diff --git a/src/cmd/compile/internal/gc/ssa.go b/src/cmd/compile/internal/gc/ssa.go
index f76b6d4c021..6b5a6a3bdf2 100644
--- a/src/cmd/compile/internal/gc/ssa.go
+++ b/src/cmd/compile/internal/gc/ssa.go
@@ -3561,12 +3561,12 @@ func init() {
 			return s.newValue2(ssa.OpCopysign, types.Types[TFLOAT64], args[0], args[1])
 		},
 		sys.PPC64, sys.Wasm)
-	addF("math", "Fma",
+	addF("math", "FMA",
 		func(s *state, n *Node, args []*ssa.Value) *ssa.Value {
-			return s.newValue3(ssa.OpFma, types.Types[TFLOAT64], args[0], args[1], args[2])
+			return s.newValue3(ssa.OpFMA, types.Types[TFLOAT64], args[0], args[1], args[2])
 		},
 		sys.ARM64, sys.PPC64, sys.S390X)
-	addF("math", "Fma",
+	addF("math", "FMA",
 		func(s *state, n *Node, args []*ssa.Value) *ssa.Value {
 			if !s.config.UseFMA {
 				a := s.call(n, callNormal)
@@ -3587,7 +3587,7 @@ func init() {
 
 			// We have the intrinsic - use it directly.
 			s.startBlock(bTrue)
-			s.vars[n] = s.newValue3(ssa.OpFma, types.Types[TFLOAT64], args[0], args[1], args[2])
+			s.vars[n] = s.newValue3(ssa.OpFMA, types.Types[TFLOAT64], args[0], args[1], args[2])
 			s.endBlock().AddEdgeTo(bEnd)
 
 			// Call the pure Go version.
@@ -3601,7 +3601,7 @@ func init() {
 			return s.variable(n, types.Types[TFLOAT64])
 		},
 		sys.AMD64)
-	addF("math", "Fma",
+	addF("math", "FMA",
 		func(s *state, n *Node, args []*ssa.Value) *ssa.Value {
 			if !s.config.UseFMA {
 				a := s.call(n, callNormal)
@@ -3622,7 +3622,7 @@ func init() {
 
 			// We have the intrinsic - use it directly.
 			s.startBlock(bTrue)
-			s.vars[n] = s.newValue3(ssa.OpFma, types.Types[TFLOAT64], args[0], args[1], args[2])
+			s.vars[n] = s.newValue3(ssa.OpFMA, types.Types[TFLOAT64], args[0], args[1], args[2])
 			s.endBlock().AddEdgeTo(bEnd)
 
 			// Call the pure Go version.
diff --git a/src/cmd/compile/internal/ssa/gen/AMD64.rules b/src/cmd/compile/internal/ssa/gen/AMD64.rules
index 65f229169a7..2d662d5ae68 100644
--- a/src/cmd/compile/internal/ssa/gen/AMD64.rules
+++ b/src/cmd/compile/internal/ssa/gen/AMD64.rules
@@ -110,7 +110,7 @@
 (Floor x)	-> (ROUNDSD [1] x)
 (Ceil x)  	-> (ROUNDSD [2] x)
 (Trunc x) 	-> (ROUNDSD [3] x)
-(Fma x y z) -> (VFMADD231SD z x y)
+(FMA x y z) -> (VFMADD231SD z x y)
 
 // Lowering extension
 // Note: we always extend to 64 bits even though some ops don't need that many result bits.
diff --git a/src/cmd/compile/internal/ssa/gen/ARM.rules b/src/cmd/compile/internal/ssa/gen/ARM.rules
index c1c73e23ec2..160545e02b8 100644
--- a/src/cmd/compile/internal/ssa/gen/ARM.rules
+++ b/src/cmd/compile/internal/ssa/gen/ARM.rules
@@ -211,7 +211,7 @@
 (Round(32|64)F x) -> x
 
 // fused-multiply-add
-(Fma x y z) -> (FMULAD z x y)
+(FMA x y z) -> (FMULAD z x y)
 
 // comparisons
 (Eq8 x y)  -> (Equal (CMP (ZeroExt8to32 x) (ZeroExt8to32 y)))
diff --git a/src/cmd/compile/internal/ssa/gen/ARM64.rules b/src/cmd/compile/internal/ssa/gen/ARM64.rules
index f0033a05269..a9bf64488e1 100644
--- a/src/cmd/compile/internal/ssa/gen/ARM64.rules
+++ b/src/cmd/compile/internal/ssa/gen/ARM64.rules
@@ -90,7 +90,7 @@
 (Round x) -> (FRINTAD x)
 (RoundToEven x) -> (FRINTND x)
 (Trunc x) -> (FRINTZD x)
-(Fma x y z) -> (FMADDD z x y)
+(FMA x y z) -> (FMADDD z x y)
 
