Transla'on Out of SSA Form

Size: px
Start display at page:

Download "Transla'on Out of SSA Form"

Transcription

1 COMP 506 Rice University Spring 2017 Transla'on Out of SSA Form Benoit Boissinot, Alain Darte, Benoit Dupont de Dinechin, Christophe Guillon, and Fabrice Rastello, Revisi;ng Out-of-SSA Transla;on for Correctness, Code Quality, and Efficiency, CGO Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice University have explicit permission to make copies of these materials for their personal use. Faculty from other educaional insituions may use these materials for nonprofit educaional purposes, provided this copyright noice is preserved CitaIon numbers refer to entries in the EaC2e bibliography.

2 SSA Deconstruc;on At some point, we need executable code Do machines implement ϕ-operaions? Need to fix up the flow of values Original idea [CFRWZ, 110] Replace ϕ-funcion with copies in predecessor blocks Simple algorithm Works in most cases Adds lots of copies Most of them coalesce away Copy-coalescing is well understood See, for example, Chai0n-Briggs GCRA [75,56] X 17 ϕ (x 10,x 11 )... x 17 X 17 x 10 X 17 x x 17 COMP 506, Rice University 2

3 Transla;on Out of SSA Form Two classic problems arise in SSA deconstruc;on Lost-copy problem Swap problem In each case, simple copy inserion produces incorrect code Cri;cal Observa;on Both problems are caused by transformaions that rewrite the code and move definiions and uses Some of the complicaion arises from the shi_ between the parallel semanics of the Φ-funcIons and the sequenial semanics of copy These problems were idenified by Cliff Click and first published by Briggs et al. in 1997 [50]. They presented ad-hoc ways of solving the problems. (Click is not an author; he found the problems.) This lecture is based on the work of Boissinot et al. and presents a more systemaic approach to solving the problems inherent in translaion out of SSA form. See RevisiIng Out-of-SSA TranslaIon for Correctness, Code Quality, and Efficiency, by Benoit Boissinot, Alain Darte, Benoit Dupont de Dinechin, Christophe Guillon, and COMP Fabrice 506, Rastello Rice University in CGO

4 TRANSLATION OUT OF SSA FORM The Lost Copy Problem i 1 In opimizaion, a cri;cal edge in a control-flow graph is one that runs from a node with muliple successors to node with muliple predecessors. In out-of-ssa transla;on, cri;cal edges create serious problems for copy inserion. i 1 i 0 cri;cal edge y i i i + 1 i 1 Φ(i 0,i 2 ) y 0 i 1 i 1 Φ(i 0,i 2 ) i 1 i 2 d y + z 0 y 0 + Original code In SSA form With copies folded Copies naïvely inserted Copy folding is a simple transformaion that creates the problem. Other transformaions can have the same effect. The assignment to z now receives the wrong value. To fix this problem, the compiler either needs to create a temporary name to hold the penulimate value of i, or to insert the copy in the loop-closing branch. COMP 506, Rice University 4 *

5 TranslaIon Out of SSA Form Copy folding can only produce this code because of the parallel seman;cs of Φ-func;ons The Swap Problem x y x 1 x 0 y 1 y 0 Ignored the parallel seman;cs of the Φ func;ons t x x y y t x 1 Φ(x 0,x 2 ) y 1 Φ(y 0,y 2 ) t 0 x 1 x 2 y 1 y 2 t 0 x 1 Φ(x 0,y 1 ) y 1 Φ(y 0,x 1 ) x 1 y 1 y 1 x Original code In SSA Form Copies folded Copies naïvely inserted Code is incorrect This problem arises when a Φ-funcIon argument is defined by a Φ-funcIon in the same block. To generate correct code, the compiler will need to insert one or more additonal copy operaions and temporary names. * COMP 506, Rice University 5

6 Transla;on Out of SSA Form The Big Picture Swap problem & lost-copy problem arise from messing with Φ-funcIon parameters Renaming Φ-funcIon parameters, moving their definiions, Underlying issue is the parallel semanics of Φ-funcIon evaluaion One way to simplify out-of-ssa translaion is to separate the parallelism from the Φ-funcIons and tackle it directly Convert to a form of SSA where eliminaing the subscripts produces correct code We call that form ConvenIonal SSA or CSSA COMP 506, Rice University 6

7 Transla;on Out of SSA Form The Big Picture Insert parallel copies to convert SSA to convenional SSA In CSSA, we can just drop the subscripts on SSA name Introduces a new set of names Rename out of CSSA by replacing introduced names Eliminate Φ funcions as in original paper Insert copies at end of the predecessor blocks SequenIalize parallel copies (may introduce new temporaries) Aggressive copy coalescing to remove copies Can coalesce copies before renaming or a_er we are done Coalescing parallel copies requires some care May be easier, and clearer, to coalesce a_er sequenializaion Moves the issues created by parallel Φ funcion semanics back into pred. blocks (in parallel copies). e.g., Budimlic et. al. PLDI 2002, or unrestricted coalescing [75,56] Sketch of ideas from Revisi0ng Out-of-SSA Transla0on for Correctness, Efficiency, and Speed, Boissinot, Darte, de Dinechin, Guillon, and Rastello, Proceedings of CGO COMP 506, Rice University 7

