Re: [ng-spice] Support for SMP ??
On Tue, 30 Jan 2001, James Swonger wrote:
> If you're doing looped analyses (like parametric or Monte Carlo analyses)
> then parallelism at the "job" level will help you out and this could be
> script-automated or -assisted. If you're looking to speed up individual,
> linear runs, then I think that you're probably out of luck and only faster
> hardware / more memory will help. The analog simulations are inherently
> serial solutions of large matrices and I'm not optimistic that you can get
> the matrix solution to be shared. Even if you could, you would need a lot
>of
> brain-to-brain bandwidth; throwing the whole matrix back & forth over a
> network would be even nastier.
James, what about matrix partition on multiple CPU ?
I have spotted a sparse matrix package (superLU_mt) that uses threads
to run on parallel shared memory machines. Even this should be an
improvement.
> Check into what types of SPICE jobs the big waits are incurred for;
> are they, say, nested temperature/process/parameter loops, that you could
> flatten and parcel out to individual nodes in a compute cluster, and then
> reassemble the output data? Or just hugely long single runs?
Ngspice inherits its code from cider. Cider had some support for parallel
machines, may be you are interested.
Paolo
Partial thread listing: