Includes bibliographical references and index.

Vendor-supplied metadata.

Annotation "Power and Performance: Software Analysis and Optimization" is a guide to solving performance problems in modern Linux systems. Power-efficient chips are no help if the software those chips run on is inefficient. Starting with the necessary architectural background as a foundation, the book demonstrates the proper usage of performance analysis tools in order to pinpoint the cause of performance problems, and includes best practices for handling common performance issues those tools identify. Provides expert perspective from a key member of Intel s optimization team on how processors and memory systems influence performancePresents ideas to improve architectures running mobile, desktop, or enterprise platformsDemonstrates best practices for designing experiments and benchmarking throughout the software lifecycleExplains the importance of profiling and measurement to determine the source of performance issues."

""Front Cover""; ""Power and Performance: Software Analysis and Optimization""; ""Copyright""; ""Dedication""; ""Contents""; ""Introduction""; "" Performance Apologetic""; "" Performance Is Power Efficiency""; "" Performance and Usage Patterns""; "" A Word on Premature Optimization""; "" The Roadmap""; "" Reference""; ""Part 1: Background Knowledge""; ""Chapter 1: Early Intel�A� Architecture""; ""1.1 Intel�A� 8086""; ""1.1.1 System State""; ""Address space""; ""1.1.2 Registers""; ""1.1.3 Instructions""; ""Data movement""; ""Integer arithmetic""; ""Boolean logic""; ""Flow control""; ""String""

""1.1.4 Machine Code Format""""1.2 Intel�A� 8087""; ""1.2.1 IEEE 754 Floating Point""; ""Formats, precision, and environment""; ""C99 support""; ""1.2.2 x87 Floating Point""; ""1.3 Intel�A� 80286 and 80287""; ""1.3.1 Protected and Real Mode""; ""1.3.2 Protected Mode Segmentation""; ""1.3.3 Task Control""; ""1.4 Intel�A� 80386 and 80387""; ""1.4.1 32-Bit Mode""; ""1.4.2 Paging""; "" References""; ""Chapter 2: Intel�A� Pentium�A� Processors""; ""2.1 Intel�A� Pentium�A�""; ""2.1.1 Superscalar""; ""2.2 Intel�A� Pentium�A� Pro""; ""2.2.1 PAE""; ""2.2.2 �I?ops ""; ""2.2.3 Out-of-Order Execution""

""Register Renaming""""2.3 Intel�A� Pentium�A� 4""; ""2.3.1 IA-32e Mode""; ""2.3.2 Multi-Core""; ""2.3.3 Intel�A� Hyper-Threading""; "" References""; ""Chapter 3: Intel�A� Core�a?� Processors""; ""3.1 Intel�A� Pentium�A� M""; ""3.1.1 ACPI""; ""P states and Intel SpeedStep�A� Technology""; ""C States""; ""3.2 Second Generation Intel�A� Core�a?�Processor Family""; ""3.2.1 Intel�A� HD Graphics""; ""3.2.2 Intel�A� Flex Memory Technology""; ""3.2.3 Intel�A� Turbo Boost Technology""; ""3.2.4 Intel�A� RAPL""; "" References""; ""Chapter 4: Performance Workflow""; ""4.1 Step 0: Defining the Problem""

""4.2 Step 1: Determine the Source of the Problem""""4.3 Step 2: Determine Whether the Bottleneck Can Be Avoided""; ""4.4 Step 3: Design a Reproducible Experiment""; ""4.5 Step 4: Check Upstream""; ""4.5.1 Who""; ""4.5.2 Where and How""; ""4.5.3 What""; ""Git bisect""; ""Cleaning patches with Git""; ""Sending patches with Git""; ""4.6 Step 5: Algorithmic Improvement""; ""4.7 Step 6: Architectural Tuning""; ""4.8 Step 7: Testing""; ""4.9 Step 8: Performance Regression Testing""; "" References""; ""Chapter 5: Designing Experiments""; ""5.1 Choosing a Metric""

""5.2 Dealing with External Variables""""5.2.1 Controllable External Variables""; ""5.2.2 Uncontrollable External Variables""; ""5.3 Timing""; ""5.3.1 CPU Cycles""; ""5.3.2 Clock Time and Unix Time""; ""5.4 Phoronix Test Suite""; ""5.4.1 Running Phoronix""; ""Configuration""; ""Installing Tests""; ""Running Tests""; ""Batch Mode""; ""5.4.2 Working with Results""; ""5.4.3 Creating Custom Tests""; ""5.4.4 Phoronix Resources""; "" References""; ""Part 2: Monitors""; ""Chapter 6: Introduction to Profiling""; ""6.1 PMU""; ""6.1.1 Event Counters""; ""6.1.2 Using Event Counters""