Search found 1028 matches

by smatovic
Wed Jan 23, 2013 6:19 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Kogge Stone, Vector Based
Replies: 8
Views: 2875

Re: Kogge Stone, Vector Based

the idea for the generalized shift is taken from the chessprogramming wiki, Do you have any links to the concrete SIMD instruction set? For Nvidia there is the PTX instruction set: www.nvidia.com/content/CUDA-ptx_isa_1.4.pdf for AMD CAL http://coachk.cs.ucf.edu/courses/CDA6938/AMD_course/M6%20-%20CA...
by smatovic
Tue Jan 22, 2013 11:30 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Kogge Stone, Vector Based
Replies: 8
Views: 2875

Kogge Stone, Vector Based

Maybe it is of interest for someone, the 8 directions for Kogge-Stone are done via an ulong8 vector datatype, on my gpu it performs as fast as magic bitboards... // generate pseudo legal moves while( bbWork ) bbTemp = 0; bbMoves = 0; // pop 1st bit pos = pop_1st_bit(bbWork); piece = getP...
by smatovic
Thu Jan 17, 2013 2:40 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: How beautiful is your code?
Replies: 59
Views: 11742

Re: How beautiful is your code? simplicity

hmm, difficult question,
imo it is also about simplicity,
when you express something complex in an simple manner
it can be be beautiful....

in this sense, my code needs a pretty up ;)

--
Srdja
by smatovic
Thu Jan 17, 2013 12:47 am
Forum: Computer Chess Club: General Topics
Topic: Zeta097x on Nvidia GTX580 @ 1800 ELO
Replies: 1
Views: 422

Zeta097x on Nvidia GTX580 @ 1800 ELO

I have done some fixing and modding and Zeta097x reaches now clearly the 1800 ELO mark on an Nvidia GTX 580 GPU...

https://github.com/smatovic/Zeta/tree/zeta_cl_097x

Zeta097x uses an slightly improved Evaluation Function from ZetaDva @ 2000 ELO...

--
Srdja
by smatovic
Mon Jan 14, 2013 10:15 pm
Forum: Computer Chess Club: General Topics
Topic: Intel Xeon Phi coprocessor cards
Replies: 6
Views: 1485

Re: Intel Xeon Phi coprocessor cards - dr dobbs links

http://www.drdobbs.com/parallel/cuda-vs-phi-phi-programming-for-cuda-dev/240144545 http://www.drdobbs.com/parallel/programming-intels-xeon-phi-a-jumpstart/240144160 1 core can run 4 threads via hyperthreading, so 240 threads should be running, the question is how to utilize the 512 bit vector unit e...
by smatovic
Tue Jan 01, 2013 9:38 pm
Forum: Computer Chess Club: General Topics
Topic: Best 2012 chess GUI ?
Replies: 37
Views: 6644

Re: Best 2012 chess GUI ?

as an XBoard user i feel a bit discriminated.....

---
Srdja
by smatovic
Mon Dec 31, 2012 11:13 pm
Forum: Computer Chess Club: General Topics
Topic: HAPPY NEW YEAR 2013 CCC FORUM !
Replies: 9
Views: 597

Re: HAPPY NEW YEAR 2013 CCC FORUM !

Srecna Nova godina :)

--
Srdja
by smatovic
Sun Dec 30, 2012 2:46 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: How do you spend time?
Replies: 10
Views: 2616

Re: How do you spend time?

my engine is still under construction and unstable,
so i need to watch most of the games played with debug output to
see how an modification behaves :oops:

--
Srdja
by smatovic
Fri Dec 28, 2012 7:56 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: MCab and depth based score bonus?
Replies: 3
Views: 595

Re: MCab and depth based score bonus?

Position A, score: 45, depth 10
Position B, score: 50, depth 3
i think the trick is to avoid such an difference in depth at close scores by an smarter select phase....

--
Srdja
by smatovic
Fri Dec 28, 2012 7:23 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: MCab and depth based score bonus?
Replies: 3
Views: 595

Re: MCab and depth based score bonus?

Hmm,
assumed these two are root node positions, which one would you choose to play and why?

--
Srdja