Author: Robert Hyatt
Date: 19:46:28 09/02/03
Go up one level in this thread
On September 02, 2003 at 18:26:44, Jeremiah Penery wrote: >On September 02, 2003 at 18:08:40, Vincent Diepeveen wrote: > >>On September 02, 2003 at 00:14:02, Jeremiah Penery wrote: >> >>>On September 01, 2003 at 23:23:18, Robert Hyatt wrote: >>> >>>>On September 01, 2003 at 09:39:55, Jeremiah Penery wrote: >>>> >>>>>Any large (multi-node) SMP machine will have the same problem as NUMA with >>>>>respect to inter-node latency. SMP doesn't magically make node-to-node >>>>>communication any faster. >>>> >>>>Actually it does. SMP means symmetric. >>>> >>>>NUMA is _not_ symmetric. >>> >>>Of course. The acronym means "non uniform memory access". >>> >>>But if you think "symmetric" necessarily means "faster", maybe you'd better look >>>in a dictionary. >> >>You're wrong by a factor 2 or so in latency and up to factor 5 for 128 cpu's. > >Blah blah blah. I snipped the rest of your drivel. > >All I said is that being SMP doesn't magically make your latency better. It >depends on how the machine is built. There's no reason a big NUMA machine >couldn't be built where the average memory access wouldn't be as fast or faster >than a similar-sized SMP machine. There are several reasons why such a machine >isn't built, but not a real technical reason. The reason is $. NUMA scales well with respect to price per additional CPU. Crossbars to not. It is possible to build a NUMA box that switches just as fast as a Crossbar. There's no reason to do so, because you end up with a crossbar, and its associated high scaling cost.
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.