Reviews

  • Allan McNaughton. The One-Stop Shop for 64-bit Development

    24.03.2006
    Would you like to see a one-stop shop providing everything you need for 64-bit development? There is such a place - AMD Developer Central. With the help of Allan McNaughton you'll learn about the advantages of this source and a great variety of services and tools provided by it. For example, very good news for many developers is that such popular tools as AMD CodeAnalyst, AMD Core Math Library and AMD SimNow used for increasing performance, writing low-level code and other purposes are now free to obtain. And there are many other spheres you can benefit in from AMD Developer Central.

  • 64-bit Insider Volume 1, Issues 1-12

    23.03.2006
    In his brief blog-note the author dwells upon an idea to report various issues of 64-bit programming for a wide public. As we can see, this idea resulted in a kind of weekly e-magazine "64-bit Insider" providing interesting information on everything connected with 64-bit programming. The main plus of these publications is their easy-to-digest language so that even developers new to program migration and 64-bit software development could benefit from these materials.

  • Brad Martin, Anita Rettinger, and Jasmit Singh. Multiplatform Porting to 64 Bits

    01.02.2006
    In this article, the authors share their experience of porting an application with support of 11 operation systems to a 64-bit environment. Proceeding from this experience they single out the two major cross-platform migration issues: integer-long-pointer mismatches and implicit function declarations. They also touch upon some more hidden issues and in the end of the article, give valuable tips on coding and a list of useful 64-bit tools and techniques. The article will be interesting for all programmers and especially for those who face the task of porting cross-platform applications with large code to 64-bit systems.

  • Rich Newman. Removing Memory Errors from 64-Bit Platforms

    01.10.2005
    Memory errors may cause a lot of problems when porting to 64-bit platforms. Rich Newman is going to share some guidelines on how to fix memory errors before porting an application. He touches upon the nature of troubles occurring when moving to 64-bits and explains step by step how to find and fix errors with the help of the so called mutation testing. You will also find some tips on how to detect problems of the 64-bit processor itself.

  • Conrad Carter. C/C++ Compiler Switches for Fast, Portable 64-Bit Code

    07.03.2005
    In a brief but informative article Conrad Carter focuses on efficient use of switches in C/C++ compilers for providing high performance and good portability of 64-bit code. In particular, he considers Visual C++ 2005 and gcc 3.3.3. For each compiler a set of the most useful switches is given as well as thorough instructions on their use. In conclusion the author touches briefly upon AMD Core Math Library.

  • The Old New Thing: Why did the Win64 team choose the LLP64 model?

    31.01.2005
    One more brief blog-note by Raymond Chen where he explains the reasons why Win64 team chose LLP64 data model, showing by an example the rationality of their choice.

  • Converting 32-bit Applications Into 64-bit Applications: Things to Consider

    11.01.2005
    As it is known, the main cause of problems when converting 32-bit applications to 64-bit ones is change of sizes of long and pointer types. Proceeding from that the authors of the article offer a set of some very useful rules of using types following which will help you avoid these problems. They cover various methods of improving your code and are illustrated by detailed examples.

  • Matt Tyrlik, Making the transition to 64 bits

    19.10.2004
    Porting applications to the 64-bit mode of the PowerPC 970FX processor is a specific task and Matt Tyrlik is going to encourage developers and help them eliminate main issues relating to porting existing 32-bit code to the new computing model. The article consists of three parts. The first one describes main issues you may encounter when porting C language software to the 64-bit computation mode of the PowerPC 970FX. This section covers data type changes, pointer arithmetic, function pointers and compiler options. The second part discusses issues of porting 32-bit assembly language software to the 64-bit computation mode and deals with TOC programming, assembly language data objects and function calls etc. The third part covers changes needed for supervisor-level software. The text is supplied with many code samples.

  • Harsha Jaquasia. Porting and Optimizing Multimedia Codecs for AMD64 architecture on Microsoft Windows

    21.07.2004
    Porting of multimedia codecs to 64-bit environment involves some challenges but done skillfully it provides a great increase of performance. The article focuses on the porting issues and ways to solve them. It consists of two main sections - "Porting Challenges" and "Optimization Techniques". In the first one, the author describes obstacles a developer can face before the porting process and gives some tips to avoid them: these tips include using SSE/SSE2 Instructions instead of MMX and 3DNow! ones, using intrinsics instead of inline assembly and using portable scalable data types. The second section dwells upon several techniques of optimizing the process of porting. They include extended 64-bit general purpose registers, loop unrolling, using aligned memory accesses and some more. Each technique is thoroughly described in illustrated by schemes and example of various codecs' codes. Besides, there is a brief note on porting and performance evaluation tools for codecs with an emphasis on CodeAnalyst.

  • Gregory Tarsy and Neil Toda. Floating-Point Computing: A Comedy of Errors?

    20.01.2004
    This article considers floating-point arithmetic for numerical computing on the example of summing square-roots on different architectures. The results of computation are thoroughly compared and described in several tables. The issues of the precision level and what to consider a "correct" result are also examined. The article will be interesting for everyone involved in floating-point computation.