This course teaches the methodology and tools used by Intel's Application Solution Center (ASC) engineers to identify and eliminate bottlenecks in hardware, operating systems, and applications in order to tune performance. The course combines in-depth instruction with hands-on labs on Intel processor-based systems.

You will learn a complete and time-tested process for performance optimization. The course starts with a high-level overview, then guides you through hardware and operating system tuning, followed by application tuning with better thread management, memory management, and disk I/O.

Along the way you will also get hands-on experience using the tools that will enable you to successfully tune your applications on Windows NT. Performance Monitor will surface the bottlenecks in your hardware, operating system, and applications, while VTune will help you identify the hot spots that make your applications less efficient. While the methodology and tools are appropriate to both current and next-generation Intel architectures, the course includes a close look at the IA-32 Micro-architecture and performance opportunities it provides.


Course Outline

• Performance Terminology & Concepts Overview
• Performance Tuning Methodology
• Using Windows NT Performance Monitor
• System Architecture
  • Overview of the SMP Intel Architecture (IA)
  • SMP System Bus
  • Memory Configuration
  • I/O Buses
  • Disk Architecture
  • Typical Memory Hierarchy
  • Tuning Guide
• Application Performance Issues
  • Typical Application Architecture
  • Application Tuning Guidelines and Recommendations
    • Thread management
    • Memory management
    • Disk I/O
    • System Services
    • Fibers
• IA-32 Micro-Architecture
  • Pentium Pro, Pentium II, Pentium III Processor Architecture Overview
    • Pipeline stages
    • MMX and XMM
    • L1 and L2 cache design
    • Streaming SIMD Extensions
  • Recommended ways of improving performance for C/C++ apps
    • Code optimization techniques
    • Common pitfalls
• Intel’s Proton Compiler
  • Multi-pass optimizations
  • Optimizing for SIMD architectures
  • Processor dispatch and targeting
  • Optimizing for specific applications
  • Code Examples
• Software Performance Tuning with the VTune Performance Analyzer
• Performance and Web Applications
• Putting It All Together