Various fixes related to inlining #4464
Conversation
In readLater the context is not preserved, but reconstituted using the owner of the context building the closure. We need to pass the mode as well, or else ReadPositions might be reset. I noticed this when I saw that positions were not read correctly when inlining.
…nitions This was done by accident before, because we did not keep the mode bits of the context calling a readLater.
When printing trees, modifiers were always taken from the tree. This is wrong for typed trees, since their flags and annotations come from the symbol, the tree modifiers are no longer relevant.
Was forgotten before, which meant that labels were forgotten in inlined code. The code was still correct, but a method would be created for each label.
Dealias them and try again.
The idea is to have a compiler-specific list data type that optimizes for small lists. Where the overhead of normal lists is one Cons node per element, the overhead of `Lst` is 0 for lists of length <= 1, and 1 array node for longer lists. Furthermore, we'll try to inline heavily, which is easier for specialized data structures. The main downsides are inefficient `tail` and `::`, in particular for longer lists. Compiler code will need to be adapted to avoid these where it is performance critical. This also serves as test case for the inline related fixes in the previous commits.
|
test performance please |
|
performance test scheduled: 1 job(s) in queue, 0 running. |
|
Performance test finished successfully: Visit http://dotty-bench.epfl.ch/4464/ to see the changes. Benchmarks is based on merging with master (6f6816d) |
|
Are all the fixes tested by |
|
|
||
| def length: Int = elems match { | ||
| case null => 0 | ||
| case elems: Arr => elems.length |
There was a problem hiding this comment.
This would not behave correctly if we have Lst[Array[Any]].
There was a problem hiding this comment.
Same for every other pattern match.
There was a problem hiding this comment.
I have an updated version of Lst in the "Try some Optimizations" pull request that handles this case. The version in this PR serves just as a test case for the inliner.
| def sharedOp(x: T) = op(x) | ||
| elems match { | ||
| case null => | ||
| case elems: Arr => def elem(i: Int) = elems(i).asInstanceOf[T] |
There was a problem hiding this comment.
Why is this cast in a method?
There was a problem hiding this comment.
It's an automatic rewrite, to avoid having to cast at multiple places in the body. I rewrote:
case elems: Array[T] =>
to what you see here.
There was a problem hiding this comment.
Now it makes sense :)
| private def eq(x: Any, y: Any) = x.asInstanceOf[AnyRef] `eq` y.asInstanceOf[AnyRef] | ||
| private val eqFn = (x: Any, y: Any) => eq(x, y) | ||
|
|
||
| val Empty = new Lst[Nothing](null) |
There was a problem hiding this comment.
Why not an inline def? That will just be null in the bytecode
There was a problem hiding this comment.
Good point. I'll update it in the newer version of Lst.
yes. |
No description provided.