image/color: optimize RGBToYCbCr

Apply optimizations used to speed up YCbCrToRGB from
https://go-review.googlesource.com/#/c/21910/
to RGBToYCbCr.

name             old time/op  new time/op  delta
RGBToYCbCr/0-2   6.81ns ± 0%  5.96ns ± 0%  -12.48%  (p=0.000 n=38+50)
RGBToYCbCr/Cb-2  7.68ns ± 0%  6.13ns ± 0%  -20.21%  (p=0.000 n=50+33)
RGBToYCbCr/Cr-2  6.84ns ± 0%  6.04ns ± 0%  -11.70%  (p=0.000 n=39+42)

Updates #15260

Change-Id: If3ea5393ae371a955ddf18ab226aae20b48f9692
Reviewed-on: https://go-review.googlesource.com/22411
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ralph Corderoy <ralph@inputplus.co.uk>

2 files changed, 53 insertions, 19 deletions

diff --git a/src/image/color/ycbcr.go b/src/image/color/ycbcr.go
index 55549c3fce..3df5d3675d 100644
--- a/src/image/color/ycbcr.go
+++ b/src/image/color/ycbcr.go
@@ -15,24 +15,37 @@ func RGBToYCbCr(r, g, b uint8) (uint8, uint8, uint8) {
 	r1 := int32(r)
 	g1 := int32(g)
 	b1 := int32(b)
+
+	// yy is in range [0,0xff].
 	yy := (19595*r1 + 38470*g1 + 7471*b1 + 1<<15) >> 16
-	cb := (-11056*r1 - 21712*g1 + 32768*b1 + 257<<15) >> 16
-	cr := (32768*r1 - 27440*g1 - 5328*b1 + 257<<15) >> 16
-	if yy < 0 {
-		yy = 0
-	} else if yy > 0xff {
-		yy = 0xff
-	}
-	if cb < 0 {
-		cb = 0
-	} else if cb > 0xff {
-		cb = 0xff
+
+	// The bit twiddling below is equivalent to
+	//
+	// cb := (-11056*r1 - 21712*g1 + 32768*b1 + 257<<15) >> 16
+	// if cb < 0 {
+	//     cb = 0
+	// } else if cb > 0xff {
+	//     cb = ^int32(0)
+	// }
+	//
+	// but uses fewer branches and is faster.
+	// Note that the uint8 type conversion in the return
+	// statement will convert ^int32(0) to 0xff.
+	// The code below to compute cr uses a similar pattern.
+	cb := -11056*r1 - 21712*g1 + 32768*b1 + 257<<15
+	if uint32(cb)&0xff000000 == 0 {
+		cb >>= 16
+	} else {
+		cb = ^(cb >> 31)
 	}
-	if cr < 0 {
-		cr = 0
-	} else if cr > 0xff {
-		cr = 0xff
+
+	cr := 32768*r1 - 27440*g1 - 5328*b1 + 257<<15
+	if uint32(cr)&0xff000000 == 0 {
+		cr >>= 16
+	} else {
+		cr = ^(cr >> 31)
 	}
+
 	return uint8(yy), uint8(cb), uint8(cr)
 }
 
diff --git a/src/image/color/ycbcr_test.go b/src/image/color/ycbcr_test.go
index 1b110691a2..561699f4e0 100644
--- a/src/image/color/ycbcr_test.go
+++ b/src/image/color/ycbcr_test.go
@@ -172,7 +172,7 @@ func TestPalette(t *testing.T) {
 	}
 }
 
-var sinkr, sinkg, sinkb uint8
+var sink uint8
 
 func BenchmarkYCbCrToRGB(b *testing.B) {
 	// YCbCrToRGB does saturating arithmetic.
@@ -180,17 +180,38 @@ func BenchmarkYCbCrToRGB(b *testing.B) {
 	// different paths through the generated code.
 	b.Run("0", func(b *testing.B) {
 		for i := 0; i < b.N; i++ {
-			sinkr, sinkg, sinkb = YCbCrToRGB(0, 0, 0)
+			sink, sink, sink = YCbCrToRGB(0, 0, 0)
 		}
 	})
 	b.Run("128", func(b *testing.B) {
 		for i := 0; i < b.N; i++ {
-			sinkr, sinkg, sinkb = YCbCrToRGB(128, 128, 128)
+			sink, sink, sink = YCbCrToRGB(128, 128, 128)
 		}
 	})
 	b.Run("255", func(b *testing.B) {
 		for i := 0; i < b.N; i++ {
-			sinkr, sinkg, sinkb = YCbCrToRGB(255, 255, 255)
+			sink, sink, sink = YCbCrToRGB(255, 255, 255)
+		}
+	})
+}
+
+func BenchmarkRGBToYCbCr(b *testing.B) {
+	// RGBToYCbCr does saturating arithmetic.
+	// Different values can take different paths
+	// through the generated code.
+	b.Run("0", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			sink, sink, sink = RGBToYCbCr(0, 0, 0)
+		}
+	})
+	b.Run("Cb", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			sink, sink, sink = RGBToYCbCr(0, 0, 255)
+		}
+	})
+	b.Run("Cr", func(b *testing.B) {
+		for i := 0; i < b.N; i++ {
+			sink, sink, sink = RGBToYCbCr(255, 0, 0)
 		}
 	})
 }