> On Feb 19, 2018, at 7:04 AM, 'Paulo Matos' via Racket Users
> <racket-users@googlegroups.com> wrote:
>
> Hi,
>
> There was a time, many moons ago, not too far from the initial
> implementations of typed racked when Sam was working on it that moving
> from Racket to Typed Racket had a huge impact on performance.
>
> Nowadays that might not be the case. I also noticed that using typed
> racket might in itself improve performance (optimization coach seems to
> mention Typed-Racket specific optimizations).
>
> So, could anyone elaborate on their experiences to using typed racket
> nowadays with regards to performance? Can it still be a step backwards
> these days?
This characterization is still good:
— fully typed programs might run significantly faster than untyped programs
(30%)
— the performance of partially typed programs is all over the map (10% slower
to 10000% slower)
— it remains difficult to predict whether the conversion of a single module
introduces a serious slowdown
— slowdowns are caused by high frequency or large composite “crossings” of
typed-untyped module boundaries
the allocation cost of wrappings of mutable and/or h-o values, and the
wrappers themselves when these values are accessed.
If you want more specific information, we can post a journal submission that
compares several Racket implementations (6.2, 6.3, and 6.4) on a range of
programs using functional to ho object-oriented style.
— Matthias
--
You received this message because you are subscribed to the Google Groups
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
