- The Intel® Integrated Performance Primitives (Intel® IPP) is a cross-platform signal
and multimedia software library with a low-level layer that abstracts functionality from the processor underneath.
This library enables you to transparently use recent Intel architecture enhancements such as MMX™
technology, Streaming SIMD Extensions, Streaming SIMD Extensions 2, Streaming SIMD Extensions 3
as well as Itanium architecture and the Intel® XScale™ microarchitecture for your applications.
Supporting a variety of data types and layouts for each function and minimizing data structures,
Intel IPP delivers a rich set of options to choose from while designing and optimizing an application,
without having to write assembly code.
- The Intel IPP library is optimized for the Itanium 2 microarchitecture, Intel® Xeon™,
Pentium III, and Pentium 4 processors plus the Pentium M processor component of Intel®
Centrino™ mobile technology. Intel IPP is also optimized for the Intel®
Personal Client Architecture (Intel® PCA).
With a single application program interface (API) across the range of platforms, you can achieve compatibility and reduced cost of development.
- Intel IPP is compatible with the Windows* and Linux* operating system environments.
These include Microsoft* Windows* 95, 98, 2000, CE, ME, and Windows* NT. Linux support includes Linux* distributions
with 2.x kernels. Special kernel support is required for support of the Pentium 4 processors.
- Intel IPP provides a range of functions for multimedia: audio, video, and speech codecs;
image processing; signal processing; math support routines; and computer vision. You may construct standardized codecs
using Intel IPP including MP3 and AAC audio, H.263, H.264, MPEG-1, MPEG-2 and MP3 and
MPEG-4 video, JPEG and JPEG2000 image, as well as G.723.1 and G.729 speech. You have learnt now what every element
means in the name of any Intel IPP function.
- The Intel IPP functions are organized into three principal domains: operations on one-dimentional (IPPS) and
two-dimensional (IPPI) data, and operations on small matrices (IPPM). There is also a fourth set of functions
that are intended for cryptography operations (IPPCP) and structurally belong to the single processing domain.
- Intel IPP library automatically selects processor-specific data
link libraries (DLLs) and uses architecture-specific instruction sets. For applications working on a known CPU setting and requiring
strict memory control the static merged library is used. When using only a few Intel IPP functions the better chioce is
the libraries statically linked via emereged libraries.
- Memory aligning, thresholding denormal data before FIR filtering, buffer reusing, and using FFT to optimize algorithms
can boost the application performance on your machine.