AMD Logo AMD Developer Central

Framewave Project
Skip Navigation LinksHome > CPU Tools > AMD Performance Libraries > Framewave Project

Framewave Banner

Framewave revolutionizes the way silicon manufacturers deliver performance and optimization tools into the hands of software developers. Sponsored by AMD, the Framewave project offers developers unparalleled, code-level access to a vast array of signal and image functions and routines designed with "code faster, faster code" in mind.

Overview
Derived from the AMD Performance Library, Framewave is a free and open-source collection of popular image and signal processing routines designed to accelerate application development, debugging, and optimization on x86-class processor platforms.
  • API compatible with the Intel® Integrated Performance Primitives (Intel® IPP)
  • New! Mac OS support
  • New! MPEG1 and MPEG2 Decode support
  • Open source — Developers are no longer beholden to a proprietary code base
  • Self-supporting community of contributors and enthusiasts
  • Apache licensing model — create derivative and even proprietary works or gain personal fame by contributing to the Framewave project
  • ISVs can implement just those functions and routines needed to ensure optimal performance benefits
  • Better application performance on x86 class processors
  • Compressed development time and time-to-market

Re-factored to use SSEPlus. Please follow the SSEPlus for additional details.

  • » Framewave project on SourceForge.net
  • » Framewave documentation
  • » SSEPlus project on SourceForge.net
  •  

     

     


     

     

    Framewave News Feeds

     Quick Poll

    How were you introduced to Framewave?
     Submit
     Clear
     Results

    Now available: Framewave 1.3.1

    Some highlights for the 1.3.1 release:

    » Introduced Image Statistical functions like Min, Max, MinMax, Sum, mean and CountInRange
     
    » Important Bug fixes

    » Download | Release Notes

    Some highlights for the 1.3 release:

    » Enhanced color space conversion using 3D look up, in terms of space usage optimization and improved performance
     
    » Introduced memory management convenience functions (malloc), in Signal library

    » Introduced convert functions in Image and Signal:
    - involving 32f to 8u, 8s, 16u, 16s
    - involving unsigned 8, 16, 32 to 16s, 32s, 32f and 64f

    » Advanced optimization of JPEG routines involving:
    - RGB to Y and BGR to Y
    - YCbCr 411, 422, 444 to RGB and BGR MCU functions
    - RGB to YCbCr 422 MCU functions

    Retrieving Data
    Retrieving Data
    AMD Libraries Blog
    » SuperComputing 2009 - Day 2
    Thursday, November 19, 2009
    » SuperComputing 2009
    Wednesday, November 18, 2009
    » Framewave Multipass Build System
    Monday, September 07, 2009
    » ACML 4.3.0 Performance Data
    Tuesday, June 30, 2009
    » See More