Intel® - Integrated Performance Primitives (Intel® IPP)

	Intel® Software Development Products for Intel® Platforms and Technologies

Intel® Integrated Performance Primitives 5.3 for Windows, Linux, and Mac OS* X

Overview	New In This Release	Print
Features	Compatibility
Performance-Optimized Functions	System Requirements

Overview

Intel® Integrated Performance Primitives 5.2<br>for Windows*, Linux*, and Mac OS*

Deliver rich multimedia experiences to your customers with Intel® Integrated Performance Primitives (Intel® IPP) — a library of thousands of highly optimized software functions for multimedia and data processing applications. Intel IPP 5.3 brings expanded functionality in data compression, image processing and more.

Intel IPP is one of the Intel® Performance Libraries products, which provide optimized software building blocks to complement Intel’s optimizing compilers and performance optimization tools. Intel IPP is available as a standalone product, or with the Intel® Compiler Professional Editions for a more complete and cost-effective solution.

Features

Multi-Core Processor Support

Figure 1. Multi-Core Processors
Multi-Core Processors Enable the True Parallel Execution of Multi-Threaded Software Applications
	Multi-Core Optimized Functions: Many critical functions for Vector and Statistical mathematics, Signal Filtering, Fourier Transforms, Image/JPEG Compression and Color Conversion are internally threaded using OpenMP* for the highest performance on multi-core systems.
	Multi-Core Optimized Code Samples: Many of the Intel IPP code samples are threaded to illustrate the effective use of Intel IPP functions in applications such as video encoding and decoding.
	Fully Thread-Safe Functions: All Intel IPP functions are fully thread-safe, simplifying integration into threaded applications.
To learn more about threading and Intel IPP functions visit our Threading Frequently Asked Questions (FAQ) page.

Performance-Optimized Functions

Intel IPP functions are designed to deliver performance beyond what optimized compilers alone can deliver, by matching the function algorithms to low-level optimizations based on the processor’s available features such as Streaming SIMD Extensions (SSE) and other optimized instructions sets.

Intel® IPP Function Domain List

Vector/Matrix Operations

Color Conversion

Audio Coding

Cryptography

Ray-Tracing/Rendering (NEW!)

Video Coding: Key algorithmic components for DV25/50/100, MPEG-2, MPEG-4, H.263, and MPEG-4 Part 10 (H.264) codecs. Figure 2 shows where Intel IPP video coding components (represented by the blue boxes) fit into the H.264 codec process flow. Functions include:
	Motion Compensation
	Motion Estimation
	Modified Discrete Cosine Transforms
	Quantization and Inverse Quantization
	Entropy Coding

Figure 2. Intel® Integrated Performance Primitives (Intel® IPP) Components in the H.264 Codec Process Flow
The Video and Audio code samples illustrate sample codec implementations using Intel IPP functions.

Back to Function Domain List

Image and 2D Signal Processing: Intel IPP is the premier library of image and 2D signal processing algorithms, and includes a rich selection of algorithms operating on images and regions within images (ROIs):

Transforms:
	Wavelet
	Fourier (FFT/DFT, real/complex)
	Windowing (Hamming, Bartlett)
	Discrete Cosine (DCT)

Image Statistics:
	Sum, Integral, Tilted Integral
	Mean, Min, Max, Histogram, StDev
	Image Moments
	Image Norms (L1, L2, infinity)
	Image Quality Index calculation
	Proximity Measures (Cross-corrrelations, Square Distance)
	Threshold/Compare Operations

Filtering Functions:
	General linear filtering
	Convolution/Deconvolution (LR and FFT)
	Box, Min, Max, Median
	Wiener Filters
	Fixed Filters (Prewitt, Sobel, Laplace, Gauss, Scharr, Roberts)
	Sharpening/Hipass/Lowpass Filters

Image Arithmetic/Logic Operations:
	Alpha composition
	Arithmetic operations
	(add/sub/mul/div/ sqrt/sqr/ln/exp/abs)
	Logical operations (AND, OR, XOR, Shift, NOT)

Geometric Transforms:
	Resize, Mirror, Rotate,, Shear
	Affine transforms
	Perspective transformations
	Bilinear warping
	Coordinate remapping

Image Data Exchange/Initialization:
	Copy/Set/Transpose
	Channel swapping
	Jaehne/Ramp/Zigzag initialization
	Memory allocation for multiple image types

Back to Function Domain List

Computer Vision: Includes optimized functions for many key computer-vision operations, for applications in security, machine control, media management, media annotation and more:

