Author: K. Burcham
Date: 17:55:34 12/28/03
Go up one level in this thread
On December 28, 2003 at 19:51:45, Robert Hyatt wrote: >On December 28, 2003 at 00:26:44, K. Burcham wrote: > >> >>interesting article on the dual core AMD. >>what do you think about these numbers. maybe too optimistic? >> >>based on Crafty 19.8 benchmarks posted here. >>If the 248 opteron will get 1640 knps, >>If the dual 248 will get 2296 knps, > >These are bad. IE the dual 248 comes very close to doubling the >nps of the single 248, based on the 1-2-4 846 numbers I posted here >a couple of weeks ago. > >>If the future dual core AMD gets 2296 knps, > > >It should do better than that. Whether it will equal a dual 84x is >speculation (but not likely) but it ought to get reasonably decent >performance. IE close to 2x... > > >>then maybe the dual core in smp will get 3214. > > >two dual-cores ought to be within spitting distance of four single >processors, the question is how much will the memory contention for >a single controller impact performance... The AMD 64 has 105.9 million transistors now with a die size of 190mm. I know they plan on another shrink soon, but if AMD is going to put 211.8 million transistors on a single die, then it would seem there are many issues. the processor would be large with 211.8 million on one processor. maybe only commercial. maybe no ATX motherboard, server. no ATX power supply, server. four processors in one case, several fans, noise. large ram, maybe 2-4 gigs or more, expensive. for the home builders, this may be too expensive. kburcham > > > >> >>http://www.theinquirer.net/?article=13344 >> >>kburcham
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.