8 The Individual Steps Convert SSA to CSSA For a ϕ-funcion a 0 ϕ(a 1,a 2,, a n ) : 1. Insert a parallel copy a 1 a 1 at the end of the block corresponding to a 1 2. Replace a 0 ϕ(a 1,a 2,, a n ) with a 0 ϕ(a 1,a 2,, a n ) 3. Insert a parallel copy a 0 a 0 a_er the ϕ-funcion Rename Out Of CSSA primed name, i j, drop the subscripts and replace each i with a new name Remove Φ Func;ons Replace Φ-funcIons with copies in predecessor blocks as in CFRWZ paper Use parallel copies for good measure Sequen;alize The Parallel Copies Build a dependence graph and break cycles with a new name COMP Copies 506, inserted Rice University for different Φ-funcIons in the same block form a parallel copy group. 8

9 Sequen;alizing Parallel Copies A parallel copy group forms a graph a d a 4 b; b 4 c; c 4 a; d 4 a Parallel copies b c Corresponding graph Graph is either a tree or a cycle Schedule a tree, bopom up from the leaves Must break each cycle with an extra copy May require a new name Other copies may avoid the extra name If possible, break on the value preserved in a non-cycle copy d a a b b c c d Serialized copies Subscripts on the copies indicate parallel copy group membership. COMP 506, Rice University 9

10 Reminder from earlier slide TRANSLATION OUT OF SSA FORM The Lost Copy Problem i 1 i 1 i 0 y i i i + 1 i 1 Φ(i 0,i 2 ) y 0 i 1 i 1 Φ(i 0,i 2 ) i 1 i 2 d y + z 0 y 0 + Original code In SSA form With copies folded Copies naïvely inserted Copy folding is a simple transformaion that creates the problem. Other transformaions can have the same effect. The assignment to z now receives the wrong value. To fix this problem, the compiler needs to create a temporary name to hold the penulimate value of i. COMP 506, Rice University 10

11 TRANSLATION OUT OF SSA FORM The Lost Copy Problem via CSSA i 0 i o x i o i 1 y i i i + 1 i 1 Φ(i 0,i 2 ) i 1 Φ(i 0,i 2 ) i 1 i 1 i 2 i 2 x Φ(x,x) i 1 x x i 2 z y + Original code In SSA with copies folded Insert parallel copies to create CSSA Rename out of CSSA With only one Φ, parallel is a singleton. Each use of i replaced with x. Use of i 1 to compute z 0 is correct. COMP 506, Rice University 11

12 TRANSLATION OUT OF SSA FORM The Lost Copy Problem via CSSA x i o x i o x x x 1 x Φ(x,x) i 1 x x i 2 x Φ(x,x) i 1 x x i 2 x x i 1 x x i i 1 y i i i + 1 z 0 y + From previous slide Replace Φ s with copies A]er coalescing The original code Copies look rather stupid because it is a simple case, but it is correct. COMP 506, Rice University 12

13 Reminder from earlier slide TranslaIon Out of SSA Form The Swap Problem x y x 1 x 0 y 1 y 0 t x x y y t x 1 Φ(x 0,x 2 ) y 1 Φ(y 0,y 2 ) t 0 x 1 x 2 y 1 y 2 t 0 x 1 Φ(x 0,y 1 ) y 1 Φ(y 0,x 1 ) x 1 y 1 y 1 x 1 Original code In SSA Form Copies folded Copies naïvely inserted Code is incorrect This problem arises when a Φ-funcIon argument is defined by a Φ-funcIon in the same block. To generate correct code, the compiler will need to insert one or more additonal copy operaions and temporary names. COMP 506, Rice University 13

14 TRANSLATION OUT OF SSA FORM The Swap Problem x 0 1 x 0 y 0 1 y 0 a 1 x 0 b 1 y 0 a 1 x 0 b 1 y 0 a 4 a b 4 b x 1 Φ(x 0,y 1 ) y 1 Φ(y 0,x 1 ) x 1 Φ(x 0,y 1 ) y 1 Φ(y 0,x 1 ) x 1 2 x 1 y 1 2 y 1 x 1 3 x 1 y 1 3 y 1 a Φ(a,b) b Φ(b,a) x 1 2 a y 1 2 b a 3 x 1 b 3 y 1 x 1 2 a y 1 2 b a 3 x 1 b 3 y 1 a 5 b b 5 a Star;ng Point (code in SSA form) Convert to CSSA Rename primed variables Eliminate Φ func;ons (used parallel copies) COMP 506, Rice University Subscript on indicates a parallel copy group 14

15 TRANSLATION OUT OF SSA FORM The Swap Problem a x 0 a 1 x 0 b 1 y 0 a 4 a b 4 b b a b y 0 a b a x 0 b y 0 a a b b x 1 a x 1 2 a y 1 2 b a 3 x 1 b 3 y 1 a 5 b b 5 a y 1 a b b x 1 y 1 x 1 a y 1 b a x 1 b y 1 a 5 b b 5 a Result from previous slide Parallel copy groups 1, 2, 3, & 4 have acyclic dependence graphs. They can be scheduled in either order. Groups 1, 2, 3, & 4 sequen;alized COMP 506, Rice University 15

16 TRANSLATION OUT OF SSA FORM The Swap Problem a x 0 b y 0 a a b b These copies coalesce away These copies are useless a x 0 b y 0 a a b b x 1 a y 1 b a x 1 b y 1 a 5 b b 5 a a b Cyclic dependence graph for parallel copy group 5 These copies are LIVE These copies are useless x 1 a y 1 b a x 1 b y 1 t a a b b t Must break this cycle with an extra copy & a new name. Groups 1, 2, 3, & 4 sequen;alized All copies are now sequen;al COMP 506, Rice University 16