Feature Detection (Corner, Canny Edge Detection)

Pattern Recognition (Haar classifiers)

Distance Transforms

Pyramid functions (Gaussian/Laplacian pyramids)

Image Gradients

Universal Pyramid functions

Flood Filling

Camera Calibration

Motion Templates generation

3D Reconstruction[ST2]

Optical Flow calculation (Lucas-Kanade)

Back to Function Domain List

Color Conversion:Intel IPP provides a rich set of optimized color-conversion routines on 32/24/16-bit-per-pixel formats:

Color Module Conversion:
	RGB, YUV, YCbCr, BGR, CbYCr, HSV, LUV, Lab, YCC, HLS, SBGR, YCoCg, YCCK, XYZ, CMYK

Lookup Table Conversions (Linear/Cublic/Palette)

Color to Greyscale Conversions (Fixed/Custom coefficients)

Image Bit Resolution Reductions

Color Format Conversions:
	YCbCr422, YCbCr420, YCbCr411, CbYCr422, BGR565, BGR555, BGR565Dither

Color Twist Conversions (integer/float pixel values)

Gamma Corrections (Forward/Backward)

Back to Function Domain List

String Processing: Build optimized text database management, search and retrieval, or document indexing processing into your applications using Intel IPP’s optimized string operations.

Substring substitution/insertion

String/Substring matching

String concatenation/splitting

Pyramid functions (Gaussian/Laplacian pyramids)

Upper/lower case conversions

Hash value calculation

Back to Function Domain List

JPEG Coding: Key algorithmic components for JPEG, JPEG 2000, and Motion JPEG codecs. Figure 3 shows where Intel IPP JPEG coding components (represented by the blue boxes) fit into the JPEG and JPEG 2000 codec process flow.

Figure 3. Multi-Core Processors Components in the JPEG And JPEG 2000 Codec Process Flow

Back to Function Domain List

Speech Coding: Includes functions for the following:

G.722.1

G729

G.722 Sub-Band ADPCM

GSM/AMR

G.723.1

AMR-Wideband

G.726

GSM Full Rate

G.728

Companding

Echo Cancellation

Back to Function Domain List

Signal Processing: Includes signal processing features for the following:

Filtering and Convolution:
	Finite Impulse Response (FIR)
	Infinite Impulse Response (IIR)
	Median filtering
	Cyclic convolution
	Auto/Cross-correlation

Array/Signal Initialization/Manipulation:
	Move/Copy/Set/Zero
	Tone/Triangle/Ramp/Jaehne Generation
	Random vector generation (Uniform/Gaussian)
	Array allocation
	Real/Complex conversion
	Polar/Cartesian conversion

Transforms:
	Fourier (FFT, DFT, Goertzel)
	Discrete Cosine Transforms (DCT)
	Hilbert Transforms
	Wavelet Transforms (fixed/custom filters)
	Power spectrum calculation

Array/Signal statistics :
	Sum/Max/Min/Mean/StdDev/Norm
	Dot products
	Thresholding
	Viterbi decoding

Windowing/Sampling:
	Upsampling/Downsampling
	Windowing (Bartlett/Blackman/ Hamming/Hann/Kaiser)

Array Arithmetic/Logic Operations:
	Arithmetic operations (add/sub/mul/div/ sqrt/sqr/ln/exp/abs)
	Logical operations (AND, OR, XOR, Shift, NOT)
	Array sorting
	Magnitude/Phase

Back to Function Domain List

Data Compression:In addition to video, audio and image compression with codecs, Intel IPP provides functions for lossless compression menthods, such as those used in the popular “zlib” (inflate and deflate) and “libbzip2” libraries:

Burrows-Wheeler-Transform techniques:
	Burrows-Wheeler Transform (BWT)
	Generalized Interval Transform
	Move-to-Front (MTF)
	Run-length encoding (RLE)

Entropy coding:
	Huffman coding
	Variable-Length Coding (VLC)

Dictionary-based Compression:
	LZSS encode/decode
	LZ77 encode/decode

Back to Function Domain List

Audio Coding: Key algorithmic components for MP3, AAC, and AC3 codecs. Figure 4 shows where Intel IPP JPEG coding components (represented by the blue boxes) fit into the AAC codec process flow. Functions include:
	Huffman Coding
	Motion Estimation
	Modified Discrete Cosine Transforms
	Block Filtering
	Frequency Domain Prediction
	Spectral Band Replication
	Fast Fourier Transforms

Figure 4. Intel® Integrated Performance Primitives (Intel® IPP) Components in the AAC Codec Process Flow
The Video and Audio code samples illustrate sample codec implementations using Intel IPP functions.

