Search icon CANCEL
Subscription
0
Cart icon
Your Cart (0 item)
Close icon
You have no products in your basket yet
Arrow left icon
Explore Products
Best Sellers
New Releases
Books
Videos
Audiobooks
Learning Hub
Conferences
Free Learning
Arrow right icon
Arrow up icon
GO TO TOP
LLVM Cookbook

You're reading from   LLVM Cookbook Over 80 engaging recipes that will help you build a compiler frontend, optimizer, and code generator using LLVM

Arrow left icon
Product type Paperback
Published in May 2015
Publisher
ISBN-13 9781785285981
Length 296 pages
Edition 1st Edition
Tools
Arrow right icon
Toc

Table of Contents (11) Chapters Close

Preface 1. LLVM Design and Use FREE CHAPTER 2. Steps in Writing a Frontend 3. Extending the Frontend and Adding JIT Support 4. Preparing Optimizations 5. Implementing Optimizations 6. Target-independent Code Generator 7. Optimizing the Machine Code 8. Writing an LLVM Backend 9. Using LLVM for Various Useful Projects Index

Optimizing SelectionDAG


A SelectionDAG representation shows data and instructions in the form of nodes. Similar to the InstCombine pass in the LLVM IR, these nodes can be combined and optimized to form a minimized SelectionDAG. But, it's not just a DAGCombine operation that optimizes the SelectionDAG. A DAGLegalize phase may generate some unnecessary DAG nodes, which are cleaned up by subsequent runs of the DAG optimization pass. This finally represents the SelectionDAG in a more simple and elegant way.

How to do it…

There are lots and lots of function members (most of them are named like this: visit**()) provided in the DAGCombiner class to perform optimizations by folding, reordering, combining, and modifying SDNode nodes. Note that, from the DAGCombiner constructor, we can guess that some optimizations require alias analysis information:

class DAGCombiner {
SelectionDAG &DAG;
const TargetLowering &TLI;
CombineLevel Level;
CodeGenOpt::Level OptLevel;
bool LegalOperations;
bool LegalTypes...
lock icon The rest of the chapter is locked
Register for a free Packt account to unlock a world of extra content!
A free Packt account unlocks extra newsletters, articles, discounted offers, and much more. Start advancing your knowledge today.
Unlock this book and the full library FREE for 7 days
Get unlimited access to 7000+ expert-authored eBooks and videos courses covering every tech area you can think of
Renews at $19.99/month. Cancel anytime