Overview

The AMD APP Profiler is a performance analysis tool that gathers data from the OpenCL™ run-time and AMD Radeon™ GPUs during the execution of an OpenCL™ application. We can then use this information to discover bottlenecks in an application and find ways to optimize the application’s performance for AMD platforms.

• What's New
• Features
• Getting Started
• Support
• Resources
• Download

OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.

What's New

What's New in Version 2.4

• Support for AMD APP SDK v2.6.
• Added a kernel occupancy analyzer, which calculates and displays a kernel occupancy number estimating the number of in-flight wavefronts on a compute unit as a percentage of the theoretical maximum number of wavefronts that the compute unit can support
• Added support for collecting symbol information when collecting an application trace, allowing navigation from the API Trace view to the source code that called an API
• Improved OpenCL™ analysis module:
  • Added detection of non-optimized data transfer operations
  • Added detection of redundant synchronization operations
  • Improved detection of unnecessary blocking write operations
  • Improved analysis in multithreaded applications (fixed false positives)
• Added support for specifying which OpenCL™ APIs will be traced
• Added ability to rename sessions in the Session Explorer Window
• Added ability to automatically delete profiler sessions when closing a Microsoft® Visual Studio® solution
• Added support for modifying the parameters used to initiate a profiler session
• Added support for multiple-GPU systems when collecting performance counters
• Improved the CLPerfMarkerAMD library
• Improved performance when using timeout mode
• In the session window, "GPRs" column has been renamed "VGPRs" (vector GPRs)
• Fixed a problem with loading saved counters from a file
• Fixed a problem where the performance counter values for some kernel dispatch operations were reported as all zeros
• Fixed a problem with missing GPU timestamps in an application trace when enabling the "Write trace data periodically during program execution" option
• Removed Data Transfer data from the Session view for OpenCL™ applications. It is recommended that you use the Application Trace view to get information on data transfers
• Preview: Support for profiling with AMD Radeon™ HD7000 series GPUs (requires AMD APP SDK v2.6 and an AMD Catalyst version that supports this hardware)

Features

• Collect OpenCL™ Application Trace
  • View and debug the input parameters and output results for all OpenCL™ API calls
  • Search the API calls
  • Navigate to the source code that called an OpenCL™ API
  • Specify which OpenCL™ APIs will be traced
• Collect GPU Performance Counters of AMD Radeon™ graphics cards
  • Show kernel resource usages
  • Show the number of instructions executed by the GPU
  • Show the GPU utilization
  • Show the GPU memory access characteristics
  • Measure kernel execution time
• OpenCL™ Timeline visualization
  • Visualize the application high level structure
  • Visualize kernel execution and data transfer operations
  • Visualize host code execution
  • Annotate host code in the timeline with performance markers using the included CLPerfMarkerAMD library
• OpenCL™ Application Summary pages
  • Find incorrect or inefficient usage of the OpenCL™ API using the OpenCL™ analysis module
  • Find the API hotspots
  • Find the bottleneck between kernel execution and data transfer operations
  • Find the top 10 data transfer and kernel execution operations
• OpenCL™ Kernel Occupancy Viewer
  • Calculates and displays a kernel occupancy number, which estimates the number of in-flight wavefronts on a compute unit as a percentage of the theoretical maximum number of wavefronts that the compute unit can support
  • Find out which kernel resource (GPR usage, LDS size, or Work-group size) is currently limiting the number of in-flight wavefronts
  • Displays graphs showing how kernel occupancy would be affected by changes in each kernel resource
• Display the AMD IL and ISA (hardware disassembly) code of the kernel for OpenCL™ kernels and DXASM code for DirectCompute kernels.

Getting Started

Requirements

• Microsoft Windows Vista or 7 (32 bit/64 bit) or Linux 32-bit/64-bit
• [Optional] Microsoft Visual Studio 2008 and 2010 (Standard/Professional/Team System Edition)
• To profile OpenCL™ applications:
  • AMD APP SDK v2.6 or later
  • [GPU device] AMD Catalyst with OpenCL™ GPU support (11.7 or newer)
  • [GPU device] AMD Radeon™ HD 4000 series or newer
• To profile DirectCompute applications:
  • Microsoft DirectX run-time (June 2010 or later)
  • AMD Radeon™ HD 5000 series or newer

AMD APP Profiler Screenshot


Figure 1: Click to enlarge.


Support

• AMD APP Profiler Knowledge Base
• AMD APP Profiler Help Desk
• GPU Developer Tool Forum

Resources

• AMD APP Profiler User Guide
• OpenCL™ Programming Guide
• AMD APP Profiler's article at Siggraph 2010
• OpenCL™ Zone

Download