mutable arrays of tuples

So I was thinking about a mutable array of tuples, but to avoid allocating tuples to modify their fields, I guess I really want an immutable array of tuples of STRefs. Just how much less efficient is this than a plain mutable array? might it even make sense to use parallel mutable arrays? The thought of that is disgusting to me, but at least one of the arrays could likely be made unboxed...

David Feuer
So I was thinking about a mutable array of tuples, but to avoid allocating tuples to modify their fields, I guess I really want an immutable array of tuples of STRefs. Just how much less efficient is this than a plain mutable array? might it even make sense to use parallel mutable arrays? The thought of that is disgusting to me, but at least one of the arrays could likely be made unboxed...
Maybe you could use a tuple of (unboxed) arrays instead? Or if you use Vector instead of Array, I think tuples are member of Unbox (as long as the tuple elements are), and can be used directly in unboxed vectors. -k -- If I haven't seen further, it is by standing in the footprints of giants

In fact, unboxed arrays of tuples are represented in vector as tuples of
unboxed arrays.
On Aug 9, 2012 4:35 AM, "Ketil Malde"
David Feuer
writes: So I was thinking about a mutable array of tuples, but to avoid allocating tuples to modify their fields, I guess I really want an immutable array of tuples of STRefs. Just how much less efficient is this than a plain mutable array? might it even make sense to use parallel mutable arrays? The thought of that is disgusting to me, but at least one of the arrays could likely be made unboxed...
Maybe you could use a tuple of (unboxed) arrays instead? Or if you use Vector instead of Array, I think tuples are member of Unbox (as long as the tuple elements are), and can be used directly in unboxed vectors.
-k -- If I haven't seen further, it is by standing in the footprints of giants
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
participants (3)
-
David Feuer
-
Jake McArthur
-
Ketil Malde