 // lowering rotates
 (RotateLeft8 <t> x (MOVDconst [c])) -> (Or8 (Lsh8x64 <t> x (MOVDconst [c&7])) (Rsh8Ux64 <t> x (MOVDconst [-c&7])))
diff --git a/src/cmd/compile/internal/ssa/gen/PPC64.rules b/src/cmd/compile/internal/ssa/gen/PPC64.rules
index 13fe1ab2e91..fc370741176 100644
--- a/src/cmd/compile/internal/ssa/gen/PPC64.rules
+++ b/src/cmd/compile/internal/ssa/gen/PPC64.rules
@@ -68,7 +68,7 @@
 (Round x) -> (FROUND x)
 (Copysign x y) -> (FCPSGN y x)
 (Abs x) -> (FABS x)
-(Fma x y z) -> (FMADD x y z)
+(FMA x y z) -> (FMADD x y z)
 
 // Lowering constants
 (Const(64|32|16|8)  [val]) -> (MOVDconst [val])
@@ -1349,7 +1349,7 @@
 
 // The 2 byte store appears after the 4 byte store so that the
 // match for the 2 byte store is not done first.
-// If the 4 byte store is based on the 2 byte store then there are 
+// If the 4 byte store is based on the 2 byte store then there are
 // variations on the MOVDaddr subrule that would require additional
 // rules to be written.
 
diff --git a/src/cmd/compile/internal/ssa/gen/S390X.rules b/src/cmd/compile/internal/ssa/gen/S390X.rules
index 2c56c665814..3635aeb9154 100644
--- a/src/cmd/compile/internal/ssa/gen/S390X.rules
+++ b/src/cmd/compile/internal/ssa/gen/S390X.rules
@@ -139,7 +139,7 @@
 (Trunc       x) -> (FIDBR [5] x)
 (RoundToEven x) -> (FIDBR [4] x)
 (Round       x) -> (FIDBR [1] x)
-(Fma     x y z) -> (FMADD z x y)
+(FMA     x y z) -> (FMADD z x y)
 
 // Atomic loads and stores.
 // The SYNC instruction (fast-BCR-serialization) prevents store-load
diff --git a/src/cmd/compile/internal/ssa/gen/genericOps.go b/src/cmd/compile/internal/ssa/gen/genericOps.go
index 1ffca8118fe..aa9d5703963 100644
--- a/src/cmd/compile/internal/ssa/gen/genericOps.go
+++ b/src/cmd/compile/internal/ssa/gen/genericOps.go
@@ -314,7 +314,7 @@ var genericOps = []opData{
 	//
 	// When the multiply is an infinity times a zero, the result is NaN.
 	// See section 7.2 in ieee754.
-	{name: "Fma", argLength: 3}, // compute (a*b)+c without intermediate rounding
+	{name: "FMA", argLength: 3}, // compute (a*b)+c without intermediate rounding
 
