feat(asm): add VEX shuffle, extract, permute, move and FP forms

Assisted-by: Qwen 3.8 Max Preview

asm/assemble_test.go

@@ -199,3 +199,48 @@ TEXT ·withframe(SB), NOSPLIT, $16-16
 		t.Errorf("frame translation mismatch:\n got:  %s\n want: %s", hexBytes(code), hexBytes(want))
 	}
 }
+
+// TestAssembleVexKernel assembles the horizontal-sum reduction the go-flac
+// kernels end with — exercising the VEX moves, shuffle and extract forms
+// through the full parser → encoder path — and checks the output is
+// byte-identical to the Go assembler's.
+func TestAssembleVexKernel(t *testing.T) {
+	fn := firstText(t, `
+#include "textflag.h"
+TEXT ·hsum(SB), NOSPLIT, $0
+	VPADDQ       Y8, Y9, Y8
+	VEXTRACTI128 $1, Y8, X9
+	VPADDQ       X9, X8, X8
+	VPSHUFD      $0xEE, X8, X9
+	VPADDQ       X9, X8, X8
+	VMOVQ        X8, AX
+	VZEROUPPER
+	RET
+`)
+	code, _, err := Assemble(fn)
+	if err != nil {
+		t.Fatalf("Assemble: %v", err)
+	}
+	// From the Go-assembled function:
+	//   VPADDQ Y8, Y9, Y8      c44135d4c0
+	//   VEXTRACTI128 $1, Y8, X9  c4437d39c101
+	//   VPADDQ X9, X8, X8      c44139d4c1
+	//   VPSHUFD $0xEE, X8, X9  c4417970c8ee
+	//   VPADDQ X9, X8, X8      c44139d4c1
+	//   VMOVQ X8, AX           c461f97ec0
+	//   VZEROUPPER             c5f877
+	//   RET                    c3
+	want := []byte{
+		0xc4, 0x41, 0x35, 0xd4, 0xc0,
+		0xc4, 0x43, 0x7d, 0x39, 0xc1, 0x01,
+		0xc4, 0x41, 0x39, 0xd4, 0xc1,
+		0xc4, 0x41, 0x79, 0x70, 0xc8, 0xee,
+		0xc4, 0x41, 0x39, 0xd4, 0xc1,
+		0xc4, 0x61, 0xf9, 0x7e, 0xc0,
+		0xc5, 0xf8, 0x77,
+		0xc3,
+	}
+	if hexBytes(code) != hexBytes(want) {
+		t.Errorf("VEX kernel mismatch:\n got:  %s\n want: %s", hexBytes(code), hexBytes(want))
+	}
+}