Author: Dann Corbit
Date: 10:44:07 09/29/99
Go up one level in this thread
On September 29, 1999 at 06:10:16, Francesco Di Tolla wrote: >>>i say all the software should be rated on equal hardware , even the old ones >>>that you only see at 486 and 90 mhz pent , to me it makes zero sence to try >>>and take the programs ssdf thinks is good or new and only test them on the new >>>hardware that the average person does not have anyway . >>You don't know what you are talking about. The "old" tests were run on slower >>hardware because it was state of the art at the time. They now have newer >>faster hardware and are testing the new programs with the new hardware. > >I think the idea was another: why not test also older software on new platforms. >This would clearly show that the difference in elo between some older version of >some programs are only slightly inferior to more modern software, while the fact >that often programs of different generation are tested only on machines >available when they first came out tends to be misinterpreted as there was a big >leap in the quality of the software, while, apart the 16bit-32bit, the biggest >difference is due to hardware. While this is a great idea mathematically, I think you will find that people (in general) could care less about what older programs can do. They are curious about what is on the shelves right now. >E.g. I pretty muth guess that Genius 2.0 would be a strong opponent on new >hardware for any program in the list. I think that a careful examination will show that hardware advances will show the lion's share of the gain in strenght for chess programs. Be that as it may, I suspect that if the SSDF spent a great deal of time testing the older programs there would be a mighty hue and cry that would make the current din pale by contrast.
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.