iOS and macOS Performance Tuning

In iOS and macOS(TM) Performance Tuning, Marcel Weiher drills down to the code level to help you systematically optimize CPU, memory, I/O, graphics, and program responsiveness in any Objective-C, Cocoa, or CocoaTouch program.

This guide focuses entirely on performance optimization for macOS and iOS. Drawing on more than 25 years of experience optimizing Apple device software, Weiher identifies concrete performance problems that can be discovered empirically via measurement. Then, based on a deep understanding of fundamental principles, he presents specific techniques for solving them.

Weiher presents insights you won't find anywhere else, most of them applying to both macOS and iOS development. Throughout, he reveals common pitfalls and misconceptions about Apple device performance, explains the realities, and helps you reflect those realities in code that performs beautifully.

• Understand optimization principles, measurement, tools, pitfalls, and techniques
• Recognize when to carefully optimize, and when it isn't worth your time
• Balance performance and encapsulation to create efficient object representations, communication, data access, and computation
• Avoid mistakes that slow down Objective-C programs and hinder later optimization
• Fix leaks and other problems with memory and resource management
• Address I/O issues associated with drives, networking, serialization, and SQLite
• Code graphics and UIs that don't overwhelm limited iOS device resources
• Learn what all developers need to know about Swift performance
  This book's source code can be downloaded at github.com/mpw/iOS-macOS-performance.

Register your product at informit.com/register for convenient access to downloads, updates, and corrections as they become available.

Autorentext

Marcel Weiher is a software engineeer and researcher with more than 25 years of experience with Cocoa-related technologies. Marcel's work has always been performance-focused, ranging from solving impossible pre-press problems on the machines of the day via optimizing one of the world's busiest web properties at the BBC to helping other Apple engineers improve the performance of their code on Apple's Mac OS X performance team.

In addition to helping established companies and startups create award-winning software and turn around development teams, Marcel also teaches, blogs, speaks at conferences, contributes to open source, and invents new techniques such as Higher Order Messaging. He also works on programming languages, starting with an Objective-C implementation in 1987 and culminating in the Objective-Smalltalk architecture research language. Marcel currently works as a principal software engineer at Microsoft Berlin and maintains his own software and consulting company, metaobject ltd.

Inhalt

About the Author xv

Introduction xvii

Chapter 1: CPU: Principles 1

A Simple Example 2

The Power of Hybrids 10

Trends 11

Cost of Operations 12

Computational Complexity 14

Summary 16

Chapter 2: CPU: Measurement and Tools 17

Command-Line Tools 18

Xcode Gauges 22

Instruments 22

Internal Measurement 35

Dtrace 38

Optimization Beyond the Call of Duty 38

Summary 39

Chapter 3: CPU: Pitfalls and Techniques 41

Representation 41

Objects 48

Messaging 64

Methods 71

Multicore 72

Mature Optimization 75

Chapter 4: CPU Example: XML Parsing 79

An HTML Scanner 80

Mapping Callbacks to Messages 83

Objects 85

Objects, Cheaply 87

Evaluation 90

Tune-Ups 93

Optimizing the Whole Widget: MAX 94

MAX Implementation 96

Summary 97

Chapter 5: Memory: Principles 99

The Memory Hierarchy 99

Mach Virtual Memory 105

Heap and Stack 106

Resource Management 113

Summary 117

Chapter 6: Memory: Measurement and Tools 119

Xcode Gauges 119

Command-Line Tools 120

Internal Measurement 125

Memory Instruments 126

Summary 136

Chapter 7: Memory: Pitfalls and Techniques 137

Reference Counting 137

Foundation Objects vs. Primitives 141

Smaller Structures 142

Memory and Concurrency 146

Architectural Considerations 147

Temporary Allocations and Object Caching 151

NSCache and libcache 153

Memory-Mapped Files 153

madvise 156

iOS Considerations 157

Optimizing ARC 157

Summary 160

Chapter 8: Memory Example: FilterStreams 161

Unix Pipes and Filters 161

Object-Oriented Filters 163

DescriptionStream 164

Stream Hierarchy 170

Summary 171

Chapter 9: Swift 173

Swift Performance: Claims 173

Assessing Swift Performance 177

Larger Examples 188

Compile Times 191

Optimizer-Oriented Programming 195

A Sufficiently Smart Compiler 197

Summary 204

Chapter 10: I/O: Principles 205

Hardware 205

Operating System 208

Summary 214

Chapter 11: I/O: Measurement and Tools 215

Negative Space: top and time 216

Summary Information: iostat and netstat 217

Instruments 218

Detailed Tracing: fs_usage 221

Summary 224

Chapter 12: I/O: Pitfalls and Techniques 225

Pushing Bytes with NSData 225

Unixy I/O 230

Network I/O 232

Serialization 242

CoreData 253

SQLite 261

Event Posting 264

Segregated Stores 265

Summary 266

Chapter 13: I/O: Examples 267

iPhone Game Dictionary 267

Fun with Property Lists 271

Comma-Separated Values 282

Public Transport Schedule Data 283

Faster CSV Parsing 288

Summary 293

Chapter 14: Graphics and UI: Principles 295

Responsiveness 295

Software and APIs 296

Quartz and the PostScript Imaging Model 299

OpenGL 300

Metal 301

Graphics Hardware and Acceleration 301

From Quartz Extreme to Core Animation 305

Summary 308

Chapter 15: Graphics and UI: Measurement and Tools 309

CPU Profiling with Instruments 310

Quartz Debug 311

Core Animation Instrument 312

What Am I Measuring? 318

Summary 323

Chapter 16: Graphics and UI: Pitfalls and Techniques 325

Pitfalls 325

Techniques 326

Too Much Communication Slows Down Installation 327

Overwhelming an iPhone 330

It's Just an Illusion 332

Line Drawing on iPhone 338

Summary 341

Chapter 17: Graphics and UI: Examples 343

Beautiful Weather App 343

Wunderlist 3 350

Summary 357

Index 359