17 Transla;on Out of SSA Form To recap, the algorithm is Insert parallel copies to convert SSA to convenional SSA In CSSA, we can just drop the subscripts on SSA name Introduces a new set of names Rename out of CSSA by replacing introduced names Eliminate Φ funcions by insering parallel copies in prior blocks May look redundant, but is necessary to handle the swap problem SequenIalize parallel copies (may introduce new temporaries) Aggressive copy coalescing to remove copies Can coalesce copies before renaming or a_er we are done Coalescing parallel copies requires some care May be easier, and clearer, to coalesce a_er sequenializaion e.g., ChaiIn-Briggs unrestricted coalescing or Budimlic et. al. PLDI COMP 506, Rice University 17

18 Using SSA Form in a Compiler Need to translate source or IR into SSA form Algorithm from earlier lecture [110,50] IR needs to represent both Φ-funcIons and Φ-argument to CFG edge correspondance Working with SSA Form The SSA name space makes some kinds of transformaions harder Code moion past Φ-funcIons is problemaic SSA provides a useful sparse representaion that can lead to efficiency Wegman-Zadeck SCP and SCCP as examples [347] Need to translate out of SSA form Algorithms from this lecture LLVM comes out of SSA in code generator (selec0on, alloca0on, scheduling) COMP 506, Rice University 18

Towards an SSA based compiler back-end: some interesting properties of SSA and its extensions

Towards an SSA based compiler back-end: some interesting properties of SSA and its extensions Towards an SSA based compiler back-end: some interesting properties of SSA and its extensions Benoit Boissinot Équipe Compsys Laboratoire de l Informatique du Parallélisme (LIP) École normale supérieure

More information

Combining Optimizations: Sparse Conditional Constant Propagation

Combining Optimizations: Sparse Conditional Constant Propagation Comp 512 Spring 2011 Combining Optimizations: Sparse Conditional Constant Propagation Copyright 2011, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice University

More information

Local Optimization: Value Numbering The Desert Island Optimization. Comp 412 COMP 412 FALL Chapter 8 in EaC2e. target code

Local Optimization: Value Numbering The Desert Island Optimization. Comp 412 COMP 412 FALL Chapter 8 in EaC2e. target code COMP 412 FALL 2017 Local Optimization: Value Numbering The Desert Island Optimization Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon,

More information

Register Allocation. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice.

Register Allocation. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice. Register Allocation Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP at Rice. Copyright 00, Keith D. Cooper & Linda Torczon, all rights reserved.

More information

Instruction Scheduling Beyond Basic Blocks Extended Basic Blocks, Superblock Cloning, & Traces, with a quick introduction to Dominators.

Instruction Scheduling Beyond Basic Blocks Extended Basic Blocks, Superblock Cloning, & Traces, with a quick introduction to Dominators. Instruction Scheduling Beyond Basic Blocks Extended Basic Blocks, Superblock Cloning, & Traces, with a quick introduction to Dominators Comp 412 COMP 412 FALL 2016 source code IR Front End Optimizer Back

More information

Intermediate Representations

Intermediate Representations Most of the material in this lecture comes from Chapter 5 of EaC2 Intermediate Representations Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP

More information

Static single assignment

Static single assignment Static single assignment Control-flow graph Loop-nesting forest Static single assignment SSA with dominance property Unique definition for each variable. Each definition dominates its uses. Static single

More information

Introduction to Optimization. CS434 Compiler Construction Joel Jones Department of Computer Science University of Alabama

Introduction to Optimization. CS434 Compiler Construction Joel Jones Department of Computer Science University of Alabama Introduction to Optimization CS434 Compiler Construction Joel Jones Department of Computer Science University of Alabama Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved.

More information

Instruction Selection: Preliminaries. Comp 412

Instruction Selection: Preliminaries. Comp 412 COMP 412 FALL 2017 Instruction Selection: Preliminaries Comp 412 source code Front End Optimizer Back End target code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled

More information

Parsing II Top-down parsing. Comp 412

Parsing II Top-down parsing. Comp 412 COMP 412 FALL 2018 Parsing II Top-down parsing Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled

More information

CS415 Compilers. Intermediate Represeation & Code Generation

CS415 Compilers. Intermediate Represeation & Code Generation CS415 Compilers Intermediate Represeation & Code Generation These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University Review - Types of Intermediate Representations

More information

Loop Invariant Code Mo0on

Loop Invariant Code Mo0on COMP 512 Rice University Spring 2015 Loop Invariant Code Mo0on A Simple Classical Approach Copyright 2015, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice University

More information

Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit

Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit Intermediate Representations Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit permission to make copies

More information

The Software Stack: From Assembly Language to Machine Code

The Software Stack: From Assembly Language to Machine Code COMP 506 Rice University Spring 2018 The Software Stack: From Assembly Language to Machine Code source code IR Front End Optimizer Back End IR target code Somewhere Out Here Copyright 2018, Keith D. Cooper

More information

Revisiting Out-of-SSA Translation for Correctness, Code Quality, and Efficiency

Revisiting Out-of-SSA Translation for Correctness, Code Quality, and Efficiency Revisiting Out-of-SSA Translation for Correctness, Code Quality, and Efficiency Benoit Boissinot, Alain Darte, Fabrice Rastello, Benoît Dupont de Dinechin, Christophe Guillon To cite this version: Benoit

