You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jun 1, 2023. It is now read-only.
We typed now all ops, and can optimize (up and downgrade) on them.
But we want to pass native typed data (e.g. INT_PADSV) to subs also, when they allow
native types.
We want to keep optimized sub args when the sub allows it, and don't want to downgrade all typed ops within pushmark.
Dispatch (run-time)
Our compile-time type double-dispatch is slow but needs not much memory, we just cycle all type variants until we find it.
The user-based dispatch needs to be fast, so we need dispatch tables, or even a dynamically adjusted tree. it is also multi-dispatch, not just a simple double dispatch
Damian went in Dios with a too trivial arity check first, while he really should have started from left.
The text was updated successfully, but these errors were encountered:
Optimizer (compile-time)
We typed now all ops, and can optimize (up and downgrade) on them.
But we want to pass native typed data (e.g. INT_PADSV) to subs also, when they allow
native types.
We want to keep optimized sub args when the sub allows it, and don't want to downgrade all typed ops within pushmark.
Dispatch (run-time)
Our compile-time type double-dispatch is slow but needs not much memory, we just cycle all type variants until we find it.
The user-based dispatch needs to be fast, so we need dispatch tables, or even a dynamically adjusted tree. it is also multi-dispatch, not just a simple double dispatch
Damian went in Dios with a too trivial arity check first, while he really should have started from left.
The text was updated successfully, but these errors were encountered: