Author: Dann Corbit
Date: 13:43:58 06/20/02
Go up one level in this thread
On June 19, 2002 at 23:42:17, Keith Evans wrote: >On June 19, 2002 at 22:35:19, Tom Kerrigan wrote: > >>On June 19, 2002 at 20:45:33, Keith Evans wrote: >> >>>Is there an easy way to compare a 1.4 GHz P3 to a 1 GHz McKinley and see where >>>this Crafty performance increase is coming from? I'm not at all familiar with >>>McKinley, but would it be possible to run a version of Crafty compiled for >>>32-bits on a McKinley and compare that to a Crafty compiled for 64-bits on >>>McKinley? Is this a dumb idea? If this isn't possible, then it's going to be >>>difficult to tell where the performance gain is really coming from. >> >>It's extremely possible. You can make a bitboard class that contains two 32-bit >>ints and overload all the int operators and run it on a 64-bit chip. I suggested >>this to Bob last time we had this argument and his response was that Crafty is >>written in C, not C++. (Apparently renaming his source files from .c to .cpp and >>writing this simple class is too much work.) Perhaps somebody else with access >>to a 64-bit chip and a modicum of motivation could perform this experiment. >> >>-Tom > >This seems like an interesting experiment to me. So despite your suggestion >nobody has ever done this? Not even for an Alpha? > >Does anybody see a reason that this wouldn't be a fair test? If it's fair, then >I would think that it could help to settle the bitboard performance debate. > >We do have an Alpha at work, but I'm not sure if I could get access. It's pretty >much committed to running a GPS simulator, plus it's running VMS (yuck) which >nobody at my workplace really understands. They just know enough to launch the >GUI. Here is what I think that experiment settles: For chip x and compiler y under OS z, the ratio is <blah> It answers no other queries (like what happens on a hammer or mckinley, or with a change of compiler or OS etc.) Just fetching 32 bit integers on a 64 bit machine is slower than 64 bit ones because of addressing, in general. I expect 64 bits to be a big win, at least for some machines. I have done some experiments with int sizes (but not for chess programs). On the alpha, 64 bit integers are a clear win for encryption. Not sure about other stuff, though.
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.