More information

Introduction to Optimization Local Value Numbering

Introduction to Optimization Local Value Numbering COMP 506 Rice University Spring 2018 Introduction to Optimization Local Value Numbering source IR IR target code Front End Optimizer Back End code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights

More information

Intermediate Representations

Intermediate Representations COMP 506 Rice University Spring 2018 Intermediate Representations source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students

More information

6. Intermediate Representation

6. Intermediate Representation 6. Intermediate Representation Oscar Nierstrasz Thanks to Jens Palsberg and Tony Hosking for their kind permission to reuse and adapt the CS132 and CS502 lecture notes. http://www.cs.ucla.edu/~palsberg/

More information

Lab 3, Tutorial 1 Comp 412

Lab 3, Tutorial 1 Comp 412 COMP 412 FALL 2018 Lab 3, Tutorial 1 Comp 412 source code IR IR Front End Optimizer Back End target code Copyright 2018, Keith D. Cooper, Linda Torczon & Zoran Budimlić, all rights reserved. Students enrolled

More information

Syntax Analysis, III Comp 412

Syntax Analysis, III Comp 412 COMP 412 FALL 2017 Syntax Analysis, III Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp

More information

The Processor Memory Hierarchy

The Processor Memory Hierarchy Corrected COMP 506 Rice University Spring 2018 The Processor Memory Hierarchy source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved.

More information

A Sparse Algorithm for Predicated Global Value Numbering

A Sparse Algorithm for Predicated Global Value Numbering Sparse Predicated Global Value Numbering A Sparse Algorithm for Predicated Global Value Numbering Karthik Gargi Hewlett-Packard India Software Operation PLDI 02 Monday 17 June 2002 1. Introduction 2. Brute

More information

Global Register Allocation via Graph Coloring

Global Register Allocation via Graph Coloring Global Register Allocation via Graph Coloring Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit permission

More information

Intermediate Representations Part II

Intermediate Representations Part II Intermediate Representations Part II Types of Intermediate Representations Three major categories Structural Linear Hybrid Directed Acyclic Graph A directed acyclic graph (DAG) is an AST with a unique

More information

CS153: Compilers Lecture 23: Static Single Assignment Form

CS153: Compilers Lecture 23: Static Single Assignment Form CS153: Compilers Lecture 23: Static Single Assignment Form Stephen Chong https://www.seas.harvard.edu/courses/cs153 Pre-class Puzzle Suppose we want to compute an analysis over CFGs. We have two possible

More information

Just-In-Time Compilers & Runtime Optimizers

Just-In-Time Compilers & Runtime Optimizers COMP 412 FALL 2017 Just-In-Time Compilers & Runtime Optimizers Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved.

More information

Syntax Analysis, III Comp 412

Syntax Analysis, III Comp 412 Updated algorithm for removal of indirect left recursion to match EaC3e (3/2018) COMP 412 FALL 2018 Midterm Exam: Thursday October 18, 7PM Herzstein Amphitheater Syntax Analysis, III Comp 412 source code

More information

Generating Code for Assignment Statements back to work. Comp 412 COMP 412 FALL Chapters 4, 6 & 7 in EaC2e. source code. IR IR target.

Generating Code for Assignment Statements back to work. Comp 412 COMP 412 FALL Chapters 4, 6 & 7 in EaC2e. source code. IR IR target. COMP 412 FALL 2017 Generating Code for Assignment Statements back to work Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights

More information

CSE P 501 Compilers. SSA Hal Perkins Spring UW CSE P 501 Spring 2018 V-1

CSE P 501 Compilers. SSA Hal Perkins Spring UW CSE P 501 Spring 2018 V-1 CSE P 0 Compilers SSA Hal Perkins Spring 0 UW CSE P 0 Spring 0 V- Agenda Overview of SSA IR Constructing SSA graphs Sample of SSA-based optimizations Converting back from SSA form Sources: Appel ch., also

More information

Code Shape Comp 412 COMP 412 FALL Chapters 4, 5, 6 & 7 in EaC2e. source code. IR IR target. code. Front End Optimizer Back End

Code Shape Comp 412 COMP 412 FALL Chapters 4, 5, 6 & 7 in EaC2e. source code. IR IR target. code. Front End Optimizer Back End COMP 412 FALL 2017 Code Shape Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at

More information

Implementing Control Flow Constructs Comp 412

Implementing Control Flow Constructs Comp 412 COMP 412 FALL 2018 Implementing Control Flow Constructs Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students

More information

Reuse Optimization. LLVM Compiler Infrastructure. Local Value Numbering. Local Value Numbering (cont)

Reuse Optimization. LLVM Compiler Infrastructure. Local Value Numbering. Local Value Numbering (cont) LLVM Compiler Infrastructure Source: LLVM: A Compilation Framework for Lifelong Program Analysis & Transformation by Lattner and Adve Reuse Optimization Eliminate redundant operations in the dynamic execution

More information

6. Intermediate Representation!

6. Intermediate Representation! 6. Intermediate Representation! Prof. O. Nierstrasz! Thanks to Jens Palsberg and Tony Hosking for their kind permission to reuse and adapt the CS132 and CS502 lecture notes.! http://www.cs.ucla.edu/~palsberg/!

