[mercury-users] Flattening (was: maybe(T))

Michael Day mikeday at yeslogic.com
Mon Oct 21 19:23:18 AEST 2002


> (1) Because programmers have been shown, on many occasions, to be a pretty
> bad source of information about performance characteristics of their programs.

I was suggesting to use the programmer purely as a manual feedback device
between the profiler and the pragma, not to place any responsibility for
analysis of performance issues on their humble brain. Presumably it would
be possible to manually enable an option and see if it increases or
decreases performance for a given workload? That's necessary for
optimisation and grade selection in the compiler already.

> (2) I didn't say it would be "easy" to implement; I just said "relatively
> easy", i.e. easier than something else, which in this case happens to be
> really hard.

I stand corrected. Though you did imply that this task -- which happens to
be easy only in comparison to impossibly difficult things -- would be
worth undertaking if there was a method for telling when it should be
used. Which a pragma would provide :)

Michael

--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------



More information about the users mailing list