[Beowulf] Vector coprocessors
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Craig Tierney ctierney at hypermall.netThu Mar 16 20:27:11 PST 2006
- Previous message: [Beowulf] Vector coprocessors
- Next message: [Beowulf] Vector coprocessors
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Joe Landman wrote: > > > Robert G. Brown wrote: >> On Thu, 16 Mar 2006, Joe Landman wrote: >> >>> I think (WAG here) that Clearspeed wants people to design a socket >>> onto motherboards for them. Lowers the costs all around. >> >> >> Wouldn't the sane thing to do be just putting them on an HT interface? >> Eliminate the PCI interface altogether, provide symmetric access to e.g. >> processor and memory etc resources. > > If you have a spare HT interface, yes. If you have a "socket" on the MB > you can to HT to that. Or run it out the HTX connector. This is not > something that you want in the IO channel. > > BTW: Anyone know if Intel is considering doing HT? Would be nice. > I would doubt it. Intel sees themselves as a leader, and to go follow AMD (and the consortium) wouldn't fit. Look around for information on the Intel Common System Interface (CSI). http://endian.net/details.asp?tag=csi There isn't much information there, but it indicates that CSI is to compete with HT. Craig
- Previous message: [Beowulf] Vector coprocessors
- Next message: [Beowulf] Vector coprocessors
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
