Re: [Bug-ocrad] OCRAD with Visual Studio 2010 success

[Tilman Hausherr]

I made another benchmark. I ran the 6 files on my old notebook with a
knoppix live DVD from 2008, they were done in 10 seconds (the CD was
quiet, so everything was running from the RAM).

Then I ran the VS2010 version on Windows XP 32 bit there, it was done in
14 seconds, although this time the HDD was used.

This is even faster than on my desktop. However I was running the media
center during the whole benchmark, including the one today with MinGW, I
thought it wouldn't make a difference if it is always running, i.e. that
all would have an equal handicap.

Without the media center:

32bit VS2010: 14 secs
64bit VS2010: 11 secs
MinGW: 14 or 15 secs
cygwin+gcc: 36 seconds.

So this rehabilitates MinGW :-)

But its really weird that my old notebook isn't slower than my current
PC. It probably means my current PC is "not that good" :-(

Tilman

>On Wed, 2010-08-11 at 13:54 -0400, Tilman Hausherr wrote:
>> On Wed, 11 Aug 2010 11:54:21 -0400, Igor Filippov [Contr] wrote:
>> 
>> >
>> >> > Details of a benchmark that I just did, this time without a supervising
>> >> > application:
>> >> > 
>> >> > time with gcc + cygwin: 39 secs
>> >> > time with VC++ 2010 32 bit: 16 secs
>> >> > time with VC++ 2010 64 bit: 12 secs
>> >> 
>> >> Impressive numbers! Some Cygwin expert should check this.
>> >> 
>> >
>> >It would be very interesting to see this benchmark done with MinGW g++
>> >compiler. From what I've seen MinGW-compiled code is much faster than
>> >cygwin, though still slower than the same code compiled and running on
>> >identical hardware on Linux.
>> >
>> >Igor
>> 
>> Hi Igor,
>> 
>> I just tested it, the six files were done in 112 secs. No, this is not a
>> typo, and the MinGW was a fresh download.
>> 
>> Tilman
>