Re: Question regarding the Wine Vs WineLib performance

{\rtf1\ansi\ansicpg1252\fromtext \deff0{\fonttbl
{\f0\fswiss Arial;}
{\f1\fmodern Courier New;}
{\f2\fnil\fcharset2 Symbol;}
{\f3\fmodern\fcharset0 Courier New;}}
{\colortbl\red0\green0\blue0;\red0\green0\blue255;}
\uc1\pard\plain\deftab360 \f0\fs20 On 1/3/06, Ananth M  wrote:\par
>       Then I converted the dll into .so and compiled and linked this .so\par
>
> with win32 program using WineLib and Wineg++\par
>           The timing that was taken to execute the function exported by the\par
> dll ( through .so ) is almost two times, compread with executing using wine.\par
> \par
>           Does any one has any reference for further investigation on this\par
> problem or any information\par

\par
I'm not sure if this is the reason or not, but we discussed this a bit\par
at WineConf last year.  The concensus was that MSVC produces much\par
faster code compared to GCC.  Maybe try compiling with MinGW on\par
Windows and see what happens with the DLL in Windows?\par
\par
-Brian\par
\par
}

Question regarding the Wine Vs WineLib performance

[Ananth M]

On 1/5/06, [e-mail] 
wrote:
> Thank you !!

I will try to compile the windows application using MinGW and check the
performance using the same DLL.

Here is the some more information that I would like to share with you  is :
 I am using the Binary only DLL. I am not having the source for the DLL. SO
I am not compiling the DLL in windows or in Linux. I am linking this binary
only dll with my windows application, (since my application uses the
function exported by this binary only DLL)
Then I moved the executable to Linux and executed using the wine <windows
executable>.
I checked the execution time of  the function exported by the Binary Only
DLL

In the second Case , I created the .so for the binary only dll, such that
the functions that are exported by the binary only dll will be called from
the stub functions exported by the .so
Then I called the stub function (exported by .so)  from my winelib
application.
But, In this case  also, I checked the time duration for the execution of
the dll function only.( by adding a code that gives the time duration, in
the stub function that calls the dll function).

As a third Case, I loaded  the binary dll directly (without creating the
.so),  using LoadLibrary ("Binaryonlydll"),  and compiled and linked the
application using Wine supplied DLL's and header files. Then I called the
function exported by the binary only dll directly from the application  and
checked the time duration for the execution of  the dll function .

In this case also we are  getting the same timing as in Second Case, but we
are not getting the better performance as in first case.

Question regarding the Wine Vs WineLib performance

[Boaz Harrosh]

What are the timing for the three classes ?

Ananth M wrote:
> On 1/5/06, *[e-mail] 
> * <[e-mail] 
> > wrote:
>
> Thank you !!
>
> I will try to compile the windows application using MinGW and check 
> the performance using the same DLL.
>
> Here is the some more information that I would like to share with you  
> is :
>  I am using the Binary only DLL. I am not having the source for the 
> DLL. SO I am not compiling the DLL in windows or in Linux. I am 
> linking this binary only dll with my windows application, (since my 
> application uses the function exported by this binary only DLL)
> Then I moved the executable to Linux and executed using the wine 
> <windows executable>.
> I checked the execution time of  the function exported by the Binary 
> Only DLL
>
> In the second Case , I created the .so for the binary only dll, such 
> that the functions that are exported by the binary only dll will be 
> called from the stub functions exported by the .so
> Then I called the stub function (exported by .so)  from my winelib 
> application.
> But, In this case  also, I checked the time duration for the execution 
> of the dll function only.( by adding a code that gives the time 
> duration, in the stub function that calls the dll function).
>
> As a third Case, I loaded  the binary dll directly (without creating 
> the .so),  using LoadLibrary ("Binaryonlydll"),  and compiled and 
> linked the application using Wine supplied DLL's and header files. 
> Then I called the function exported by the binary only dll directly 
> from the application  and checked the time duration for the execution 
> of  the dll function .
>
> In this case also we are  getting the same timing as in Second Case, 
> but we are not getting the better performance as in first case.

Question regarding the Wine Vs WineLib performance

[Troy Rollo]

A number of people have suggested that GCC is the problem. If so, perhaps we 
should be testing with Intel's C compiler - 
http://www.intel.com/cd/software/products/asmo-na/eng/compilers/clin/index.htm  
- to see if this offers an improvement. There is an evaluation available for 
commercial use, and a free download for non-commercial (which by Intel's 
definitions means only developers who are not being paid) use.