1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
|
inline SkPMFloat& SkPMFloat::operator=(const SkPMFloat& that) {
fColors = that.fColors;
return *this;
}
// For SkPMFloat(SkPMColor), we widen our 8 bit components (fix8) to 8-bit components in 16 bits
// (fix8_16), then widen those to 8-bit-in-32-bits (fix8_32), and finally convert those to floats.
// get() and clamped() do the opposite, working from floats to 8-bit-in-32-bit,
// to 8-bit-in-16-bit, back down to 8-bit components.
// _mm_packus_epi16() gives us clamping for free while narrowing.
inline SkPMFloat::SkPMFloat(SkPMColor c) {
SkPMColorAssert(c);
__m128i fix8 = _mm_set_epi32(0,0,0,c),
fix8_16 = _mm_unpacklo_epi8 (fix8, _mm_setzero_si128()),
fix8_32 = _mm_unpacklo_epi16(fix8_16, _mm_setzero_si128());
fColors = _mm_cvtepi32_ps(fix8_32);
SkASSERT(this->isValid());
}
inline SkPMColor SkPMFloat::get() const {
SkASSERT(this->isValid());
return this->clamped(); // Haven't beaten this yet.
}
inline SkPMColor SkPMFloat::clamped() const {
__m128i fix8_32 = _mm_cvtps_epi32(fColors), // _mm_cvtps_epi32 rounds for us!
fix8_16 = _mm_packus_epi16(fix8_32, fix8_32),
fix8 = _mm_packus_epi16(fix8_16, fix8_16);
SkPMColor c = _mm_cvtsi128_si32(fix8);
SkPMColorAssert(c);
return c;
}
inline void SkPMFloat::From4PMColors(SkPMFloat floats[4], const SkPMColor colors[4]) {
// Haven't beaten this yet.
for (int i = 0; i < 4; i++) { floats[i] = FromPMColor(colors[i]); }
}
inline void SkPMFloat::To4PMColors(SkPMColor colors[4], const SkPMFloat floats[4]) {
SkASSERT(floats[0].isValid() && floats[1].isValid()
&& floats[2].isValid() && floats[3].isValid());
// Haven't beaten this yet.
ClampTo4PMColors(colors, floats);
}
inline void SkPMFloat::ClampTo4PMColors(SkPMColor colors[4], const SkPMFloat floats[4]) {
// Same as _SSSE3.h's. We use 3 _mm_packus_epi16() where the naive loop uses 8.
__m128i c0 = _mm_cvtps_epi32(floats[0].fColors), // _mm_cvtps_epi32 rounds for us!
c1 = _mm_cvtps_epi32(floats[1].fColors),
c2 = _mm_cvtps_epi32(floats[2].fColors),
c3 = _mm_cvtps_epi32(floats[3].fColors);
__m128i c3210 = _mm_packus_epi16(_mm_packus_epi16(c0, c1),
_mm_packus_epi16(c2, c3));
_mm_storeu_si128((__m128i*)colors, c3210);
SkPMColorAssert(colors[0]);
SkPMColorAssert(colors[1]);
SkPMColorAssert(colors[2]);
SkPMColorAssert(colors[3]);
}
|
