Changes in directory llvm/test/Regression/CodeGen/X86:
vec_shuffle-5.ll updated: 1.1 -> 1.2
---
Log message:
shufps with load folding is better than movaps; movsd.
---
Diffs of the changes: (+2 -1)
vec_shuffle-5.ll |3 ++-
1 files changed, 2 insertions(+), 1 deletion(-)
Index: llvm/te
Changes in directory llvm/test/Regression/CodeGen/X86:
vec_shuffle-5.ll added (r1.1)
---
Log message:
Use movsd to shuffle in the lowest two elements of a v4f32 / v4i32 vector when
movlps cannot be used (e.g. when load from m64 has multiple uses).
---
Diffs of the changes: (+11 -0)
vec_shu
