Intel® Integrated Performance Primitives (Intel® IPP) Developer Guide and Reference

ID 790148
Date 3/22/2024
Public
Document Table of Contents

Dispatching

Intel® IPP uses multiple function implementations optimized for various CPUs. Dispatching refers to detection of your CPU and selecting the corresponding Intel IPP binary path. For example, the ippie9 library in the /bin directory contains the image processing libraries optimized for 64-bit applications on processors with Intel® Advanced Vector Extensions (Intel® AVX) enabled such as the 2nd Generation Intel® Core™ processor family.

A single Intel IPP function, for example ippsCopy_8u(), may have many versions, each one optimized to run on a specific Intel® processor with specific architecture, for example, the 64-bit version of this function optimized for the 2nd Generation Intel® Core™ processor is e9_ippsCopy_8u(), and version optimized for 64-bit applications on processors with Intel® Streaming SIMD Extensions 4.2 (Intel® SSE 4.2) is y8_ippsCopy_8u(). This means that a prefix before the function name determines CPU model. However, during normal operation the dispatcher determines the best version and you can call a generic function (ippsCopy_8u in this example).

Intel® IPP is designed to support application development on various Intel® architectures. This means that the API definition is common for all processors, while the underlying function implementation takes into account the strengths of each hardware generation.

By providing a single cross-architecture API, Intel IPP enables you to port features across Intel® processor-based desktop, server, and mobile platforms. You can use your code developed for one processor architecture for many processor generations.

The following table shows processor-specific codes that Intel IPP uses:

Description of Codes Associated with Processor-Specific Libraries
 IA-32 Intel® architecture  Intel® 64 architecture Windows* Linux* OS

Description

w7

  + +

Optimized for processors with Intel SSE2

 

m7

+ +

Optimized for processors with Intel SSE3

s8

n8

+ +

Optimized for processors with Supplemental Streaming SIMD Extensions 3 (SSSE3)

p8

y8

+ +

Optimized for processors with Intel SSE4.2

g9

e9

+ +

Optimized for processors with Intel® Advanced Vector Extensions (Intel® AVX) and Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI)

h9

l9

+ +

Optimized for processors with Intel® Advanced Vector Extensions 2 (Intel® AVX2)

 

n0

+ +

Optimized for 2nd Generation Intel® Xeon Phi™ Processor

 

k0

+ +

Optimized for processors with Intel® Advanced Vector Extensions 512 (Intel® AVX-512)

Product and Performance Information

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.

Notice revision #20201201