Author: Robert Hyatt
Date: 19:31:09 01/18/03
Go up one level in this thread
On January 18, 2003 at 02:06:06, Scott Gasch wrote: >On January 17, 2003 at 22:31:06, Robert Hyatt wrote: > >>On January 17, 2003 at 18:54:07, David Rasmussen wrote: >> >>>or whatever Sun calls their compiler these days: >>> >>>CC -c egtb.cpp >>>"egtb.cpp", line 69: Warning (Anachronism): Attempt to redefine LockInit without >>> using #undef. >>>"egtb.cpp", line 70: Warning (Anachronism): Attempt to redefine Lock without usi >>>ng #undef. >>>"egtb.cpp", line 71: Warning (Anachronism): Attempt to redefine UnLock without u >>>sing #undef. >>>"/usr/include/sys/machtypes.h", line 37: Error: A declaration does not specify a >>> tag or an identifier. >>>"/usr/include/sys/machtypes.h", line 37: Error: Use ";" to terminate declaration >>>s. >>>"/usr/include/sys/machtypes.h", line 37: Error: A declaration does not specify a >>> tag or an identifier. >>>3 Error(s) and 3 Warning(s) detected. >>>make: *** [egtb.o] Error 3 >>> >>>What can I do to make it work? >>> >>>/David >> >> >>I assume you are not doing a parallel search. >> >>If not, find the references to those things, and delete them. IE you >>don't need LockInit, Lock, Unlock, etc... >> >>Make them all "null" like this >> >>#define LockInit() >> >>and that should do it... > >Bob, > >If you _are_ doing a parallel search is it sufficient to just put one spin lock >(or waitable critical section) around the entire egtb probe call? That was what >I was planning to do. It's not as if doing two I/O operations at the same time >in two threads is going to speed anything up assuming the tablebases are on the >same disk / controller. > >Scott Don't do that. Eugene specifically wrote egtb.cpp to be threaded. Remember that _many_ calls to that code will _not_ do I/O. They find stuff in his LRU buffers and avoid I/O totally...
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.