More information

Lazy Code Motion. Comp 512 Spring 2011

Lazy Code Motion. Comp 512 Spring 2011 Comp 512 Spring 2011 Lazy Code Motion Lazy Code Motion, J. Knoop, O. Ruthing, & B. Steffen, in Proceedings of the ACM SIGPLAN 92 Conference on Programming Language Design and Implementation, June 1992.

More information

CS415 Compilers Register Allocation. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University

CS415 Compilers Register Allocation. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University CS415 Compilers Register Allocation These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University Review: The Back End IR Instruction Selection IR Register

More information

Building a Parser Part III

Building a Parser Part III COMP 506 Rice University Spring 2018 Building a Parser Part III With Practical Application To Lab One source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda

More information

Overview Of Op*miza*on, 2

Overview Of Op*miza*on, 2 OMP 512 Rice University Spring 2015 Overview Of Op*miza*on, 2 Superlocal Value Numbering, SE opyright 2015, Keith D. ooper & Linda Torczon, all rights reserved. Students enrolled in omp 512 at Rice University

More information

Syntax Analysis, VI Examples from LR Parsing. Comp 412

Syntax Analysis, VI Examples from LR Parsing. Comp 412 Midterm Exam: Thursday October 18, 7PM Herzstein Amphitheater Syntax Analysis, VI Examples from LR Parsing Comp 412 COMP 412 FALL 2018 source code IR IR target Front End Optimizer Back End code Copyright

More information

Runtime Support for OOLs Object Records, Code Vectors, Inheritance Comp 412

Runtime Support for OOLs Object Records, Code Vectors, Inheritance Comp 412 COMP 412 FALL 2017 Runtime Support for OOLs Object Records, Code Vectors, Inheritance Comp 412 source IR Front End Optimizer Back End IR target Copyright 2017, Keith D. Cooper & Linda Torczon, all rights

More information

Register Allocation. Global Register Allocation Webs and Graph Coloring Node Splitting and Other Transformations

Register Allocation. Global Register Allocation Webs and Graph Coloring Node Splitting and Other Transformations Register Allocation Global Register Allocation Webs and Graph Coloring Node Splitting and Other Transformations Copyright 2015, Pedro C. Diniz, all rights reserved. Students enrolled in the Compilers class

More information

Operator Strength Reduc1on

Operator Strength Reduc1on COMP 512 Rice University Spring 2015 Generali*es and the Cocke- Kennedy Algorithm Copyright 2015, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice University

More information

Naming in OOLs and Storage Layout Comp 412

Naming in OOLs and Storage Layout Comp 412 COMP 412 FALL 2018 Naming in OOLs and Storage Layout Comp 412 source IR IR target Front End Optimizer Back End Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in

More information

Introduction to Machine-Independent Optimizations - 6

Introduction to Machine-Independent Optimizations - 6 Introduction to Machine-Independent Optimizations - 6 Machine-Independent Optimization Algorithms Department of Computer Science and Automation Indian Institute of Science Bangalore 560 012 NPTEL Course

More information

Syntax Analysis, V Bottom-up Parsing & The Magic of Handles Comp 412

Syntax Analysis, V Bottom-up Parsing & The Magic of Handles Comp 412 Midterm Exam: Thursday October 18, 7PM Herzstein Amphitheater Syntax Analysis, V Bottom-up Parsing & The Magic of Handles Comp 412 COMP 412 FALL 2018 source code IR Front End Optimizer Back End IR target

More information

8. Static Single Assignment Form. Marcus Denker

8. Static Single Assignment Form. Marcus Denker 8. Static Single Assignment Form Marcus Denker Roadmap > Static Single Assignment Form (SSA) > Converting to SSA Form > Examples > Transforming out of SSA 2 Static Single Assignment Form > Goal: simplify

More information

Parsing. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice.

Parsing. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice. Parsing Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice. Copyright 2010, Keith D. Cooper & Linda Torczon, all rights reserved. Students

More information

Intermediate Representations

Intermediate Representations Intermediate Representations Intermediate Representations (EaC Chaper 5) Source Code Front End IR Middle End IR Back End Target Code Front end - produces an intermediate representation (IR) Middle end

More information

Runtime Support for Algol-Like Languages Comp 412

Runtime Support for Algol-Like Languages Comp 412 COMP 412 FALL 2018 Runtime Support for Algol-Like Languages Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students

More information

Computing Inside The Parser Syntax-Directed Translation. Comp 412 COMP 412 FALL Chapter 4 in EaC2e. source code. IR IR target.

Computing Inside The Parser Syntax-Directed Translation. Comp 412 COMP 412 FALL Chapter 4 in EaC2e. source code. IR IR target. COMP 412 FALL 2017 Computing Inside The Parser Syntax-Directed Translation Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights

More information

Arrays and Functions

Arrays and Functions COMP 506 Rice University Spring 2018 Arrays and Functions source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled

More information

Loops and Locality. with an introduc-on to the memory hierarchy. COMP 506 Rice University Spring target code. source code OpJmizer

Loops and Locality. with an introduc-on to the memory hierarchy. COMP 506 Rice University Spring target code. source code OpJmizer COMP 506 Rice University Spring 2017 Loops and Locality with an introduc-on to the memory hierarchy source code Front End IR OpJmizer IR Back End target code Copyright 2017, Keith D. Cooper & Linda Torczon,

