|
|
|||
![]() PVS-Studio Static Code Analyzer for 64-bit and parallel C/C++ code
|
|||
![]() ![]() ![]() ![]() ![]()
02.09.2010
Feeling the new Intel Parallel Studio XE 2011 beta So I've gotten to try the C++ compiler included into Intel Parallel Studio XE 2011 beta at last.»
30.08.2010
Five days for fixing a two-character error, or a myth of almighty technologies aiding software development In this blog, you may often read posts about how this or that software tool or software development technology helps make fewer errors, find them faster and correct them easier.»
30.08.2010
d'Artagnan and Internet, or working on the problem of bad links Friends, it is high time we stopped considering links only in the context of their number and buying/ selling and counting PR of the site they are laid out on.» ![]()
22.07.2010
Using PVS-Studio with continuous integration systems This article illustrates techniques required to employ the use of PVS-Studio static code analyzer together with continuous integration systems.»
06.07.2010
Comparing capabilities of PVS-Studio and Visual Studio 2010 in detecting defects in 64-bit programs In the article, we will compare three mechanisms of code analysis from the viewpoint of detecting 64-bit errors: the Visual C++ 2010 compiler, the Code Analysis for C/C++ component included into Visual Studio 2010 and Viva64 analyzer included into PVS-Studio 3.60.»
29.06.2010
A Collection of Examples of 64-bit Errors in Real Programs
This article is the most complete collection of examples of 64-bit errors in the C and C++ languages.» ![]() |
Knowledge Base![]() Why does WoW64 use register and file system redirection?So, why does WoW64 employ the mechanism of redirecting the register and file system instead of simply assigning the obsolete register keys and "%windir%\system32" to 32-bit applications and allowing 64-bit applications to define new keys and use, for instance, "%windir%\system64" for 64-bit system libraries? It would be rather logical especially if to remember that something similar was done when moving from Windows 3.x to Windows 95/NT. A bit of history at first. The first Itanium was released in 2001. At the same time the first version of Windows XP 64-bit Edition for Itanium systems appeared. The first 64-bit processors by AMD appeared two years later - in 2003. And two more years later - in 2005 - Windows XP Professional x64 Edition was released. So, Itanium had been a target 64-bit platform long before AMD64 processors appeared. Unlike x64 (AMD64), the difference between 32-bit and 64-bit code on Itanium (IA-64) is huge. Execution of x86 instructions is either emulated by the chip itself (it is a so called x86 hardware support) or the program emulator IA-32 EL. In the both cases, the mechanism of executing 32-bit applications is very similar to a virtual machine and therefore implies low performance of 32-bit applications. So, Intel makes it abundantly clear that support of 32-bit applications exists only to simplify the process of moving to 64-bit software. It did not mean to be long because there are just a couple of main applications running on a server unlike customer desktops where there are hundreds of programs of unknown origin. Well, let us put Itanium aside for a while and look at the second factor - the complexity of porting applications to a 64-bit platform. Microsoft tried to make this port as smooth as possible. Actually, it were just enough to recompile an existing 32-bit application with a 64-bit compiler and that is all. But in practice it is not so easy of course - basically because programmers do not follow all the recommendations 100%. There will always be somebody who would like to "cut off a corner", and such optimization turns out badly then. Now we may answer the question put in the title. We need to virtualize the register and file system because the names of the register keys and disk catalogues used by an application being recompiled remain the same after recompilation as these are just text constants in the program code. Well, it is not so bad with catalogues on the disk though, because you may (and must) get the names of the most system catalogues with the help of the functions GetSystemDirectory, GetWindowsDirectory, SHGetFolderPath and the like. But in case of "System32" it was too bad, because the fixed "System32" was everywhere. All in all, it was clear that mere recompilation was not enough while search and correct replacement of many string constants in all the applications intended to be ported promised if not to cost much money but at least significantly slow down the process of moving to a new platform. The main problem was not to change the code but to convince software developers to do it. Solving this problem with the least troubles (i.e. using redirection in WoW64) seemed very nice. As AMD64 appeared - or perhaps it was explained by rather poor sales of Itanium (in comparison with the forecast) - the move to a platform free from 32-bit applications dragged. Most applications are still 32-bit and there is no bottom to it. All this resulted in what we can see now. Windows XP and Vista x64 has inherited the architecture WoW64 from Windows XP 64-bit Edition for Itanium systems, the move to pure 64-bit systems is dragging while people begin to notice some architectural contradictions (despite which, we should say, the system works rather well). | ||