Back to Function Domain List

Speech Recognition: Build advanced speech recognition, Voice-over-IP, and voice annotation capabilities in applications, using Intel IPP’s broad range of speech-recognition capabilities:

Feature Processing

Acoustic Echo Cancellation (AEC)

Model Evaluation

Polyphase Resampling

Model Estimation

Advanced Aurora Functions

Model Adaptation

Ephraim-Malah Noise Suppression

Vector Quantization

Voice Activity Detection

Back to Function Domain List

Vector/Matrix Operations: Contains a rich set of matrix and vector operations fir a wide variety of applications, including physics modeling and 3D transform/lighting calculations:

Matrix Algebra :
	Eigenvalue/eigenvector calculation
	Least Squares (QR decompositions/back-sub)
	Linear Systems (LU/Cholesky)
	Region-of-Interest (ROI) extraction
	Fast copy of vectors/matrices

Vector Algebra :
	Dot products
	L2 norm calculation
	“saxpy” (ax + y) operations
	Linear combination (ax + by)
	Power/root functions
	Exponential/Logarithmic/Erf/Erfc functions
	Trigonometric/Hyperbolic functions
	Polar/Cartesian conversion

Note: For applications requiring high-performance linear algebra operations on very large data sets, the Intel® Math Kernel Library may also be of interest.

Back to Function Domain List

Cryptography: Intel IPP to quickly build robust, high-performance cryptographic modules and applications. Below are some of the many cryptographic building blocks included in Intel IPP’s cryptography functions.

Symmetric Ciphers:
	Block Ciphers (AES/Rijndael, DES, Triple DES, Blowfish, Twofish)
	Stream Ciphers (ARCFour)

Asymmetric Cryptography:
	Elliptic curve cryptography (GF(p) and GF(2m)
	RSA algorithm (RSA-OAEP, RSA-SSA)
	Discrete-Logarithm Cryptography
	Big-Number arithmetic
	Montgomery reduction
	Pseudo-random number generation
	Prime number generation

One-way Hashing:
	Generalized Hashing (MD5, SHA1-512)
	Mask Generation (MD5, SHA1-512)

Data Authentication:
	Keyed Hash (HMAC-MD5, HMAC-SHA1-512)
	Data Authentication Functions (DES, TDES, Rijndael, Blowfish, Twofish)

Intel IPP’s cryptography algorithms have been formally validated under the Cryptographic Algorithm Validation Program (CAVP):

Name:
	Discrete Logarithm (Digital Signature Standard)
	Integer Factoring (Digital Signature Standard)
	Eliptic Curve Digital Signature Algorithm (ECDSA)
	Random Number Generator

Standards/Certificate:
	FIPS 186-2 / Cert 190
	FIPS 186-2, ANSI X9.31-1998 / Cert 181
	FIPS 186-2, ANSI X9.62-1998 / Cert 40
	FIPS 186-2 / Cert 245

Note: To access the Cryptography Library, you must request access from Intel.

Back to Function Domain List

Ray-Tracing and Rendering: Core operations used in ray-tracing, realistic image rendering, and physics applications:

Bounding-box calculations

Object-ray intersections

Shadow/Reflection calculations

Back to Function Domain List

New In This Release

Expanded Data Compression Functions and Code Samples
	The Data Compression code samples have been expanded, and now include compression/decompression utilities that are command-line compatible with gzip and bzip2. Scripting applications can use these sample applications to accelerate server-side data management applications, and more.
	In addition, Intel IPP 5.3 now includes full support for Run-Length Encoding (RLE) data compression.
Optimized for the 45-nm Intel® Core™ 2 Processor Family
	Intel IPP 5.3 delivers new performance optimizations specially targetted at the 45-nm Intel® Core™2 processor family. The 45-nm Intel Core 2 processor family includes new features such as Intel® Streaming SIMD Extensions 4 (Intel® SSE4), a set of 47 new instructions representing the most significant improvement to the Intel® 64 instruction set architecture since 2001.
Optimized for the 45-nm Intel® ultra-mobile processor technologies
	Intel IPP 5.3 also adds new performance optimizations specially targetted at the 45-nm “Silverthorne” ultra-mobile processor.
New Operating System Support
	Intel IPP 5.3 expands the range of supporting operating systems, including: Apple* Mac OS X 10.5 Ubuntu* and Debian* Linux distributions
JPEG and JPEG2000 Codec support for Microsoft .NET*
	Intel IPP 5.3 now provides code samples for JPEG and JPEG2000 encoding/.decoding under the Microsoft .NET framework.

