From af28e89a134bb4f65cb2f8e7ecada8ec0be2ac90 Mon Sep 17 00:00:00 2001
From: Lioncash <mathew1800@gmail.com>
Date: Fri, 27 Apr 2018 13:36:03 -0400
Subject: [PATCH] emit_x64_vector: Vectorize fallback path in
 EmitVectorMaxU16()

---
 src/backend_x64/emit_x64_vector.cpp | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/src/backend_x64/emit_x64_vector.cpp b/src/backend_x64/emit_x64_vector.cpp
index 184f080d..68d7d7b4 100644
--- a/src/backend_x64/emit_x64_vector.cpp
+++ b/src/backend_x64/emit_x64_vector.cpp
@@ -1000,9 +1000,14 @@ void EmitX64::EmitVectorMaxU16(EmitContext& ctx, IR::Inst* inst) {
         return;
     }
 
-    EmitTwoArgumentFallback(code, ctx, inst, [](std::array<u16, 8>& result, const std::array<u16, 8>& a, const std::array<u16, 8>& b){
-        std::transform(a.begin(), a.end(), b.begin(), result.begin(), [](auto x, auto y) { return std::max(x, y); });
-    });
+    auto args = ctx.reg_alloc.GetArgumentInfo(inst);
+    const Xbyak::Xmm a = ctx.reg_alloc.UseScratchXmm(args[0]);
+    const Xbyak::Xmm b = ctx.reg_alloc.UseXmm(args[1]);
+
+    code.psubusw(a, b);
+    code.paddw(a, b);
+
+    ctx.reg_alloc.DefineValue(inst, a);
 }
 
 void EmitX64::EmitVectorMaxU32(EmitContext& ctx, IR::Inst* inst) {