| commit | 550a9d823a939366a9f776b58f18883acd905a93 | [log] [tgz] |
|---|---|---|
| author | Bob Wilson <bob.wilson@apple.com> | Tue Aug 14 17:27:04 2012 +0000 |
| committer | Bob Wilson <bob.wilson@apple.com> | Tue Aug 14 17:27:04 2012 +0000 |
| tree | 3641829ab867fc71805c5528bbe5e5ab1be925b6 | |
| parent | c41ace950dcf2254c9aa48e73647b89c35109f80 [diff] |
Avoid using i64 types for vld1q_lane/vst1q_lane intrinsics. The backend has to legalize i64 types by splitting them into two 32-bit pieces, which leads to poor quality code. If we produce code for these intrinsics that uses one-element vector types, which can live in Neon vector registers without getting split up, then the generated code is much better. Radar 11998303. git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@161879 91177308-0d34-0410-b5e6-96231b3b80d8