3-15% speedup to HardLight / Overlay xfermodes.

src/opts/SkNx_neon.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 SkNx_neon_DEFINED
 #define SkNx_neon_DEFINED
 
@@ skipping 319 matching lines @@
     SkNi operator << (int bits) const { SHIFT16(vshlq_n_u16, fVec, bits); }
     SkNi operator >> (int bits) const { SHIFT16(vshrq_n_u16, fVec, bits); }
 
     static SkNi Min(const SkNi& a, const SkNi& b) { return vminq_u16(a.fVec, b.fVec); }
 
     template <int k> uint16_t kth() const {
         SkASSERT(0 <= k && k < 8);
         return vgetq_lane_u16(fVec, k&7);
     }
 
+    SkNi thenElse(const SkNi& t, const SkNi& e) const {
+        return vorrq_u16(vandq_u16(t.fVec, fVec),
+                         vbicq_u16(e.fVec, fVec));
+    }
+
     uint16x8_t fVec;
 };
 
 template <>
 class SkNi<16, uint8_t> {
 public:
     SkNi(const uint8x16_t& vec) : fVec(vec) {}
 
     SkNi() {}
     explicit SkNi(uint8_t val) : fVec(vdupq_n_u8(val)) {}
@@ skipping 29 matching lines @@
     uint8x16_t fVec;
 };
 
 #undef SHIFT32
 #undef SHIFT16
 #undef SHIFT8
 
 }  // namespace
 
 #endif//SkNx_neon_DEFINED