float xfermodes (burn, dodge, softlight) in Sk8f, possibly using AVX.

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 20 matching lines @@
     case 25: return op(v, 25);  case 26: return op(v, 26);  case 27: return op(v, 27); \
     case 28: return op(v, 28);  case 29: return op(v, 29);  case 30: return op(v, 30); \
     case 31: return op(v, 31); } return fVec
 
 template <>
 class SkNf<2> {
 public:
     SkNf(float32x2_t vec) : fVec(vec) {}
 
     SkNf() {}
-    explicit SkNf(float val)           : fVec(vdup_n_f32(val)) {}
+    SkNf(float val)           : fVec(vdup_n_f32(val)) {}
     static SkNf Load(const float vals[2]) { return vld1_f32(vals); }
     SkNf(float a, float b) { fVec = (float32x2_t) { a, b }; }
 
     void store(float vals[2]) const { vst1_f32(vals, fVec); }
 
     SkNf approxInvert() const {
         float32x2_t est0 = vrecpe_f32(fVec),
                     est1 = vmul_f32(vrecps_f32(est0, fVec), est0);
         return est1;
     }
@@ skipping 60 matching lines @@
 
     float32x2_t fVec;
 };
 
 template <>
 class SkNi<4, int> {
 public:
     SkNi(const int32x4_t& vec) : fVec(vec) {}
 
     SkNi() {}
-    explicit SkNi(int val) : fVec(vdupq_n_s32(val)) {}
+    SkNi(int val) : fVec(vdupq_n_s32(val)) {}
     static SkNi Load(const int vals[4]) { return vld1q_s32(vals); }
     SkNi(int a, int b, int c, int d) { fVec = (int32x4_t) { a, b, c, d }; }
 
     void store(int vals[4]) const { vst1q_s32(vals, fVec); }
 
     SkNi operator + (const SkNi& o) const { return vaddq_s32(fVec, o.fVec); }
     SkNi operator - (const SkNi& o) const { return vsubq_s32(fVec, o.fVec); }
     SkNi operator * (const SkNi& o) const { return vmulq_s32(fVec, o.fVec); }
 
     SkNi operator << (int bits) const { SHIFT32(vshlq_n_s32, fVec, bits); }
     SkNi operator >> (int bits) const { SHIFT32(vshrq_n_s32, fVec, bits); }
 
     template <int k> int kth() const {
         SkASSERT(0 <= k && k < 4);
         return vgetq_lane_s32(fVec, k&3);
     }
 
     int32x4_t fVec;
 };
 
 template <>
 class SkNf<4> {
 public:
     SkNf(float32x4_t vec) : fVec(vec) {}
 
     SkNf() {}
-    explicit SkNf(float val)           : fVec(vdupq_n_f32(val)) {}
+    SkNf(float val)           : fVec(vdupq_n_f32(val)) {}
     static SkNf Load(const float vals[4]) { return vld1q_f32(vals); }
     static SkNf FromBytes(const uint8_t vals[4]) {
         uint8x8_t   fix8    = (uint8x8_t)vld1_dup_u32((const uint32_t*)vals);
         uint16x8_t  fix8_16 = vmovl_u8(fix8);
         uint32x4_t  fix8_32 = vmovl_u16(vget_low_u16(fix8_16));
         return SkNf(vcvtq_f32_u32(fix8_32));
     }
 
     SkNf(float a, float b, float c, float d) { fVec = (float32x4_t) { a, b, c, d }; }
 
@@ skipping 79 matching lines @@
 
     float32x4_t fVec;
 };
 
 template <>
 class SkNi<8, uint16_t> {
 public:
     SkNi(const uint16x8_t& vec) : fVec(vec) {}
 
     SkNi() {}
-    explicit SkNi(uint16_t val) : fVec(vdupq_n_u16(val)) {}
+    SkNi(uint16_t val) : fVec(vdupq_n_u16(val)) {}
     static SkNi Load(const uint16_t vals[8]) { return vld1q_u16(vals); }
 
     SkNi(uint16_t a, uint16_t b, uint16_t c, uint16_t d,
          uint16_t e, uint16_t f, uint16_t g, uint16_t h) {
         fVec = (uint16x8_t) { a,b,c,d, e,f,g,h };
     }
 
     void store(uint16_t vals[8]) const { vst1q_u16(vals, fVec); }
 
     SkNi operator + (const SkNi& o) const { return vaddq_u16(fVec, o.fVec); }
@@ skipping 16 matching lines @@
 
     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)) {}
+    SkNi(uint8_t val) : fVec(vdupq_n_u8(val)) {}
     static SkNi Load(const uint8_t vals[16]) { return vld1q_u8(vals); }
 
     SkNi(uint8_t a, uint8_t b, uint8_t c, uint8_t d,
          uint8_t e, uint8_t f, uint8_t g, uint8_t h,
          uint8_t i, uint8_t j, uint8_t k, uint8_t l,
          uint8_t m, uint8_t n, uint8_t o, uint8_t p) {
         fVec = (uint8x16_t) { a,b,c,d, e,f,g,h, i,j,k,l, m,n,o,p };
     }
 
     void store(uint8_t vals[16]) const { vst1q_u8(vals, fVec); }
@@ skipping 18 matching lines @@
     uint8x16_t fVec;
 };
 
 #undef SHIFT32
 #undef SHIFT16
 #undef SHIFT8
 
 }  // namespace
 
 #endif//SkNx_neon_DEFINED