.06k views • 25 slides
Introduction to CICS Course introduction Course introduction What is CICS? What is an

Introduction to CICS Course introduction Course introduction What is CICS? What is an

Introduction to CICS Course introduction Course introduction What is CICS? What is an application server? Why use an application server? Course introduction Services provided by CICS How CICS applications are defined Scaling CICS to meet

1.84k views • 146 slides
Network Visualization Introduction Presented by Shahed Introduction Introduction Basic

Network Visualization Introduction Presented by Shahed Introduction Introduction Basic

Network Visualization Introduction Presented by Shahed Introduction Introduction Basic building blocks Node Links (relationship between nodes) Spatial information Network data

584 views • 9 slides
Introduction Introduction (1) Main argument of the paper: universities are not only

Introduction Introduction (1) Main argument of the paper: universities are not only

Le jugement des pairs comme outil de management des universits Peer-review as a management tool Christine Musselin (CSO, Sciences Po et CNRS) Avril 2013, Lirrsistible ascension du capitalisme acadmique Introduction Introduction (1)

590 views • 28 slides
Introduction Introduction to storage and to storage and filesystems filesystems Introduction

Introduction Introduction to storage and to storage and filesystems filesystems Introduction

Moreno Baricevic Gilberto Daz Axel Kohlmeyer Stefano Cozzini ULA ICTP CNR-IOM DEMOCRITOS Merida, VENEZUELA Trieste, ITALY Trieste, ITALY Introduction Introduction to storage and to storage and filesystems filesystems Introduction

1.19k views • 51 slides
JABLOTRON JA-100 introduction November 2011 Introduction Todays goals First introduction of

JABLOTRON JA-100 introduction November 2011 Introduction Todays goals First introduction of

JABLOTRON JA-100 introduction November 2011 Introduction Todays goals First introduction of the brand new JA-100 alarm system Visions Basic architecture System components Setting Help us with final validations You are the FIRST ones

1.69k views • 111 slides
Changelog Changes made in this version not seen in fjrst lecture: sz is one past the end of the

Changelog Changes made in this version not seen in fjrst lecture: sz is one past the end of the

Changelog Changes made in this version not seen in fjrst lecture: sz is one past the end of the heap 0 2 November: Correct space on demand from to &lt; and &gt; to since Virtual Memory 1 1 last time deadlock thread A holding a

2.01k views • 162 slides
2.2 Transformations Hao Li http://cs420.hao-li.com 1 OpenGL Transformations Matrices

2.2 Transformations Hao Li http://cs420.hao-li.com 1 OpenGL Transformations Matrices

Fall 2017 CSCI 420: Computer Graphics 2.2 Transformations Hao Li http://cs420.hao-li.com 1 OpenGL Transformations Matrices Model-view matrix (4x4 matrix) Projection matrix (4x4 matrix) vertices in canonical 3D world coordinate

527 views • 41 slides
Completeness of Resolution 5ai We will show by construction: AUTOMATED REASONING If clauses S

Completeness of Resolution 5ai We will show by construction: AUTOMATED REASONING If clauses S

Completeness of Resolution 5ai We will show by construction: AUTOMATED REASONING If clauses S have no models then there is a resolution proof of [] from S . As shown in Slides 4 the construction has two parts : SLIDES 5: (i) find a ground

544 views • 5 slides
MASSACHUSETTS INSTITUTE OF TECHNOLOGY Physics Department 8.044 Statistical Physics I Spring Term

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Physics Department 8.044 Statistical Physics I Spring Term

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Physics Department 8.044 Statistical Physics I Spring Term 2013 Excited State Helium, He An Example of Quantum Statistics in a Two Particle System By definition He has one electron in the lowest

332 views • 5 slides
ESQPT in systems with long-range interactions Lea F. Santos Yeshiva University, New York, NY, USA

ESQPT in systems with long-range interactions Lea F. Santos Yeshiva University, New York, NY, USA

ESQPT in systems with long-range interactions Lea F. Santos Yeshiva University, New York, NY, USA Francisco Prez-Bernal Universidad de Huelva, Spain G. Luca Celardo and Fausto Borgonovi Universita Cattolica del Sacro Cuore, Italy

425 views • 31 slides
Valuing ISR Resources Tod S. Levi6 * , Kellen G. Leister * ,

Valuing ISR Resources Tod S. Levi6 * , Kellen G. Leister * ,

Valuing ISR Resources Tod S. Levi6 * , Kellen G. Leister * , Ronald F. Woodaman * , Jerrit K. Askvig * , Kathryn B. Laskey * , Rick Hayes-Roth ,

862 views • 46 slides
Head First into GlobalISel Or: How to delete SelectionDAG in 100* easy commits 1 LLVM Dev Meeting

Head First into GlobalISel Or: How to delete SelectionDAG in 100* easy commits 1 LLVM Dev Meeting

Head First into GlobalISel Or: How to delete SelectionDAG in 100* easy commits 1 LLVM Dev Meeting 2017 Justin Bogner, Aditya Nandakumar, Daniel Sanders Apple Porting to GlobalISel What is the structure of a GlobalISel backend? How

2.04k views • 177 slides
Pool::count Pool::grow() Pool::alloc() Pool_element_header Pool_element_header::next

Pool::count Pool::grow() Pool::alloc() Pool_element_header Pool_element_header::next

Pool::count Pool::grow() Pool::alloc() Pool_element_header Pool_element_header::next Pool::head Pool::free(void *) Pool Pool::purge() Block_pool_ATTLC::free() Block_header_ATTLC Block_header_ATTLC::next Block_pool_ATTLC::head

362 views • 12 slides

More recommend