> The Galaxy Ranger wrote:
>
> > With all this talk of Linux not scaling past 8 CPUs, does this
> > imply that Linux will never be ported to the connection machine?
> > It is Sparc based after all.
>
> The connection machine combines a vast array of 1-bit processors and a
> modified N-dimensional router network into a SIMD architecture. It cannot do
> symmetric multiprocessing (other than simulating it, of course). And FYI, it
> is not sparc based, never has been, and (RIP) never will be; it uses some
> very custom silicon driven by a specialized microcontroller. Yours truly
> designed the original CM-1 microcontroller documented in the DARPA proposal,
> and its first host-bus interface.
But that was only the the CM-1. Later CMs used commodity hardware and even
abandoned the hypercube model. CM-2, for instance, had Weitek floating
point units. CM5 was synchronized MIMD based on 32MHz Sparcs (up to 16384
nodes).
As for the current discussion, no massively parallel machines are SMP.
Most are much more like computing clusters, using fancy bus architectures
to communicate between nodes with their own local memory.
I'm surprised there's been little in the way of 2-way SMP clusters. I
would guess this would be slightly more cost effective than 2x nodes.
-- "Love the dolphins," she advised him. "Write by W.A.S.T.E.."
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/