[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [computer-go] KGS Tournaments: Cheating



Unfortunately, non-determinism can be a big win in a concurrent program.
If multi-core processors take off, as they seem poised to, then
eventually all the top programs will be concurrent.

A quad-Operon board that supports dual (and likely quad-) -core Opertons is alreaady avaiable.
In 2007, there will be quad-core 64-bit Opterons, making a box with 16 64-bits CPU's a reality.

Performance of such a system for comp. Go, best-case, will be around 30 x the best single-CPU 32-bit systems we have now on the desktop. A complete paradigm shift if you have the cash. David's cluster's performance in an ordinary tower or rack for 10,000 USD perhaps, all in less than 2 years.

I fully agree that "serious" Go programs should be multithreaded.
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/