Author: Dann Corbit
Date: 15:46:17 08/19/99
Go up one level in this thread
On August 19, 1999 at 18:05:39, Pete Galati wrote: [snip] >I've copied this to a txt file for reference, I've generally stumbled when >combining .c files but you seem to be #include ing them at the top of a main.c >type of file, Pierre seems to do that. > >On my revised version of SCP I've used -O7, that produced marginaly better nps >than other numbers. I general test it with 1. a4 and right off the bat it goes >into a search because my new book has nothing to cover that. On my computer your >version gets 7651 nps and the best I've gotten w/my version is 6653 nps. Probably because I use MS VC++, which produces much better binaries than GCC most of the time. >I tried the -pedantic out of your makefile and it was no help for use with >DJGPP. I allways thought that -Wall was just for throwing warnings at you. I've >never tried -ansi, I figure that SCP is ansi now, but it's original code >probably wasn't, I don't know what -ansi is doing, it's one I'll try. I'm >finding that -DFAST seems to help, don't know why, I borrowed it from Crafty's >makefile. Most of the flags are just for finding bugs. -Wall -ansi -pedantic are really a minimum set. You can get much more 'anal retentive' by an even more severe set of restrictions. I also strongly recommend LCLint for finding problems. PC-Lint is also good (as is the UNIX translation which I forget the name for it -- but you can find it at Gimpel).
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.