Question regarding the Wine Vs WineLib performance
Tom Spear (aka Dustin Navea)
speeddymon at gmail.com
Wed Jan 4 18:13:24 CST 2006
This may sound like a stupid thought, and may have already been
discussed (I couldnt attend wineconf), but doesnt g++ compile everything
with -Ox upon request, so it is size-optimized (read: Compressed), and
don't most people use that same flag on most compilations? It seems to
me that if the above 2 statements are correct, then the reason MSVC
produces faster code than g++ is because MSVC doesnt compress it's
output, you have to use something like upx to compress it, while keeping
it executable, and then once that is done, the binary runs slower...
Like I said, probably already discussed, and probably incorrect, but my
2c either way.
Brian Vincent wrote:
>On 1/3/06, Ananth M <mekaananth at gmail.com> wrote:
>
>
>> Then I converted the dll into .so and compiled and linked this .so
>>with win32 program using WineLib and Wineg++
>> The timing that was taken to execute the function exported by the
>>dll ( through .so ) is almost two times, compread with executing using wine.
>>
>> Does any one has any reference for further investigation on this
>>problem or any information
>>
>>
>
>I'm not sure if this is the reason or not, but we discussed this a bit
>at WineConf last year. The concensus was that MSVC produces much
>faster code compared to GCC. Maybe try compiling with MinGW on
>Windows and see what happens with the DLL in Windows?
>
>-Brian
>
>
More information about the wine-devel
mailing list