Michael Kuperstein 5e343e6fd0 [X86] Improve a dag-combine that handles a vector extract -> zext sequence.
The current DAG combine turns a sequence of extracts from <4 x i32> followed by zexts into a store followed by scalar loads.
According to measurements by Martin Krastev (see PR 21269) for x86-64, a sequence of an extract, movs and shifts gives better performance. However, for 32-bit x86, the previous sequence still seems better.

Differential Revision: http://reviews.llvm.org/D6501

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@223360 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-04 13:49:51 +00:00
..
2014-11-06 15:05:51 +00:00
2014-12-03 02:08:38 +00:00
2014-11-11 21:08:02 +00:00
2014-12-04 09:20:13 +00:00