feat(arch): add the scalar FP16 memory forms to the amd64 extension layer

Assisted-by: GLM 5.3 Flash

asm/extension_amd64_test.go

@@ -25,8 +25,8 @@ func TestAmd64ExtensionRegistry(t *testing.T) {
 		{"VDPBF16PS", 3},
 		{"VP2INTERSECTD", 3},
 		{"VP2INTERSECTQ", 3},
-		{"VMOVSH", 1},
-		{"VMOVW", 2},
+		{"VMOVSH", 3},
+		{"VMOVW", 4},
 		{"VADDSH", 1},
 		{"VSQRTSH", 1},
 		{"VCOMISH", 1},
@@ -54,8 +54,8 @@ func TestAmd64ExtensionRegistry(t *testing.T) {
 			t.Errorf("the %s lookup is not case-insensitive", tt.mnem)
 		}
 	}
-	if got := arch.Extensions(arch.AMD64); len(got) != 66 {
-		t.Errorf("the amd64 layer registers %d instructions, want 66", len(got))
+	if got := arch.Extensions(arch.AMD64); len(got) != 70 {
+		t.Errorf("the amd64 layer registers %d instructions, want 70", len(got))
 	}
 	if _, ok := LookupExtension(arch.AMD64, "NOSUCHINSTR"); ok {
 		t.Error("a non-extended mnemonic resolved")