 	// Data movement. Max argument length for Phi is indefinite.
 	{name: "Phi", argLength: -1, zeroWidth: true}, // select an argument based on which predecessor block we came from
diff --git a/src/cmd/compile/internal/ssa/opGen.go b/src/cmd/compile/internal/ssa/opGen.go
index e9a709467ec..f4370bc38d1 100644
--- a/src/cmd/compile/internal/ssa/opGen.go
+++ b/src/cmd/compile/internal/ssa/opGen.go
@@ -2428,7 +2428,7 @@ const (
 	OpRoundToEven
 	OpAbs
 	OpCopysign
-	OpFma
+	OpFMA
 	OpPhi
 	OpCopy
 	OpConvert
@@ -30743,7 +30743,7 @@ var opcodeTable = [...]opInfo{
 		generic: true,
 	},
 	{
-		name:    "Fma",
+		name:    "FMA",
 		argLen:  3,
 		generic: true,
 	},
diff --git a/src/cmd/compile/internal/ssa/rewriteAMD64.go b/src/cmd/compile/internal/ssa/rewriteAMD64.go
index 4f02554e1a6..4945d1974e9 100644
--- a/src/cmd/compile/internal/ssa/rewriteAMD64.go
+++ b/src/cmd/compile/internal/ssa/rewriteAMD64.go
@@ -770,8 +770,8 @@ func rewriteValueAMD64(v *Value) bool {
 		return rewriteValueAMD64_OpEqPtr_0(v)
 	case OpFloor:
 		return rewriteValueAMD64_OpFloor_0(v)
-	case OpFma:
-		return rewriteValueAMD64_OpFma_0(v)
+	case OpFMA:
+		return rewriteValueAMD64_OpFMA_0(v)
 	case OpGeq16:
 		return rewriteValueAMD64_OpGeq16_0(v)
 	case OpGeq16U:
@@ -52222,8 +52222,8 @@ func rewriteValueAMD64_OpFloor_0(v *Value) bool {
 		return true
 	}
 }
-func rewriteValueAMD64_OpFma_0(v *Value) bool {
-	// match: (Fma x y z)
+func rewriteValueAMD64_OpFMA_0(v *Value) bool {
+	// match: (FMA x y z)
 	// result: (VFMADD231SD z x y)
 	for {
 		z := v.Args[2]
diff --git a/src/cmd/compile/internal/ssa/rewriteARM.go b/src/cmd/compile/internal/ssa/rewriteARM.go
index 8cb534d8f62..fb7a7570f2d 100644
--- a/src/cmd/compile/internal/ssa/rewriteARM.go
+++ b/src/cmd/compile/internal/ssa/rewriteARM.go
@@ -538,8 +538,8 @@ func rewriteValueARM(v *Value) bool {
 		return rewriteValueARM_OpEqB_0(v)
 	case OpEqPtr:
 		return rewriteValueARM_OpEqPtr_0(v)
-	case OpFma:
-		return rewriteValueARM_OpFma_0(v)
+	case OpFMA:
+		return rewriteValueARM_OpFMA_0(v)
 	case OpGeq16:
 		return rewriteValueARM_OpGeq16_0(v)
 	case OpGeq16U:
@@ -17161,8 +17161,8 @@ func rewriteValueARM_OpEqPtr_0(v *Value) bool {
 		return true
 	}
 }
-func rewriteValueARM_OpFma_0(v *Value) bool {
-	// match: (Fma x y z)
+func rewriteValueARM_OpFMA_0(v *Value) bool {
+	// match: (FMA x y z)
 	// result: (FMULAD z x y)
 	for {
 		z := v.Args[2]
diff --git a/src/cmd/compile/internal/ssa/rewriteARM64.go b/src/cmd/compile/internal/ssa/rewriteARM64.go
index e9bde5ec8a2..c372865d0aa 100644
--- a/src/cmd/compile/internal/ssa/rewriteARM64.go
+++ b/src/cmd/compile/internal/ssa/rewriteARM64.go
@@ -573,8 +573,8 @@ func rewriteValueARM64(v *Value) bool {
 		return rewriteValueARM64_OpEqPtr_0(v)
 	case OpFloor:
 		return rewriteValueARM64_OpFloor_0(v)
-	case OpFma:
-		return rewriteValueARM64_OpFma_0(v)
+	case OpFMA:
+		return rewriteValueARM64_OpFMA_0(v)
 	case OpGeq16:
 		return rewriteValueARM64_OpGeq16_0(v)
 	case OpGeq16U:
@@ -28583,8 +28583,8 @@ func rewriteValueARM64_OpFloor_0(v *Value) bool {
 		return true
 	}
 }
-func rewriteValueARM64_OpFma_0(v *Value) bool {
-	// match: (Fma x y z)
+func rewriteValueARM64_OpFMA_0(v *Value) bool {
+	// match: (FMA x y z)
 	// result: (FMADDD z x y)
 	for {
 		z := v.Args[2]
diff --git a/src/cmd/compile/internal/ssa/rewritePPC64.go b/src/cmd/compile/internal/ssa/rewritePPC64.go
index a95364ece49..4d282a6730b 100644
--- a/src/cmd/compile/internal/ssa/rewritePPC64.go
+++ b/src/cmd/compile/internal/ssa/rewritePPC64.go
@@ -183,8 +183,8 @@ func rewriteValuePPC64(v *Value) bool {
 		return rewriteValuePPC64_OpEqPtr_0(v)
 	case OpFloor:
 		return rewriteValuePPC64_OpFloor_0(v)
-	case OpFma:
-		return rewriteValuePPC64_OpFma_0(v)
+	case OpFMA:
+		return rewriteValuePPC64_OpFMA_0(v)
 	case OpGeq16:
 		return rewriteValuePPC64_OpGeq16_0(v)
 	case OpGeq16U:
@@ -2007,8 +2007,8 @@ func rewriteValuePPC64_OpFloor_0(v *Value) bool {
 		return true
 	}
 }
-func rewriteValuePPC64_OpFma_0(v *Value) bool {
-	// match: (Fma x y z)
+func rewriteValuePPC64_OpFMA_0(v *Value) bool {
+	// match: (FMA x y z)
 	// result: (FMADD x y z)
 	for {
 		z := v.Args[2]
diff --git a/src/cmd/compile/internal/ssa/rewriteS390X.go b/src/cmd/compile/internal/ssa/rewriteS390X.go
index 645e8f2d9af..429e3d5be04 100644
--- a/src/cmd/compile/internal/ssa/rewriteS390X.go
+++ b/src/cmd/compile/internal/ssa/rewriteS390X.go
@@ -168,8 +168,8 @@ func rewriteValueS390X(v *Value) bool {
 		return rewriteValueS390X_OpEqPtr_0(v)
 	case OpFloor:
 		return rewriteValueS390X_OpFloor_0(v)
-	case OpFma:
-		return rewriteValueS390X_OpFma_0(v)
+	case OpFMA:
+		return rewriteValueS390X_OpFMA_0(v)
 	case OpGeq16:
 		return rewriteValueS390X_OpGeq16_0(v)
 	case OpGeq16U:
@@ -1939,8 +1939,8 @@ func rewriteValueS390X_OpFloor_0(v *Value) bool {
 		return true
 	}
 }
-func rewriteValueS390X_OpFma_0(v *Value) bool {
-	// match: (Fma x y z)
+func rewriteValueS390X_OpFMA_0(v *Value) bool {
+	// match: (FMA x y z)
 	// result: (FMADD z x y)
 	for {
 		z := v.Args[2]
diff --git a/src/math/all_test.go b/src/math/all_test.go
index e8fa2b8b66a..e9621e6dc97 100644
--- a/src/math/all_test.go
+++ b/src/math/all_test.go
@@ -3053,11 +3053,11 @@ func TestYn(t *testing.T) {
 	}
 }
 
