OpenChess

robbolito e4l dynamically compiled for froyo

http://www.speedyshare.com/files/268567 ... android.7z

Hi Aart,

Thanks for the details!
Sure I will mail you about the list of features I'd like to see implemented.

Best regards,
Swami

Any chance for Houdini Android ?

pj+ wrote:robbolito e4l dynamically compiled for froyo

We now have two Robbolito's for Android, see screenshots below. Can you explain the difference between g3l and e4l? Is it worth starting an eng-eng match between the two?
Btw, I would remove the "x86" annotation in the Android compilation, as that is somewhat misleading

I tested a little, e4l compile seemed to be a little faster. In my opinion e4 was the best robbolito version, at least in the limit of small time controls / slow hardware, from there it got worse. The differences between e4 and g3 is mostly time management and further code translation into English. With current available time controls in Chess for Android there should not be difference.
I compiled it mainly to see how to do it for Android with external libraries using ndk. Still did not figure out 'right ' way.

pj+ wrote:With current available time controls in Chess for Android there should not be difference.

You were right! It could not have been more even. Results of a 1-second per move match, 16MB hash, on a Nexus One.

Tried to compile latest Ivanhoe 47c with ndk, no luck. Cross-compiling static executable is straightforward though, with usual huge number of warnings.
Seems to be slow,, don't know how strong it supposed to be.
I also put in archive 'x86' corrected robbolito e4l

http://www.speedyshare.com/files/26878872/IvanHoe47c.7z

pj+ wrote:latest Ivanhoe 47c

Thanks. About 56KNs out of the opening.....

Quick little tournament to get some idea (16MB hash, 1-second-per-move).

is the tournament handled by chess itself, or do you present the stats by your own?
because a tournament feature would the next i would wish