[x264-devel] [PATCH 3/5] x86util: XOP optimized HADDD

[James Almer]

Signed-off-by: James Almer <jamrial at gmail.com>
---
 common/x86/x86util.asm | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/common/x86/x86util.asm b/common/x86/x86util.asm
index 79ff34f..1a28be9 100644
--- a/common/x86/x86util.asm
+++ b/common/x86/x86util.asm
@@ -298,11 +298,16 @@
     paddd   %1, %2
 %endif
 %if mmsize >= 16
+%if cpuflag(xop) && sizeof%1 == 16
+    vphadddq %1, %1
+%endif
     movhlps %2, %1
     paddd   %1, %2
 %endif
+%if notcpuflag(xop)
     PSHUFLW %2, %1, q0032
     paddd   %1, %2
+%endif
 %undef %1
 %undef %2
 %endmacro
-- 
1.8.3.2