[dev.simd] simd: make OpMasked machine ops only

Right now we can expect the `Op(...).Masked` idiom to lack many parts that will
make the API incomplete. But to make the API sizes smaller, we are removing these ops' frontend types and interfaces for now. We will have the peepholes and a new pass
checking the CPU features check domination relations to make these ops
picked for the right `Op(...).Masked` idiom.

Change-Id: I77f72a198b3d8b1880dcb911470db5e0089ac1ca
Reviewed-on: https://go-review.googlesource.com/c/go/+/697155
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Bypass: Junyang Shao <shaojunyang@google.com>

src/simd/_gen/simdgen/godefs.go

@@ -11,6 +11,7 @@ import (
 	"slices"
 	"strconv"
 	"strings"
+	"unicode"
 
 	"simd/_gen/unify"
 )
@@ -100,6 +101,11 @@ func (o *Operation) DecodeUnified(v *unify.Value) error {
 	o.Documentation = regexp.MustCompile(`\bNAME\b`).ReplaceAllString(o.Documentation, o.Go)
 	if isMasked {
 		o.Documentation += "\n//\n// This operation is applied selectively under a write mask."
+		if unicode.IsUpper([]rune(o.Go)[0]) {
+			trueVal := "true"
+			o.NoGenericOps = &trueVal
+			o.NoTypes = &trueVal
+		}
 	}
 
 	o.In = append(o.rawOperation.In, o.rawOperation.InVariant...)

src/simd/compare_test.go

@@ -15,44 +15,6 @@ import (
 // from > and =
 var comparisonFixed bool = simd.HasAVX512()
 
-func TestLessMasked(t *testing.T) {
-	if simd.HasAVX512() {
-		testFloat32x4CompareMasked(t, simd.Float32x4.LessMasked, lessSlice[float32])
-		testFloat32x8CompareMasked(t, simd.Float32x8.LessMasked, lessSlice[float32])
-		testFloat64x2CompareMasked(t, simd.Float64x2.LessMasked, lessSlice[float64])
-		testFloat64x4CompareMasked(t, simd.Float64x4.LessMasked, lessSlice[float64])
-
-		testInt16x16CompareMasked(t, simd.Int16x16.LessMasked, lessSlice[int16])
-		testInt16x8CompareMasked(t, simd.Int16x8.LessMasked, lessSlice[int16])
-		testInt32x4CompareMasked(t, simd.Int32x4.LessMasked, lessSlice[int32])
-		testInt32x8CompareMasked(t, simd.Int32x8.LessMasked, lessSlice[int32])
-		testInt64x2CompareMasked(t, simd.Int64x2.LessMasked, lessSlice[int64])
-		testInt64x4CompareMasked(t, simd.Int64x4.LessMasked, lessSlice[int64])
-		testInt8x16CompareMasked(t, simd.Int8x16.LessMasked, lessSlice[int8])
-		testInt8x32CompareMasked(t, simd.Int8x32.LessMasked, lessSlice[int8])
-
-		testUint16x16CompareMasked(t, simd.Uint16x16.LessMasked, lessSlice[uint16])
-		testUint16x8CompareMasked(t, simd.Uint16x8.LessMasked, lessSlice[uint16])
-		testUint32x4CompareMasked(t, simd.Uint32x4.LessMasked, lessSlice[uint32])
-		testUint32x8CompareMasked(t, simd.Uint32x8.LessMasked, lessSlice[uint32])
-		testUint64x2CompareMasked(t, simd.Uint64x2.LessMasked, lessSlice[uint64])
-		testUint64x4CompareMasked(t, simd.Uint64x4.LessMasked, lessSlice[uint64])
-		testUint8x16CompareMasked(t, simd.Uint8x16.LessMasked, lessSlice[uint8])
-		testUint8x32CompareMasked(t, simd.Uint8x32.LessMasked, lessSlice[uint8])
-
-		testFloat32x16CompareMasked(t, simd.Float32x16.LessMasked, lessSlice[float32])
-		testFloat64x8CompareMasked(t, simd.Float64x8.LessMasked, lessSlice[float64])
-		testInt8x64CompareMasked(t, simd.Int8x64.LessMasked, lessSlice[int8])
-		testInt16x32CompareMasked(t, simd.Int16x32.LessMasked, lessSlice[int16])
-		testInt32x16CompareMasked(t, simd.Int32x16.LessMasked, lessSlice[int32])
-		testInt64x8CompareMasked(t, simd.Int64x8.LessMasked, lessSlice[int64])
-		testUint8x64CompareMasked(t, simd.Uint8x64.LessMasked, lessSlice[uint8])
-		testUint16x32CompareMasked(t, simd.Uint16x32.LessMasked, lessSlice[uint16])
-		testUint32x16CompareMasked(t, simd.Uint32x16.LessMasked, lessSlice[uint32])
-		testUint64x8CompareMasked(t, simd.Uint64x8.LessMasked, lessSlice[uint64])
-	}
-}
-
 func TestLess(t *testing.T) {
 	testFloat32x4Compare(t, simd.Float32x4.Less, lessSlice[float32])
 	testFloat32x8Compare(t, simd.Float32x8.Less, lessSlice[float32])

src/simd/simd_test.go

@@ -43,7 +43,7 @@ func TestType(t *testing.T) {
 		return
 	}
 	v.z = maskT(simd.Mask32x4FromBits(0b0011))
-	*v.y = v.y.AddMasked(v.x, simd.Mask32x4(v.z))
+	*v.y = v.y.Add(v.x).Masked(simd.Mask32x4(v.z))
 
 	got := [4]int32{}
 	v.y.Store(&got)
@@ -121,7 +121,7 @@ func TestMaskConversion(t *testing.T) {
 	}
 	x := simd.LoadInt32x4Slice([]int32{5, 0, 7, 0})
 	mask := simd.Int32x4{}.Sub(x).ToMask()
-	y := simd.LoadInt32x4Slice([]int32{1, 2, 3, 4}).AddMasked(x, mask)
+	y := simd.LoadInt32x4Slice([]int32{1, 2, 3, 4}).Add(x).Masked(mask)
 	want := [4]int32{6, 0, 10, 0}
 	got := make([]int32, 4)
 	y.StoreSlice(got)
@@ -327,7 +327,7 @@ func TestBitMaskLoad(t *testing.T) {
 	results := [2]int64{}
 	want := [2]int64{0, 6}
 	m := simd.LoadMask64x2FromBits(&bits)
-	simd.LoadInt64x2Slice([]int64{1, 2}).AddMasked(simd.LoadInt64x2Slice([]int64{3, 4}), m).Store(&results)
+	simd.LoadInt64x2Slice([]int64{1, 2}).Add(simd.LoadInt64x2Slice([]int64{3, 4})).Masked(m).Store(&results)
 	for i := range 2 {
 		if results[i] != want[i] {
 			t.Errorf("Result at %d incorrect: want %v, got %v", i, want[i], results[i])
@@ -359,7 +359,7 @@ func TestBitMaskFromBits(t *testing.T) {
 	results := [2]int64{}
 	want := [2]int64{0, 6}
 	m := simd.Mask64x2FromBits(0b10)
-	simd.LoadInt64x2Slice([]int64{1, 2}).AddMasked(simd.LoadInt64x2Slice([]int64{3, 4}), m).Store(&results)
+	simd.LoadInt64x2Slice([]int64{1, 2}).Add(simd.LoadInt64x2Slice([]int64{3, 4})).Masked(m).Store(&results)
 	for i := range 2 {
 		if results[i] != want[i] {
 			t.Errorf("Result at %d incorrect: want %v, got %v", i, want[i], results[i])