Author: Jeremiah Penery
Date: 18:07:05 06/01/99
Go up one level in this thread
On June 01, 1999 at 20:50:39, Hristo wrote: >On June 01, 1999 at 20:08:41, Gregor Overney wrote: > >>There are several reasons why you got a stronger 15.20. But let's look at >>"plain" search strength. >> >>Did you run 'bench' on 15.20 and 16.8 to make sure that both engines run roughly >>the same amount of nodes/second? It is possible that 15.20 has been compiled >>with /Oa and VC++ 5.0 SP3. Until SP3 of VC++ 6.0 got released, Microsoft was >>"able" to sell a broken version of /Oa. It seems to be fixed in SP3. This little >>difference makes your code run 5% faster. >> >>Gregor > >Thanks ... >I didn't do the bench thing. But it will be done tonight. >Probably I should get the "new" 16.8 ... > >Hristo 16.8 will lkely be quite a bit faster in terms of NPS, because of the inline assembly that isn't in 15.20. Also, the newer versions were compiled with options that make it faster anyway. Jeremiah
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.