Author: Robert Hyatt
Date: 22:00:27 01/15/98
Go up one level in this thread
On January 15, 1998 at 21:51:06, Carsten Kossendey wrote: >On January 15, 1998 at 20:33:15, Robert Hyatt wrote: > >>On January 15, 1998 at 16:41:34, William Bryant wrote: >> >>>Because Metrowerks PPC Compilers now support a long long data type >>>(64 bit integer type), shouldn't most of the port of Crafty simply >>>be recompiling the program and making the integers all 64 bits. >>> >>>As I understand it, isn't crafty written in ANSI C. >>> >>>I know this may seem nieve? Any other problems with porting? >>> >> >>long long was *the* problem. If the Mac has them, then we should be >>set. What I need is for someone to compile and get it working on this >>compiler (for the Mac) and then send me any changes. > >I did so three times (10.18, 11.x, and 12.6). You never replied, so I >gave up on that. wait. As I recall, I did respond. You converted this to C++ using a class for the bitboard things, but if you recall, it would not compile cleanly on my machine. It simply crashed the gcc compiler, which made it impossible for me to use that at all. That was the only real problem I had. It was apparently somewhat slower, but when I couldn't compile it I couldn't do much with it... long long data types would certainly eliminate the C++ stuff.. > >>There likely will >>be very few, except maybe the CheckInput() function in utility.c, and >>a couple of header files. > >A number of others as well. > >>If the compiler has long long already, this is an hour job, or less, >>generally. Without long long's, it's a bear... > >Rather massive changes are required for the header files. > this shouldn't be necessary for an ANSI compiler with long long. IE this thing compiles cleanly on all sorts of computers from alphas running NT, to every unix box I have tried. If the compiler has problems with the global declarations that can be a problem however, but the ANSI standard explicitly covers the way I do globals... >>>Any assistance appreciated. >>> >>>Also, the new G3 processors with their extra integer support, >>>their increased system bus speed, and their backside catch should >>>make an ideal desktop system for a chess program. >>>Their not Alpha processors but a heck of a lot cheeper >>>(~$2000 for a well equiped desktop system). >>> >>>William Bryant >>>wbryant@ix.netcom.com
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.