div255(x) as ((x+128)*257)>>16 with SSE

src/core/Sk4px.h

 /*
  * Copyright 2015 Google Inc.
  *
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE file.
  */
 
 #ifndef Sk4px_DEFINED
 #define Sk4px_DEFINED
 
@@ skipping 48 matching lines @@
     // 1, 2, or 4 SkPMColors with 16-bit components.
     // This is most useful as the result of a multiply, e.g. from mulWiden().
     class Wide : public Sk16h {
     public:
         Wide(const Sk16h& v) : Sk16h(v) {}
 
         // Pack the top byte of each component back down into 4 SkPMColors.
         Sk4px addNarrowHi(const Sk16h&) const;
 
         // Rounds, i.e. (x+127) / 255.
-        Sk4px div255() const {
-            // Calculated as ((x+128) + ((x+128)>>8)) >> 8.
-            auto v = *this + Sk16h(128);
-            return v.addNarrowHi(v >> 8);
-        }
+        Sk4px div255() const;
 
         // These just keep the types as Wide so the user doesn't have to keep casting.
         Wide operator * (const Wide& o) const { return INHERITED::operator*(o); }
         Wide operator + (const Wide& o) const { return INHERITED::operator+(o); }
         Wide operator - (const Wide& o) const { return INHERITED::operator-(o); }
         Wide operator >> (int bits) const { return INHERITED::operator>>(bits); }
         Wide operator << (int bits) const { return INHERITED::operator<<(bits); }
         static Wide Min(const Wide& a, const Wide& b) { return INHERITED::Min(a,b); }
         Wide thenElse(const Wide& t, const Wide& e) const { return INHERITED::thenElse(t,e); }
 
@@ skipping 160 matching lines @@
     #if SK_CPU_SSE_LEVEL >= SK_CPU_SSE_LEVEL_SSE2
         #include "../opts/Sk4px_SSE2.h"
     #elif defined(SK_ARM_HAS_NEON)
         #include "../opts/Sk4px_NEON.h"
     #else
         #include "../opts/Sk4px_none.h"
     #endif
 #endif
 
 #endif//Sk4px_DEFINED