Skip navigation links
Tools
SDKs
Libraries
Samples & Demos
Docs
Zones
Community
Support
AMD CodeAnalyst Performance Analyzer
Skip Navigation LinksHome > Tools > AMD CodeAnalyst Performance Analyzer

What is AMD CodeAnalyst?

AMD CodeAnalyst Performance Analyzer helps software developers to improve the performance of applications, drivers and system software. Well-tuned software delivers a better end-user experience through shorter response time, increased throughput and better resource utilization. AMD CodeAnalyst is a profiling tool used by developers worldwide to improve the performance of games and media-oriented applications, transaction processing, and high performance scientific/engineering computation.

Key Features & Benefits:

  • Understand OpenCL API execution
  • Optimize heterogeneous computing applications
  • Quickly find performance hot-spots
  • Diagnose performance issues using AMD technology
  • Tune managed (e.g. Java, .Net) and native code (e.g. OpenCL, C, C++, Fortran)
  • Analyze programs on multi-core and NUMA platforms
  • Easy to use graphical user interface and command line tools
  • Analyze programs within Visual Studio (2003, 2005 and 2010)
  • Available for both Linux and Windows platforms

Please post feedback and questions to the CodeAnalyst Forums

Download now! Download now!

What can AMD CodeAnalyst do for you?

matrix_omp_graph.png

Optimize heterogeneous computing applications

Collect and analyze the OpenCL™ API execution performance from both CPUs and GPUs. Use the detailed lists to find calls of interest and use the timeline to understand the entire execution or drill down to the smallest API call details.Learn more in this webinar.

matrix_omp_graph.png

Find time-critical hot-spots in code

CodeAnalyst helps you find the hot-spots which are the best candidates for tuning and optimization. Time-based profiling reveals hot-spots without high overhead. Drill down to functions, source code and even individual instructions, if necessary.

matrix_omp_ipc.png

Diagnose performance issues in hot-spots

Diagnose performance issues using AMD technology. CodeAnalyst uses hardware-level performance counters and instruction-based sampling to provide valuable clues about inefficient program behavior. Use rates and ratios to quickly measure the efficiency of functions, loops and program statements.

matrix_omp_threads.png

Identify thread-affinity and core utilization problems

Capture and display thread profiles using CodeAnalyst. A thread profile shows the behavior of all run-time threads across all cores. The thread chart identifies threads with poor core affinity. Find threads with poor memory locality on non-uniform memory access (NUMA) platforms.

scimark_java_source.png

Analyze just-in-time, managed code

CodeAnalyst is not just for C/C++ and Fortran. Java and .NET developers can analyze the performance of managed (JIT compiled) code. Tune up those web-based applications.

matrix_omp_dual_core.png

Supports multi-core Windows and Linux platforms

CodeAnalyst supports all of the latest AMD processors on both Windows and Linux platforms. CodeAnalyst for Linux is open source and reflects AMD's commitment to the open source development community.

matrix_omp_vs2005.png

Extends Microsoft Visual Studio

Microsoft Visual Studio user can analyze their programs without leaving the Visual Studio environment. The CodeAnalyst Visual Studio plug-in provides all of the profiling features supported by the stand-alone CodeAnalyst for Windows GUI-based tool.

System Requirements

Operating Systems Supported:

  • AMD CodeAnalyst for Windows supports the following platforms:Windows® Server 2008, and Vista®, and Windows 7® distribution on x86 and AMD64 architectures.
  • 32-bit and 64-bit Linux distributions based around the 2.6 kernel series on x86 architecture

General Documentation

Videos

OpenCL Webinar Series: How to use Heterogeneous Features of AMD CodeAnalyst Performance Analyzer to optimize GPGPU programming.

Watch it now

Learn how to use CodeAnalyst in the Visual Studio 2010 IDE.

Learn about ACML-GPU (7:36)