[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
Re: a question about formatOf operations
But even that doesn't extrapolate to all problems, such as global
optimisation and chaotic PDEs. I sincerely hope that nobody here
thinks that defining a bitwise predictable solution to a chaotic
system actually makes any mathematical sense.
No mathematical sense for sure, but it is very useful nevertheless.
Here is an example.
I never said that it wasn't useful - a lot of harmful practices are
very useful. What is catastrophic in almost any field of engineering
is to require (standardise) a harmful practice because a politically
dominant requirement finds it useful.
But we have had the debate about the harm before, few people on this
list agree with me (though, from Email, I know that some do), and I
shall not bore people by repeating it. Let's stick to the feasibility.
It was simpler to design LHC@home for bit-exact computations, actually
all they needed was a Fortran compiler that can be coerced to do the
same optimisations on Linux and Windows (they used Lahey), and a
correctly rounded libm (they used our CRLibm).
No, that's not true. They were also lucky. There are a dozen other
unspecified system properties that have to be identical, and which
generally are - at any point in time. I have lost count of the
number of times when programs have changed behaviour after some
lapse of time, leading to screams of "the system is broken" from
their users.
Did you know that there are potential differences on Intel platforms
due to the alignment of aligned operands?
Clue: SSE.
Regards,
Nick Maclaren,
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QH, England.
Email: nmm1@xxxxxxxxx
Tel.: +44 1223 334761 Fax: +44 1223 334679