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
Right now, constructing and deconstructing ABTs is very expensive; there is basically no way to really optimize the ABT interface (this was the conclusion of Tom 7 many years ago), and it is fairly hopeless generally. But with that said, we can try making some changes.
For instance, we currently cache the calculation of free variables; while this may have been an optimization at one point, I think that it may be pretty devastating now. In RedPRL today, there are only two calls to calculate the free variables. One of these occurs between every rule, but by slightly changing RedPRL, it should be semantically correct to remove that call.
The text was updated successfully, but these errors were encountered:
Right now, constructing and deconstructing ABTs is very expensive; there is basically no way to really optimize the ABT interface (this was the conclusion of Tom 7 many years ago), and it is fairly hopeless generally. But with that said, we can try making some changes.
For instance, we currently cache the calculation of free variables; while this may have been an optimization at one point, I think that it may be pretty devastating now. In RedPRL today, there are only two calls to calculate the free variables. One of these occurs between every rule, but by slightly changing RedPRL, it should be semantically correct to remove that call.
The text was updated successfully, but these errors were encountered: