We think you have liked this presentation. If you wish to download it, please recommend it to your friends in any social system. Share buttons are a little bit lower. Thank you!
Presentation is loading. Please wait.
Published byJorden Branson
Modified over 2 years ago
© 2009 IBM Corporation1 Feedback Directed Dynamic Recompilation for Statically Compiled Languages Dorit Nuzman, Sergei Dyshel, Revital Eres IBM Research, Haifa Thematic Session on Dynamic Compilation HiPEAC Computing Systems Week Paris, May 3 rd 2013
© 2009 IBM Corporation2 Motivating Scenario (IBMs) customer Independent Software Vendor Computer System Vendor (e.g., IBM) Third party software owned by some ISV Power780 server Increase target platform level? Performance problem Increase optimization level? Apply feedback directed optimization? No Nope Cant do
© 2009 IBM Corporation3 Fat Binary Runtime Engine Profiler Intermediate Representation Dynamic execution stage Program Source Code Static Compiler Motivating Scenario (IBMs) customer Independent Software Vendor Computer System Vendor (e.g., IBM) Power780 server Performance problem Native machine code JIT compiler opt = -O2 arch = common no-profile
© 2009 IBM Corporation4 Fat Binary Runtime Engine Profiler Intermediate Representation Dynamic execution stage Program Source Code Static Compiler Motivating Scenario (IBMs) customer Independent Software Vendor Computer System Vendor (e.g., IBM) Power780 server Performance problem Native machine code JIT compiler opt = -O2 arch = common no-profile
© 2009 IBM Corporation5 Fat Binary Runtime Engine Profiler Intermediate Representation Dynamic execution stage Program Source Code Static Compiler selective profile-driven recompilation Native machine code JIT compiler Our approach: Fat Binary based, feedback-directed, dynamic recompilation Used for years in dynamic languages & Java Needed also for static languages Opposed to dynamic binary optimization: includes high-level semantic information allows aggressive, speculative transformations
© 2009 IBM Corporation6 Background Modern compilers provide sophisticated optimizations. O3 (O4, O5) Inter-procedural Auto-vect/par Feedback-directed Hardware-specific Complicates build process Prolongs development & testing cycle Requires per-customer tuning – too costly No representative input We can gain back the lost performance benefit by applying the optimizations dynamically, at runtime. These optimizations are usually not used. –Only in benchmarking and HPC
© 2009 IBM Corporation7 Dynamic Recompilation Solves the static-compiler usability issue –Transparent feedback-directed optimization for current workload. –Tuning for current hardware –Separation of optimization from software production Allows adaptive optimization. Allows iterative optimization. Virtualization & Cloud: physical resources known only at runtime, and continuously change
© 2009 IBM Corporation8 Other Approaches: Focus only on very long running programs with heavy workloads to compensate for time spent profiling. Focus on optimization across consecutive runs of repetitive programs Domain specific (focus on a specific optimization, to a small pre-selected part of the code) Trace-based binary-optimization …Our Goal: Demonstrate an execution environment with overheads that are low enough to allow the dynamic optimizer to speed up execution of the current invocation, for regular programs/workloads. Dynamic Recompilation for Static Languages …Our Goal: Demonstrate an execution environment with overheads that are low enough to allow the dynamic optimizer to speed up execution of the current invocation, for regular programs/workloads.
© 2009 IBM Corporation9 Fat Binary Runtime Engine Profiler Split-IR Dynamic execution stage Program Source Code Static Compiler Native machine code JIT compiler Our approach: Fat Binary based, feedback-directed, dynamic recompilation
© 2009 IBM Corporation10 t0t0 t1t1 t2t2 t3t3 Execution and sampling thread t4t4 t5t5 t6t6 t7t7 t8t8 t9t9 Instrumentation-based profiling sampling-based profiling for method hottness Original method version Instrumented method version Optimized method version Instrumentation Optimization Recompilation thread Runtime Monitoring and Recompilation timeline Startup cost (loading & mapping) monitoring overhead Recompilation cost Slow instrumented execution Synchronization cost
© 2009 IBM Corporation11 SPECint2006: Dynamic Optimization Overheads – ref dataset Overall not degrading performance. Stress test1: using highly statically-optimized executable (–O3 -qhot)
© 2009 IBM Corporation12 SPECint2006: Dynamic Optimization Overheads – train dataset Works also for very short running programs. Stress test2: using highly statically-optimized executable (–O3 -qhot) Currently limited gain from FDO alone.
© 2009 IBM Corporation13 Optimization effect (isolated from overheads) (1) Similar impact gained using sampled profile as with using a perfect profiles. the problem is not it the profile quality (2) offline optimizer applies link- time FDO (cross methods and modules). Our optimizer limited currently to single module
© 2009 IBM Corporation14 Fat Binary Runtime Engine Profiler Intermediate Representation Dynamic execution stage Program Source Code Static Compiler (IBMs) customer Independent Software Vendor Computer System Vendor (e.g., IBM) Power780 server Native machine code JIT compiler opt = -O2 arch = common no-profile programs are statically under-optimized / moderately-optimized
© 2009 IBM Corporation15 SPECint2006: Overall Effect of Dynamic Execution (ref) Overall 7% improvement on average moderately-optimized scenario (program statically compiled with –O2) Selected methods from the program dynamically recompiled using a higher optimization level.