More information

Parsing II Top-down parsing. Comp 412

Parsing II Top-down parsing. Comp 412 COMP 412 FALL 2017 Parsing II Top-down parsing Comp 412 source code IR Front End OpMmizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled

More information

Code Shape II Expressions & Assignment

Code Shape II Expressions & Assignment Code Shape II Expressions & Assignment Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit permission to make

More information

ECE 5775 High-Level Digital Design Automation Fall More CFG Static Single Assignment

ECE 5775 High-Level Digital Design Automation Fall More CFG Static Single Assignment ECE 5775 High-Level Digital Design Automation Fall 2018 More CFG Static Single Assignment Announcements HW 1 due next Monday (9/17) Lab 2 will be released tonight (due 9/24) Instructor OH cancelled this

More information

Lexical Analysis - An Introduction. Lecture 4 Spring 2005 Department of Computer Science University of Alabama Joel Jones

Lexical Analysis - An Introduction. Lecture 4 Spring 2005 Department of Computer Science University of Alabama Joel Jones Lexical Analysis - An Introduction Lecture 4 Spring 2005 Department of Computer Science University of Alabama Joel Jones Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved.

More information

Compiler Construction 2009/2010 SSA Static Single Assignment Form

Compiler Construction 2009/2010 SSA Static Single Assignment Form Compiler Construction 2009/2010 SSA Static Single Assignment Form Peter Thiemann March 15, 2010 Outline 1 Static Single-Assignment Form 2 Converting to SSA Form 3 Optimization Algorithms Using SSA 4 Dependencies

More information

Handling Assignment Comp 412

Handling Assignment Comp 412 COMP 412 FALL 2018 Handling Assignment Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp

More information

Static Single Assignment (SSA) Form

Static Single Assignment (SSA) Form Static Single Assignment (SSA) Form A sparse program representation for data-flow. CSL862 SSA form 1 Computing Static Single Assignment (SSA) Form Overview: What is SSA? Advantages of SSA over use-def

More information

The ILOC Virtual Machine (Lab 1 Background Material) Comp 412

The ILOC Virtual Machine (Lab 1 Background Material) Comp 412 COMP 12 FALL 20 The ILOC Virtual Machine (Lab 1 Background Material) Comp 12 source code IR Front End OpMmizer Back End IR target code Copyright 20, Keith D. Cooper & Linda Torczon, all rights reserved.

More information

Procedure and Function Calls, Part II. Comp 412 COMP 412 FALL Chapter 6 in EaC2e. target code. source code Front End Optimizer Back End

Procedure and Function Calls, Part II. Comp 412 COMP 412 FALL Chapter 6 in EaC2e. target code. source code Front End Optimizer Back End COMP 412 FALL 2017 Procedure and Function Calls, Part II Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon, all rights reserved. Students

More information

Compiler Design. Register Allocation. Hwansoo Han

Compiler Design. Register Allocation. Hwansoo Han Compiler Design Register Allocation Hwansoo Han Big Picture of Code Generation Register allocation Decides which values will reside in registers Changes the storage mapping Concerns about placement of

More information

Lexical Analysis. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice.

Lexical Analysis. Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice. Lexical Analysis Note by Baris Aktemur: Our slides are adapted from Cooper and Torczon s slides that they prepared for COMP 412 at Rice. Copyright 2010, Keith D. Cooper & Linda Torczon, all rights reserved.

More information

Computing Inside The Parser Syntax-Directed Translation. Comp 412

Computing Inside The Parser Syntax-Directed Translation. Comp 412 COMP 412 FALL 2018 Computing Inside The Parser Syntax-Directed Translation Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2018, Keith D. Cooper & Linda Torczon, all rights

More information

Parsing III. CS434 Lecture 8 Spring 2005 Department of Computer Science University of Alabama Joel Jones

Parsing III. CS434 Lecture 8 Spring 2005 Department of Computer Science University of Alabama Joel Jones Parsing III (Top-down parsing: recursive descent & LL(1) ) (Bottom-up parsing) CS434 Lecture 8 Spring 2005 Department of Computer Science University of Alabama Joel Jones Copyright 2003, Keith D. Cooper,

More information

What If. Static Single Assignment Form. Phi Functions. What does φ(v i,v j ) Mean?

What If. Static Single Assignment Form. Phi Functions. What does φ(v i,v j ) Mean? Static Single Assignment Form What If Many of the complexities of optimization and code generation arise from the fact that a given variable may be assigned to in many different places. Thus reaching definition

More information

Introduction to Parsing. Comp 412

Introduction to Parsing. Comp 412 COMP 412 FALL 2010 Introduction to Parsing Comp 412 Copyright 2010, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit permission to make

More information

Data Flow Analysis and Computation of SSA

Data Flow Analysis and Computation of SSA Compiler Design 1 Data Flow Analysis and Computation of SSA Compiler Design 2 Definitions A basic block is the longest sequence of three-address codes with the following properties. The control flows to

More information

Fast Copy Coalescing and Live-Range Identification

Fast Copy Coalescing and Live-Range Identification Fast Copy Coalescing and Live-Range Identification Zoran Budimlić, Keith D. Cooper, Timothy J. Harvey, Ken Kennedy, Timothy S. Oberg, and Steven W. Reeves Department of Computer Science Rice University,

More information

Instruction Selection, II Tree-pattern matching

Instruction Selection, II Tree-pattern matching Instruction Selection, II Tree-pattern matching Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 4 at Rice University have explicit permission

