Author: Uri Blass
Date: 01:12:57 12/04/02
When I do optimization for speed without changing the algorithm I need to check that there are not new bugs. A possible test is to give the program a test suite and if the program solve the test suite with the same number of nodes for every position except smaller times then it means that the optimization probably have no bugs. EPD2WB of bruce moreland can generate the relevant files but I think that a program that compares the relevant files to check if the same number of nodes and the same main lines were for all positions can be productive. The fact that the program get better results in the test suite is not enough and it is possible that a possible new bug that did not influence the times of the test suite may be a problem in a game and by comparing the files instead of comparing the times the probability for a new bug can be reduced. I think that it may be productive if somebody gives a free program only for that task. The program should ignore cases when the new program generates more main lines because if the new program is faster then it means that it may do it. The program should tell where is the mistake in case of a mistake. Uri
This page took 0.01 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.