Intel Integrated Performance Primitives

Intel Integrated Performance Primitives

Origin of development	Intel
The latest edition	It is (2016)[1] (one year ago) for 2017 Update 1 / 2,016 years
Classification	Library
License	Proprietary
Official site	Intel®Integrated Performance Primitives (Intel®IPP)
I display a template

Intel Integrated Performance Primitives (IPP) is software library consisting of the basic function group optimized for multimedia, image processing, signal processing. The library supports Intel processor and a compatible processor and is available on Windows, Linux, macOS, each platform of the Android operating system.

Table of contents

Dispatch function

Plural IPP support a dynamic dispatch function to choose with the thing which is most suitable for run time depending on environment from the plural most suitable cords implemented by every processor of the generation [²]. For example, I carry out the cord which used AVX with the processor supporting AVX order with a cord using SSE/SSE2 with the processor supporting SSE/SSE2 order. Dynamic in IPP; when link, the module (with the Windows version DLL) of the joint ownership library which is most suitable for run time automatically is chosen, but it is necessary to call ippInit() function explicitly beforehand when link static [³]. It is an entry point of the application program (main function), but, as for the timing to call ippInit(), ippInit() summonses by the startup process (by Windows DllMain function) of the user library may be usually necessary when I use IPP in the inside of the dynamic user library linked to [⁴].

Multi-thread-adaptive

There are a multi-thread version and singles red version in IPP. The multi-thread version can draw performance of the multi-core CPU by dividing processing using the inside plural threads of each function. But the big multi-thread of the particle size can cope on the application code side using singles red version in one [⁵] that the IPP function is all thread safe more. In addition, an internal thread namely the multi-thread version is abolished in IPP 7.1, and the use of the singles red version is recommended [⁶]. In addition, functions such as ippSetNumThreads() are targeted for the abolition (deprecated) and, with this, become [⁷].

IPPCV

In OpenCV 3.0, a subset of Intel IPP was donated as IPPCV.

Footnote

[Help]

1. ^ Intel®Integrated Performance Primitives (Intel®IPP) 2017 Release Notes | Intel®Software
2. ^ Before starting use of Intel ®IPP
3. ^ core function
4. ^ Intel®IPP - Intel®IPP linkage models - quick reference guide | Intel®Developer Zone
5. ^ Support of Intel ®IPP threading and OpenMP*
6. ^ Before starting use of Intel ®IPP
7. ^ ippSetNumThreads deprecated - suggested replacement?

Allied item

• Intel C++ Compiler
• Intel Parallel Studio
• OpenMP
• OpenCV
• Intel Math Kernel Library
• Intel Threading Building Blocks

It is acquired by "https://ja.wikipedia.org/w/index.php?title=Intel_Integrated_Performance_Primitives&oldid=62695567"

This article is taken from the Japanese Wikipedia Intel Integrated Performance Primitives

This article is distributed by cc-by-sa or GFDL license in accordance with the provisions of Wikipedia.

Wikipedia and Tranpedia does not guarantee the accuracy of this document. See our disclaimer for more information.

In addition, Tranpedia is simply not responsible for any show is only by translating the writings of foreign licenses that are compatible with CC-BY-SA license information.