- stardom414 wrote:
- Stockfish 2.2.2h-GL 64-bit
Susposedly shows a littel improvement but no guarntees, may be just minor bugfix I cannot trust retards for results in chess2u.
Every1 test result and fine out, then post ur scores here please .
http://www.spaadyshare.com/AD1VAY9BP2DL/Stockfish_2.2.2h-GL_x64.rar
Thank you stardom414
Here is what the compiler posted on the German forum , translated into English...
Stockfish 2.2.2h-GL - with hash ad
After a long break, I went back a bit busy chess engines and it struck me that the show floor newer versions Fish no information about the level of the hash memory.
There may be a reason why this was removed in newer versions, however, I have this hash info now as in the 1.7x versions implemented again, so that in the Stockfish 2.2.2h-GL version can recognize how the hash memory fills in the course of thinking time. A performance drawback Resulting in the not so good.
The Stockfish 2.2.2h-GL version is available here:
http://www.file-upload.net/download-4105864/stockfish-222h-GL.zip.htmland in addition to the complete modification of the listed sources with a
Windows 32-bit exe in:
stockfish-222H-GL \ Windows \ 32 \ Stockfish-222H-32-GL.exe
(Compiled with Microsoft Visual C + + 2010 Express)
On my Core 2 Duo 3.0, this version comes with Arena almost the same number of nodes / sec. as compiled by Jim Ablett version 2.2.2 YES. Seems sooo bad for the Microsoft compiler that is not his.
Regards
Gerd Lahnstein
Posted By Marsell Date 2012-02-11 18:38
Thank you, great thing. Give it a 64bit version too, because I test with 64-bit programs running have and can make comparisons.
Martin Steinwandter
Hello Marsell,
exe with a Win 64-bit I can unfortunately not be used, since I can compile with my system, only 32-bit exe, but the sources are indeed available, so you should be sure to find someone who does this job, perhaps.
As for comparisons ..., well, this Stockfish 2.2.2h-GL version should actually behave exactly like the original version by Jim Ablett compiled, even if quite naturally with a new compilation with another compiler, and if only some minor modifications "versaubeuteln" may.
In my tests, I found, however, no differences. As already mentioned, reached the GL version 2.2.2h with me on my Core 2 Duo almost the same number of nodes / sec. Interestingly, it is faster on my sytem even a bit when you can expect only one thread
Regards
Gerd Lahnstein