Peter Shenkin said:
> However, the degree of optimization you can expect depends
> greatly on the compiler.
Indeed, but it is generally agreed that certain features cause slowdown
on most systems.
For instance, accessing individual array elements (as opposed to whole
slices) of assumed-shape arrays is almost always slower because the
potential discontiguity requires an extra integer multiply (sometimes
this can be strength-reduced, sometimes not - or the resulting code
bloat hurts anyway; certain fft formulations are notoriously difficult
to optimise in this respect).
In fact we offer optimisation options for those who expect their assumed-
shape arrays to be contiguous.
> This is hearsay
Rather than offering hearsay, I've asked the author to send us his
program so that I can examine it (and then I'll offer my reasoned
comments). Without looking at his program, I find it rather difficult to
comment on why it is so slow...
...and yes, we do like to get comments from users where we've made (or
might have made) a "performance mistake". At least then we can suggest
optimisation options, coding alternatives, or fix the compiler.
Cheers,
--
...........................Malcolm Cohen, NAG Ltd., Oxford, U.K.
([log in to unmask])
_____________________________________________________________________
This message has been checked for all known viruses by Star Internet delivered
through the MessageLabs Virus Control Centre. For further information visit
http://www.star.net.uk/stats.asp
|