Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: fwd from Jim Demmel: More on repeatability

On Aug 5 2011, Arnold Neumaier wrote:


It makes sense to require repeatability on the same machine. I expect
that, too, for debugging purposes.


That's unclear. We are moving back into a world of coprocessors, and it
is worth learning from the experience of when we were there before.


So let me be more precise:

It makes sense to require that there is a (possibly a little slower)
repeatable mode on the same machine. I expect that, too, for debugging
purposes.


Yes.  That's feasible, though a regrettable number of users will assume
that it is always the case :-(

I believe that the last serious machine where it was infeasible was
the CDC 6600, though I could be wrong.  The performance hit might be
very high on some systems, if they have to ignore the coprocessor for
that mode.  The current GPU people claim a factor of a 100 improvement
on some codes.


Regards,
Nick Maclaren.