That's for just one very specific test, repeated 100 times. Total compile time is 50s with Shapeless/Kittens, and 4s with Magnolia. Hot JVM.
Show this thread
After a day optimizing ASTs, Magnolia now gets approximately 12x better compile-time performance for typeclass derivation than Shapeless!
-
-
-
On a cold JVM, I get 16s on Magnolia vs 77s for Kittens/Shapeless. Hotspot seems to be doing wonders here...
Show this thread
End of conversation
New conversation -
-
-
wow, congrats, this is very impressive!
-
I mean, it's perhaps a game changer. One order of magnitude better perf allows completely new workflows/use cases. Think git for ex.
- Show replies
New conversation -
-
-
Did you manage to deduplicate ?
-
Yes. For starters, I was generating the code for every typeclass twice.
But I also reduced my anonymous class count. - Show replies
New conversation -
-
-
Great, I knew you would do it!!! You're now a 12x developer!
-
Tu avais plus de confiance que moi...
- Show replies
New conversation -
-
-
This Tweet is unavailable.
-
Magnolia has a significant advantage over Shapeless in that it knows a lot more at the same time, so it's able to do more optimizations.
- Show replies
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.