about fast interconnects and SCI in particular

Eugene Leitl eugene.leitl@lrz.uni-muenchen.de
Mon, 14 Jun 1999 16:21:18 -0400


Greg Lindahl writes:

 > I would hesitate to jump to such a grand conclusion. Even for fine
 > grained programs such as Charmm, I've gotten supercomputer-beating
 > results with my AlphaLinux/Myrinet cluster.

Do you know how NAMD2 scales, perchance? Particularly on FastEthernet?
I'm talking about 8 nodes.

TIA,

Eugene