Compatibility

Operating Systems
Intel IPP version 5.3 expands on previous versions, by adding new support Windows* Vista*, and 64-bit Mac OS* X applications. Code re-use and cross-platform operating system development is simplified by having one library API for Windows*, Linux* and Mac OS* X.

Development Environments
Intel IPP is fully compatible with other development tools from Intel, such as compilers, performance and threading analyzers, and other Intel® Performance Libraries. In addition, Intel IPP is easily used and integrated with popular development tools and environments, such as Microsoft Visual Studio*, Xcode*, Eclipse*, and the GNU Compiler Collection (GCC).

Processors

	Multi-core processors, including Intel® Core™ 2 Quad/Extreme Editions, Intel® Core™ 2 Duo, Intel® Core™ Duo, Intel® Xeon® and Intel® Pentium® D processors.
	Intel®64-based systems, including Intel Core 2 processors, Intel Xeon processors and Intel Pentium D processors, and compatible AMD* processors.
	IA-32 Intel Architecture systems, including Intel® Core™ processor family, Intel® Pentium® processors, and compatible AMD* processors.
	Intel XScale® Microarchitecture-based processors, including Intel® IXP4xx processors.

System Requirements

	Application Target Platforms
	Installation Requirements

Application Target Platforms

Please refer to the section below that matches your application's target computing platform:

32-bit Intel Architecture Platforms

	Intel® Core™2 processor family (including 45nm processors)
	Intel® Core Processor Family
	Intel® Pentium® M processors
	Intel ultra-mobile processors (A100, A110, and 45 nm processors)
	Intel® Pentium® D processor
	Intel® Pentium® 4 processor
	Intel® Xeon® processors (3000, 5000, 7000 series)
	Other compatible processors, e.g. AMD Athlon/Opteron processors

64-bit Intel� 64-based platforms

	Intel Core 2 processor family (including 45nm processors)
	Intel ultra-mobile processors (45 nm processors)
	Intel® Pentium® D processor
	Intel® Pentium® 4 processor
	Intel® Xeon® processors (3000, 5000, 7000 series)
	Other compatible processors, e.g. AMD Athlon/Opteron processors

64-bit Intel® Itanium® 2 processor Family Platforms

Intel® Itanium® 2 processors

Intel® XScale® Microarchitecture-based Platforms

Intel® IXP4xx Processors with Intel XScale® technology

32-bit IA-32 Intel® Architecture-Based and Compatible Platforms

(Intel® Core™ 2 Processor Family, Intel® Core™ Processor Family, Intel® Pentium® Processors, and compatible AMD* processors)

OS Version

Supported Compilers

Microsoft Windows*

	Windows* XP, SP1 or SP2
	Windows Vista*
	Windows Server* 2003
	Windows CE 5.0

	Intel® C++ Compiler Versions 10.1, 10.0, 9.1 for Windows
	Microsoft Visual C++* 2003, 2005
	Microsoft Visual Studio* 6.0
	Microsoft eMbedded Visual C++ 4.0 with Service Pack 4

Linux*

	Ubuntu* 6.10, 7.1
	Debian 3.1r5, 4.0
	Red Hat Fedora Core 4, 5
	Red Hat Enterprise Linux* 3, 4 and 5
	SUSE Linux Enterprise Server* 9 or 10
	Red Flag DC Server 5.0
	Mandriva 10.1
	Monta Vista* 3.1, 4.0
	Linux Standard Base 3.1

	Intel C++ Compiler version 8.1 for Linux*
	Intel C++ Compiler 9.0 for Linux or later
	Intel C++ Compiler 10.0, 10.1 for Linux
	Linux Developer Tools component installed, including gcc, g++, and related tools

Mac OS* X

Mac OS X 10.4.3 or later

	Intel C++ Compiler 10.1 for Mac OS X, including Professional Edition
	GNU Compiler Collection 4.0 or later

64-bit Intel® 64 Architecture-Based and Compatible Platforms

(Intel® Core™ 2 Processor Family, Intel® Pentium® D Processors, Intel® Xeon® Processors, and processors from AMD compatible with the Intel 64 architecture)

OS Version

Supported Compilers

Microsoft Windows*

	Windows* XP 64-bit Edition
	Windows* Vista

	Intel® C++ Compiler 10.1 and 10.0 for Windows
	Windows Server* 2003 with Service Pack 1 SDK

