AVX指令集中的32种浮点比较关系详解

　　在传统印象中，数字的比较关系只有6种。但在AVX指令集中，Intel一下给出了32种浮点比较谓词，详见下图--

（Intel手册：Table 3-9. Comparison Predicate for VCMPPD and VCMPPS Instructions）

　　为什么会有这么多种比较谓词呢？我为此困惑困惑了很久。
　　直到最近翻阅了不少资料后，才终于将它们弄懂了。

一、浮点数据类型

　　Intel使用的是IEEE 754规范的浮点数据类型。对于浮点数据类型来说，除了可以存储数字、无穷之外，还可以存储 NaN（not a number。非数）。

　　NaN（非数）分为两大类--

Getting Started with Intel® Cilk™ Plus SIMD Vectorization and Elemental Functions

A tutorial on how to use #pragma simd and elemental function features in Intel® Cilk™ Plus.

Linux*

Apple Mac OS X*

Microsoft Windows* (XP, Vista, 7)

C/C++

Compilateur Intel® C++

Intel® C++ Composer XE

Intel® Composer XE

simd

vector

Vectorization

pragma

Cilk Plus

simd loop

vectorlength

Optimisation

Vectorisation

Intel® Integrated Performance Primitives 6.1 Code Samples

Intel IPP 6.1 samples

Bibliothèque Intel® Integrated Performance Primitives (IPP)

sample code

simd

IPP

Using Intel® Streaming SIMD Extensions and Intel® Integrated Performance Primitives to Accelerate Algorithms

This paper examines how to use Intel® SSE to accelerate an image processing application, how to take advantage of it using compiler directives for optimal SIMD code generation, and how to employ Intel® IPP to fully exploit SIMD acceleration.

simd

IPP

visual computing

Intel IPP

image processing

Intel SSE

vcsource_type_techarticle

vcsource_domain_media

vcsource_os_windows

vcsource_platform_desktoplaptop

vcsource_product_icc

vcsource_os_linux

vcsource_index

Graphiques

Développement de jeu

Optimizing the Rendering Pipeline of Animated Models Using the Intel Streaming SIMD Extensions

by J.M.P. van Waveren, Id Software, Inc.

Introduction

The following articles describe the individual routines used in the rendering pipeline and how they are optimized using the Intel Streaming SIMD Extensions.

simd

specular map

visual computing

vcsource_type_techarticle

vcsource_os_windows

vcsource_platform_desktoplaptop

vcsource_domain_graphics

vcsource_domain_gamedev

vcsource_index

vcsource

Développement de jeu

Graphiques

How Special Silicon Facilitates Parallel Arithmetic

SIMD (for single instruction, multiple data) enables one arithmetic instruction to be performed across multiple data items simultaneously. This article lightly touches on the advantages of SIMD for parallel applications.

Compilateur Intel® C++

simd

Multi-thread apps for Multi-Core

SSE

game development

physics

visual computing

vcsource_type_techarticle

vcsource_os_windows

vcsource_platform_desktoplaptop

vcsource_domain_graphics

vcsource_product_icc

vcsource_index

vcsource

Développement de jeu

Graphiques

Informatique parallèle

3D Vector Normalization Using 256-Bit Intel® Advanced Vector Extensions (Intel® AVX)

This article shows how to use 256-bit Intel® Advanced Vector Extensions (Intel® AVX) to normalize an array of 3D vectors. We describe a shuffle approach to convert between AOS & SOA on-the-fly in order to make data ready for up to 8-wide SIMD processing.

SOA

simd

AVX

game development

visual computing

transpose

shuffle

AOS

vcsource_type_techarticle

vcsource_os_windows

vcsource_platform_desktoplaptop

vcsource_domain_graphics

vcsource_index

Intel® Advanced Vector Extensions

Développement de jeu

Graphiques

Vectorization - Find out what it is, Find out More!

One of my performance focus areas for this year is vectorization. I am excited to start creating more content and spreading the message about this technology, as it has been a little bit underappreciated in the past. So to kick things off, I am going to launch a blog series and a 1-hour overview webinar.

-------------
First, information about the webinar.