diff options
| author | Bob Wilson <bob.wilson@apple.com> | 2013-06-25 19:09:50 +0000 |
|---|---|---|
| committer | Bob Wilson <bob.wilson@apple.com> | 2013-06-25 19:09:50 +0000 |
| commit | a1fe2948ed4039e68d1784494c3b23a4ce4126b4 (patch) | |
| tree | 4a63fabad494aa789ebf7c51c28166308920a07a /test/Analysis | |
| parent | 2560e242c87b205143ca5e0e87a216e06f51a434 (diff) | |
| download | external_llvm-a1fe2948ed4039e68d1784494c3b23a4ce4126b4.zip external_llvm-a1fe2948ed4039e68d1784494c3b23a4ce4126b4.tar.gz external_llvm-a1fe2948ed4039e68d1784494c3b23a4ce4126b4.tar.bz2 | |
Fix SROA to avoid unnecessary scalar conversions for 1-element vectors.
When a 1-element vector alloca is promoted, a store instruction can often be
rewritten without converting the value to a scalar and using an insertelement
instruction to stuff it into the new alloca. This patch just adds a check
to skip that conversion when it is unnecessary. This turns out to be really
important for some ARM Neon operations where <1 x i64> is used to get around
the fact that i64 is not a legal type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@184870 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/Analysis')
0 files changed, 0 insertions, 0 deletions