[Remind-Fans] Performance (was Re: New release of Remind coming soon; testers wanted)

[Dianne Skoll]

On Sat, 1 Jun 2024 08:49:17 -0500
Tim Chase via Remind-fans <remind-fans at lists.skoll.ca> wrote:

> Also, though I'd have to scaffold out an actual timing benchmark
> between them,  my gut feel for `rem -n` producing all my events was
> that it ran a LOT faster.  Any thoughts on whether that feels right
> to you?

It depends on the reminder file.  On my reminders, it's a tad slower.
Running "time remind -p40 ~/.reminders" for me gives:

New: real	0m0.720s
Old: real	0m0.717s

It's a tiny difference.  However, running on the file at the end of
this email makes a huge difference because the short-circuit "&&" helps
tons:

time remind -p40 remind-on-blue2.rem
New: real	0m1.457s
Old: real	0m2.873s

Moral is if you have a lot of expressions, expecially ones that use
logical operators, choose() and iif() a lot, and especially if you
have a lot of user-defined functions, you'll likely see a speedup.

Regards,

Dianne.

#-----------------------------------------------------------------
# remind-on-blue2.rem: Issue a reminder on the second full moon of
# a calendar month
SET $MaxSatIter 10000

FSET is_second_full(date) date == moondate(2, date) && monnum(date) == monnum(moondate(2, date-30))

REM 30 SATISFY [is_second_full(trigdate())] MSG %"Blue Moon%"
REM 31 SATISFY [is_second_full(trigdate())] MSG %"Blue Moon%"