-func TestFma(t *testing.T) {
+func TestFMA(t *testing.T) {
 	for _, c := range fmaC {
-		got := Fma(c.x, c.y, c.z)
+		got := FMA(c.x, c.y, c.z)
 		if !alike(got, c.want) {
-			t.Errorf("Fma(%g,%g,%g) == %g; want %g", c.x, c.y, c.z, got, c.want)
+			t.Errorf("FMA(%g,%g,%g) == %g; want %g", c.x, c.y, c.z, got, c.want)
 		}
 	}
 }
@@ -3793,10 +3793,10 @@ func BenchmarkFloat32frombits(b *testing.B) {
 	GlobalF = float64(x)
 }
 
-func BenchmarkFma(b *testing.B) {
+func BenchmarkFMA(b *testing.B) {
 	x := 0.0
 	for i := 0; i < b.N; i++ {
-		x = Fma(E, Pi, x)
+		x = FMA(E, Pi, x)
 	}
 	GlobalF = x
 }
diff --git a/src/math/fma.go b/src/math/fma.go
index 76249229b2b..db78dfa5e15 100644
--- a/src/math/fma.go
+++ b/src/math/fma.go
@@ -90,8 +90,9 @@ func split(b uint64) (sign uint32, exp int32, mantissa uint64) {
 	return
 }
 
-// Fma returns x * y + z, computed with only one rounding.
-func Fma(x, y, z float64) float64 {
+// FMA returns x * y + z, computed with only one rounding.
+// (That is, FMA returns the fused multiply-add of x, y, and z.)
+func FMA(x, y, z float64) float64 {
 	bx, by, bz := Float64bits(x), Float64bits(y), Float64bits(z)
 
 	// Inf or NaN or zero involved. At most one rounding will occur.
diff --git a/test/codegen/math.go b/test/codegen/math.go
index 751406d732e..80e5d60d963 100644
--- a/test/codegen/math.go
+++ b/test/codegen/math.go
@@ -114,7 +114,7 @@ func fma(x, y, z float64) float64 {
 	// s390x:"FMADD"
 	// ppc64:"FMADD"
 	// ppc64le:"FMADD"
-	return math.Fma(x, y, z)
+	return math.FMA(x, y, z)
 }
 
 func fromFloat64(f64 float64) uint64 {