Linux*

	Ubuntu 6.10, 7.10
	Debian 3.1r5, 4.0
	Red Hat Fedora Core 4, 5
	Red Hat Enterprise Linux* 3, 4, and 5
	SUSE Linux Enterprise Server* 9 or 10
	Red Flag DC Server 5.0
	Mandriva 10.1
	Linux Standard Base 3.1

	Intel C++ Compiler 10.1 and 10.0 for Linux, for processors with Intel®64
	Linux Developer Tools component installed, including GCC, g++, and related tools

Mac OS* X

Mac OS X 10.4.3 or later

	Intel C++ Compiler 10.1 and 10.0 for Mac OS X
	GNU Compiler Collection 4.0 or later

64-bit Intel® Itanium® 2 Processor family Platforms

OS Version

Supported Compilers

Microsoft Windows*

Windows Server* 2003

	Intel C++ Compiler 10.1 and 10.0 for Windows
	Microsoft Platform SDK February 2003

Linux

	Ubuntu 6.10, 7.10
	Debian 3.1r5, 4.0
	Red Hat Enterprise Linux 3, 4 and 5
	SUSE Linux Enterprise Server* 9 or 10
	Mandriva 10.1
	SGI Propack 4.0, 5.0
	Linux Standard Base 3.1

	Intel C++ Compiler 10.1 and 10.0 for Linux
	Linux Developer Tools component installed, including gcc, g++, and related tools

Embedded Applications on Intel XScale® Microarchitecture-based Platforms

(Intel® IXP4xx Product Line of Network Processors)

OS Version

Supported Compilers

Microsoft Windows*

Windows CE 5.0

Microsoft eMbedded Visual C++* 4.0 with Service Pack 4

Linux

	MontaVista Linux* 3.0 CEE LE
	MontaVista Linux 3.1 Pro BE

	iwmmxt_le-gcc (MontaVista) for LE applications
	xscale_be-gcc (MontaVista) for BE applications

Installation Requirements

Supported Software / Hardware

Operating Systems

*Microsoft Windows:**
	Windows Vista
	Windows XP, SP1 or SP2
	Windows Server* 2003, SP1 or SP2
*Linux:**
	Ubuntu 6.10, 7.10
	Debian 3.1r5, 4.0
	Red Hat Fedora Core 4, 5
	Red Hat Enterprise Linux 3, 4 and 5
	SUSE Linux Enterprise Server* 9 or 10
	Red Flag DC Server 5.0
	Mandriva 10.1
*Mac OS X:**
	Versions 10..4.3 to10.4.6 for Intel®
	Versions 10.5 for Intel®

Minimum Processor Requirements

IA-32 Intel® Architecture & compatible platforms:
	Intel® Pentium® III processor or later, 400MHz or faster
	Any Intel® Core™ Family processor
	Any Intel® Xeon® processor
	Compatible AMD* processors
Intel® 64 architecture or compatible platforms:
	Any Intel® Core™ Family processor
	Any Intel® Pentium® D processor
	Any Intel® Xeon® processor
	Compatible AMD* processors
Intel® Itanium® 2 Platforms:
	Any Intel Itanium 2 processor

Minimum Disk Space for Install

400 MB for each application target platform

Support

Every purchase of an Intel® Software Development Product includes a year of support services, which provides access to Intel® Premier Support and all product updates during that time. Intel Premier Support gives you online access to technical notes, application notes, and documentation. Install the product, and then register to get support and product update information.

Intel provides both the tools and support to enhance the performance, functionality and efficiency of software applications.

Compatible with leading Windows* and Linux* development environments, Intel® Software Development Products are the fastest and easiest way to take advantage of the latest features of Intel processors. Intel Software Development Products are designed for use in the full development cycle, and include Intel® Performance Libraries, Intel Compilers (C++, Fortran for Windows, Linux, and Mac OS* X), Intel® VTune™ Analyzer, and Intel® Threading Tools and Intel® Cluster Tools.

The Intel® Premier Customer Support Web site provides expert technical support for all Intel software products, product updates and related downloads. For additional product information visit: www.intel.com/software/products.

Intel, the Intel logo, Itanium, Pentium, Intel Centrino, Intel Xeon, Intel XScale, VTune, Celeron, Intel NetBurst, and MMX are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries.

*Other brands and names may be claimed as the property of others. Visit our Legal Information Web site for more information.