More information

The Development of Static Single Assignment Form

The Development of Static Single Assignment Form The Development of Static Single Assignment Form Kenneth Zadeck NaturalBridge, Inc. zadeck@naturalbridge.com Ken's Graduate Optimization Seminar We learned: 1.what kinds of problems could be addressed

More information

Computing Static Single Assignment (SSA) Form. Control Flow Analysis. Overview. Last Time. Constant propagation Dominator relationships

Computing Static Single Assignment (SSA) Form. Control Flow Analysis. Overview. Last Time. Constant propagation Dominator relationships Control Flow Analysis Last Time Constant propagation Dominator relationships Today Static Single Assignment (SSA) - a sparse program representation for data flow Dominance Frontier Computing Static Single

More information

Code Placement, Code Motion

Code Placement, Code Motion Code Placement, Code Motion Compiler Construction Course Winter Term 2009/2010 saarland university computer science 2 Why? Loop-invariant code motion Global value numbering destroys block membership Remove

More information

Grammars. CS434 Lecture 15 Spring 2005 Department of Computer Science University of Alabama Joel Jones

Grammars. CS434 Lecture 15 Spring 2005 Department of Computer Science University of Alabama Joel Jones Grammars CS434 Lecture 5 Spring 2005 Department of Computer Science University of Alabama Joel Jones Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled

More information

A Bad Name. CS 2210: Optimization. Register Allocation. Optimization. Reaching Definitions. Dataflow Analyses 4/10/2013

A Bad Name. CS 2210: Optimization. Register Allocation. Optimization. Reaching Definitions. Dataflow Analyses 4/10/2013 A Bad Name Optimization is the process by which we turn a program into a better one, for some definition of better. CS 2210: Optimization This is impossible in the general case. For instance, a fully optimizing

More information

ECE 5775 (Fall 17) High-Level Digital Design Automation. Static Single Assignment

ECE 5775 (Fall 17) High-Level Digital Design Automation. Static Single Assignment ECE 5775 (Fall 17) High-Level Digital Design Automation Static Single Assignment Announcements HW 1 released (due Friday) Student-led discussions on Tuesday 9/26 Sign up on Piazza: 3 students / group Meet

More information

C-SSAPRE: Eliminating redundancy over copies

C-SSAPRE: Eliminating redundancy over copies C-SSAPRE: Eliminating redundancy over copies Tom Murphy VII, Brendan McMahan 7 May 2003 1 Introduction Partial Redundancy Elimination (PRE) is an optimization that prevents, where possible, the re-computation

More information

Local Register Allocation (critical content for Lab 2) Comp 412

Local Register Allocation (critical content for Lab 2) Comp 412 Updated After Tutorial COMP 412 FALL 2018 Local Register Allocation (critical content for Lab 2) Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2018, Keith D. Cooper & Linda

More information

Lecture 21 CIS 341: COMPILERS

Lecture 21 CIS 341: COMPILERS Lecture 21 CIS 341: COMPILERS Announcements HW6: Analysis & Optimizations Alias analysis, constant propagation, dead code elimination, register allocation Available Soon Due: Wednesday, April 25 th Zdancewic

More information

Construc)on of Sta)c Single- Assignment Form

Construc)on of Sta)c Single- Assignment Form COMP 512 Rice University Spring 2015 Construc)on of Sta)c Single- Assignment Form Copyright 2015, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice University

More information

Syntax Analysis, VII One more LR(1) example, plus some more stuff. Comp 412 COMP 412 FALL Chapter 3 in EaC2e. target code.

Syntax Analysis, VII One more LR(1) example, plus some more stuff. Comp 412 COMP 412 FALL Chapter 3 in EaC2e. target code. COMP 412 FALL 2017 Syntax Analysis, VII One more LR(1) example, plus some more stuff Comp 412 source code IR Front End Optimizer Back End IR target code Copyright 2017, Keith D. Cooper & Linda Torczon,

More information

Optimizer. Defining and preserving the meaning of the program

Optimizer. Defining and preserving the meaning of the program Where are we? Well understood Engineering Source Code Front End IR Optimizer IR Back End Machine code Errors The latter half of a compiler contains more open problems, more challenges, and more gray areas

More information

CS415 Compilers. Procedure Abstractions. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University

CS415 Compilers. Procedure Abstractions. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University CS415 Compilers Procedure Abstractions These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University Where are we? Well understood Engineering Source Code

More information

The View from 35,000 Feet

The View from 35,000 Feet The View from 35,000 Feet This lecture is taken directly from the Engineering a Compiler web site with only minor adaptations for EECS 6083 at University of Cincinnati Copyright 2003, Keith D. Cooper,

More information

Middle End. Code Improvement (or Optimization) Analyzes IR and rewrites (or transforms) IR Primary goal is to reduce running time of the compiled code

Middle End. Code Improvement (or Optimization) Analyzes IR and rewrites (or transforms) IR Primary goal is to reduce running time of the compiled code Traditional Three-pass Compiler Source Code Front End IR Middle End IR Back End Machine code Errors Code Improvement (or Optimization) Analyzes IR and rewrites (or transforms) IR Primary goal is to reduce

More information

The So'ware Stack: From Assembly Language to Machine Code

The So'ware Stack: From Assembly Language to Machine Code Taken from COMP 506 Rice University Spring 2017 The So'ware Stack: From Assembly Language to Machine Code source code IR Front End OpJmizer Back End IR target code Somewhere Out Here Copyright 2017, Keith

More information

The Static Single Assignment Form:

The Static Single Assignment Form: The Static Single Assignment Form: Construction and Application to Program Optimizations - Part 3 Department of Computer Science Indian Institute of Science Bangalore 560 012 NPTEL Course on Compiler Design

More information

Computing Inside The Parser Syntax-Directed Translation, II. Comp 412 COMP 412 FALL Chapter 4 in EaC2e. source code. IR IR target.