© 2009 IBM Corporation16 Selected methods from the program dynamically recompiled using a higher optimization level. Recompilation Statistics Default recompilation mode (default method hotness threshold) Aggressive recompilation mode (lower method hotness threshold) moderately-optimized scenario (program statically compiled with –O2) Overall 7% improvement on average Overall 8% improvement on average
© 2009 IBM Corporation17 More Benchmarks: SQlite SQlite: –Static version compiled with default compiler options: -O2 warm. –Using 1G of TPC-H tables. (smallest dataset) –Using TPC-H queries: Stream of 13 instances of query #1 13% improvement from dynamic FDO Most improvement comes from higher optimization level.
© 2009 IBM Corporation18 Overall cost of runtime optimization environment, including – environment startup cost – recompilation – profiling overheads is less than 2% on average (SPECint2006) For highly optimized native binaries, on average, there is no overall degradation These low overheads imply that the fat-binary based approach is practical for real-world use-cases and workloads –Feedback directed optimization can easily surpass these costs Aggressive optimization level for selected methods at runtime brings up to 20% speedup, and an 8% average speedup Much more potential available: – more aggressive optimizations: loop-nest, memory-hierarchy, parallelization – more profiling (event based?) – more synergy with static compiler more synergy with underlying (virtual) environment, to adapt to changes Summary and Conclusions
© 2009 IBM Corporation19 Thematic Session on Dynamic Compilation 1) What is the dynamic optimization stage? During program execution 2) What triggers the dynamic compilation cycle? A method gets warm 3) How are these triggers being detected? sampling execution/PCs (via time interrupts & code instrumentation) to monitor application behavior 4) How/when are the above triggers being inserted? at run-time 5) What is the recompilation scope/granularity? method 6) What is the target application domain? general purpose/commercial applications 7) What is the input code for the dynamic optimization? fat-binary (binary + IR) 8) What is the programming language of the target applications? statically compiled languages (C/C++...) 9) What specific adaptation / optimization / code-transformation is applied? general feedback-directed optimizations (BB ordering, …)
© 2009 IBM Corporation Session one: 1."Asynchronous Dynamic Code Adaptation for Generic Data-Parallel Array Programming" Clemens Grelck (U. of Amsterdam)
© 2012 IBM Corporation On the Benefits and Pitfalls of Extending a Statically Typed Language JIT Compiler for Dynamic Scripting Languages Jose Castanos,
Copyright 2004 Bernd Brügge TUM Software Engineering WS TUM System Design II Bernd Brügge Technische Universität München Applied Software Engineering.
The Client/Server Database Environment CS263 Lecture 12.
Programming Technologies, MIPT, April 7th, 2012 Introduction to Binary Translation Technology Roman Sokolov SMWare
® IBM Software Group © 2008 IBM Corporation A new feature providing mainframe development flexibility David Myers Rational Developer for System z Product.
1 Advanced Database Application Development Performance Tuning Performance Benchmarks Standardization E-Commerce Legacy Systems.
Debugging Tools Towards better use of system tools to weed the nasty critters out of your programs.
Performance Analysis and Optimization through Run-time Simulation and Statistics Philip J. Mucci University Of Tennessee
© Krste Asanovic, 2014CS252, Spring 2014, Lecture 16 CS252 Graduate Computer Architecture Spring 2014 Lecture 16: Virtual Machines Krste Asanovic
1 Wake Up and Smell the Coffee: Performance Analysis Methodologies for the 21st Century Kathryn S McKinley Department of Computer Sciences University of.
SQL tuning and the Dynamic Statement Cache. What Will We Talk About? Some SQL Tuning Fundamentals Dynamic SQL in More Detail Introduction to DB2 Statement.
IBM Tivoli JVM Monitoring – Best Practices Steve Klopfer Technical Specialist, IBM
Parallel computing and message-passing in Java Bryan Carpenter NPAC at Syracuse University Syracuse, NY 13244
1 Computer Systems & Architecture Lesson 3 5. Designing the Architecture.
A load testing solution for web applications This presentation should be viewed as a Slide Show. Choose “Slide Show” from the PowerPoint menu and click.
Data-Awareness and Low- Latency on the Enterprise Grid Getting the Most out of Your Grid with Enterprise IMDG Shay Hassidim Deputy CTO Oct 2007.
Software & Services Group Developer Products Division Copyright© 2011, Intel Corporation. All rights reserved. *Other brands and names are the property.
Chapter 7 – Design and Implementation Lecture 1 1Chapter 7 Design and implementation.
RMS and Scheduling for Future Generation Grids Ramin Yahyapour University Dortmund Leader CoreGRID Institute on Resource Management and Scheduling CoreGRID.
Toward Innovative Model based Enterprise IT Outsourcing NGEBIS Workshop at CAISE 2013 Vinay Kulkarni and Sagar Sunkle.
1 © 2008 Universität Karlsruhe (TU), System Architecture Group 3 Scheduling Problems Kernel Scheduler User Level Scheduler.
© The AspectWerkz Team 2004 What are the key issues for commercial AOP use - how does AspectWerkz address them? Jonas Bonér Senior Software Engineer BEA.
Introduction to cloud computing Jiaheng Lu Department of Computer Science Renmin University of China
What is an Operating System? A program that acts as an intermediary between a user of a computer and the computer hardware. Operating system goals: Execute.
Chapter 7 – Design and Implementation 1Chapter 7 Design and implementation Note: These are a modified version of Ch 7 slides available from the authors.
Silberschatz, Galvin and Gagne ©2013 Operating System Concepts – 9 th Edition Chapter 2: Operating-System Structures.
©Ian Sommerville 2000 Software Engineering, 6th edition. Chapter 10Slide 1 Chapter 10 Architectural Design.
Legal Entity/Division - Date Multicore For Avionics Certification Issue 2013 – 03 – 22.
© 2016 SlidePlayer.com Inc. All rights reserved.