Computing Inside The Parser Syntax-Directed Translation, II. Comp 412 COMP 412 FALL Chapter 4 in EaC2e. source code. IR IR target. COMP 412 FALL 20167 Computing Inside The Parser Syntax-Directed Translation, II Comp 412 source code IR IR target Front End Optimizer Back End code Copyright 2017, Keith D. Cooper & Linda Torczon, all

More information

Mechanized Verification of SSA-based Compilers Techniques

Mechanized Verification of SSA-based Compilers Techniques Mechanized Verification of SSA-based Compilers Techniques Delphine Demange - U. Rennes 1 / IRISA / Inria Rennes Joint work with G. Barthe, S. Blazy, Y. Fernandez de Retana, D. Pichardie, L. Stefanesco

More information

Mechanizing Conventional SSA for a Verified Destruction with Coalescing

Mechanizing Conventional SSA for a Verified Destruction with Coalescing Mechanizing Conventional SSA for a Verified Destruction with Coalescing Delphine Demange University of Rennes 1 / IRISA / Inria, France delphine.demange@irisa.fr Yon Fernandez de Retana University of Rennes

More information

Sustainable Memory Use Allocation & (Implicit) Deallocation (mostly in Java)

Sustainable Memory Use Allocation & (Implicit) Deallocation (mostly in Java) COMP 412 FALL 2017 Sustainable Memory Use Allocation & (Implicit) Deallocation (mostly in Java) Copyright 2017, Keith D. Cooper & Zoran Budimlić, all rights reserved. Students enrolled in Comp 412 at Rice

More information

Lecture 3 Local Optimizations, Intro to SSA

Lecture 3 Local Optimizations, Intro to SSA Lecture 3 Local Optimizations, Intro to SSA I. Basic blocks & Flow graphs II. Abstraction 1: DAG III. Abstraction 2: Value numbering IV. Intro to SSA ALSU 8.4-8.5, 6.2.4 Phillip B. Gibbons 15-745: Local

More information

Code Merge. Flow Analysis. bookkeeping

Code Merge. Flow Analysis. bookkeeping Historic Compilers Copyright 2003, Keith D. Cooper, Ken Kennedy & Linda Torczon, all rights reserved. Students enrolled in Comp 412 at Rice University have explicit permission to make copies of these materials

More information

Lecture 23 CIS 341: COMPILERS

Lecture 23 CIS 341: COMPILERS Lecture 23 CIS 341: COMPILERS Announcements HW6: Analysis & Optimizations Alias analysis, constant propagation, dead code elimination, register allocation Due: Wednesday, April 25 th Zdancewic CIS 341:

More information

Goals of Program Optimization (1 of 2)

Goals of Program Optimization (1 of 2) Goals of Program Optimization (1 of 2) Goal: Improve program performance within some constraints Ask Three Key Questions for Every Optimization 1. Is it legal? 2. Is it profitable? 3. Is it compile-time

More information

Intermediate Representation (IR)

Intermediate Representation (IR) Intermediate Representation (IR) Components and Design Goals for an IR IR encodes all knowledge the compiler has derived about source program. Simple compiler structure source code More typical compiler

More information

Register Alloca.on via Graph Coloring

Register Alloca.on via Graph Coloring COMP 512 Rice University Spring 2015 Register Alloca.on via Graph Coloring Beyond Chai,n Briggs Copyright 2015, Keith D. Cooper & Linda Torczon, all rights reserved. Students enrolled in Comp 512 at Rice

More information

Register Allocation 3/16/11. What a Smart Allocator Needs to Do. Global Register Allocation. Global Register Allocation. Outline.

Register Allocation 3/16/11. What a Smart Allocator Needs to Do. Global Register Allocation. Global Register Allocation. Outline. What a Smart Allocator Needs to Do Register Allocation Global Register Allocation Webs and Graph Coloring Node Splitting and Other Transformations Determine ranges for each variable can benefit from using

More information

Preference-Guided Register Assignment

Preference-Guided Register Assignment Preference-Guided Register Assignment Matthias Braun 1, Christoph Mallon 2, and Sebastian Hack 2 1 Karlsruhe Institute of Technology matthias.braun@kit.edu 2 Computer Science Department Saarland University

More information

: Advanced Compiler Design. 8.0 Instruc?on scheduling

: Advanced Compiler Design. 8.0 Instruc?on scheduling 6-80: Advanced Compiler Design 8.0 Instruc?on scheduling Thomas R. Gross Computer Science Department ETH Zurich, Switzerland Overview 8. Instruc?on scheduling basics 8. Scheduling for ILP processors 8.

More information

CS415 Compilers Overview of the Course. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University

CS415 Compilers Overview of the Course. These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University CS415 Compilers Overview of the Course These slides are based on slides copyrighted by Keith Cooper, Ken Kennedy & Linda Torczon at Rice University Critical Facts Welcome to CS415 Compilers Topics in the

More information