- Add support for the rest of AVX SSE3 instructions

- Fix VEX prefix to be emitted with 3 bytes whenever VEX_5M
represents a REX equivalent two byte leading opcode

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@107523 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/X86/X86InstrSSE.td b/lib/Target/X86/X86InstrSSE.td
index a06b72a25178476c378c12ff4b8e7abb129a7516..31cf1963b801e8bf10cab23a5782c20f8af14cca 100644 (file)
--- a/lib/Target/X86/X86InstrSSE.td
+++ b/lib/Target/X86/X86InstrSSE.td
@@ -3380,6 +3380,18 @@ multiclass SS3I_unop_rm_int<bits<8> opc, string OpcodeStr,
                        (bitconvert (mem_frag128 addr:$src))))]>, OpSize;
 }
 
+let isAsmParserOnly = 1, Predicates = [HasAVX, HasSSE3] in {
+  defm VPABSB  : SS3I_unop_rm_int<0x1C, "vpabsb", memopv8i8, memopv16i8,
+                                  int_x86_ssse3_pabs_b,
+                                  int_x86_ssse3_pabs_b_128>, VEX;
+  defm VPABSW  : SS3I_unop_rm_int<0x1D, "vpabsw", memopv4i16, memopv8i16,
+                                  int_x86_ssse3_pabs_w,
+                                  int_x86_ssse3_pabs_w_128>, VEX;
+  defm VPABSD  : SS3I_unop_rm_int<0x1E, "vpabsd", memopv2i32, memopv4i32,
+                                  int_x86_ssse3_pabs_d,
+                                  int_x86_ssse3_pabs_d_128>, VEX;
+}
+
 defm PABSB       : SS3I_unop_rm_int<0x1C, "pabsb", memopv8i8, memopv16i8,
                                     int_x86_ssse3_pabs_b,
                                     int_x86_ssse3_pabs_b_128>;
@@ -3433,6 +3445,47 @@ multiclass SS3I_binop_rm_int<bits<8> opc, string OpcodeStr,
           (bitconvert (memopv16i8 addr:$src2))))]>, OpSize;
 }
 
+let isAsmParserOnly = 1, Predicates = [HasAVX, HasSSE3] in {
+let isCommutable = 0 in {
+  defm VPHADDW    : SS3I_binop_rm_int<0x01, "vphaddw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_phadd_w,
+                                      int_x86_ssse3_phadd_w_128, 0>, VEX_4V;
+  defm VPHADDD    : SS3I_binop_rm_int<0x02, "vphaddd", memopv2i32, memopv4i32,
+                                      int_x86_ssse3_phadd_d,
+                                      int_x86_ssse3_phadd_d_128, 0>, VEX_4V;
+  defm VPHADDSW   : SS3I_binop_rm_int<0x03, "vphaddsw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_phadd_sw,
+                                      int_x86_ssse3_phadd_sw_128, 0>, VEX_4V;
+  defm VPHSUBW    : SS3I_binop_rm_int<0x05, "vphsubw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_phsub_w,
+                                      int_x86_ssse3_phsub_w_128, 0>, VEX_4V;
+  defm VPHSUBD    : SS3I_binop_rm_int<0x06, "vphsubd", memopv2i32, memopv4i32,
+                                      int_x86_ssse3_phsub_d,
+                                      int_x86_ssse3_phsub_d_128, 0>, VEX_4V;
+  defm VPHSUBSW   : SS3I_binop_rm_int<0x07, "vphsubsw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_phsub_sw,
+                                      int_x86_ssse3_phsub_sw_128, 0>, VEX_4V;
+  defm VPMADDUBSW : SS3I_binop_rm_int<0x04, "vpmaddubsw", memopv8i8, memopv16i8,
+                                      int_x86_ssse3_pmadd_ub_sw,
+                                      int_x86_ssse3_pmadd_ub_sw_128, 0>, VEX_4V;
+  defm VPSHUFB    : SS3I_binop_rm_int<0x00, "vpshufb", memopv8i8, memopv16i8,
+                                      int_x86_ssse3_pshuf_b,
+                                      int_x86_ssse3_pshuf_b_128, 0>, VEX_4V;
+  defm VPSIGNB    : SS3I_binop_rm_int<0x08, "vpsignb", memopv8i8, memopv16i8,
+                                      int_x86_ssse3_psign_b,
+                                      int_x86_ssse3_psign_b_128, 0>, VEX_4V;
+  defm VPSIGNW    : SS3I_binop_rm_int<0x09, "vpsignw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_psign_w,
+                                      int_x86_ssse3_psign_w_128, 0>, VEX_4V;
+  defm VPSIGND    : SS3I_binop_rm_int<0x0A, "vpsignd", memopv2i32, memopv4i32,
+                                      int_x86_ssse3_psign_d,
+                                      int_x86_ssse3_psign_d_128, 0>, VEX_4V;
+}
+defm VPMULHRSW    : SS3I_binop_rm_int<0x0B, "vpmulhrsw", memopv4i16, memopv8i16,
+                                      int_x86_ssse3_pmul_hr_sw,
+                                      int_x86_ssse3_pmul_hr_sw_128, 0>, VEX_4V;
+}
+
 // None of these have i8 immediate fields.
 let ImmT = NoImm, Constraints = "$src1 = $dst" in {
 let isCommutable = 0 in {
@@ -3484,26 +3537,43 @@ def : Pat<(X86pshufb VR128:$src, (bc_v16i8 (memopv2i64 addr:$mask))),
 // SSSE3 - Packed Align Instruction Patterns
 //===---------------------------------------------------------------------===//
 
-let Constraints = "$src1 = $dst" in {
-  def PALIGNR64rr  : SS3AI<0x0F, MRMSrcReg, (outs VR64:$dst),
-                           (ins VR64:$src1, VR64:$src2, i8imm:$src3),
-                           "palignr\t{$src3, $src2, $dst|$dst, $src2, $src3}",
-                           []>;
-  def PALIGNR64rm  : SS3AI<0x0F, MRMSrcMem, (outs VR64:$dst),
-                           (ins VR64:$src1, i64mem:$src2, i8imm:$src3),
-                           "palignr\t{$src3, $src2, $dst|$dst, $src2, $src3}",
-                           []>;
-
-  def PALIGNR128rr : SS3AI<0x0F, MRMSrcReg, (outs VR128:$dst),
-                           (ins VR128:$src1, VR128:$src2, i8imm:$src3),
-                           "palignr\t{$src3, $src2, $dst|$dst, $src2, $src3}",
-                           []>, OpSize;
-  def PALIGNR128rm : SS3AI<0x0F, MRMSrcMem, (outs VR128:$dst),
-                           (ins VR128:$src1, i128mem:$src2, i8imm:$src3),
-                           "palignr\t{$src3, $src2, $dst|$dst, $src2, $src3}",
-                           []>, OpSize;
+multiclass sse3_palign<string asm, bit Is2Addr = 1> {
+  def R64rr  : SS3AI<0x0F, MRMSrcReg, (outs VR64:$dst),
+      (ins VR64:$src1, VR64:$src2, i8imm:$src3),
+      !if(Is2Addr,
+        !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
+        !strconcat(asm,
+                  "\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}")),
+      []>;
+  def R64rm  : SS3AI<0x0F, MRMSrcMem, (outs VR64:$dst),
+      (ins VR64:$src1, i64mem:$src2, i8imm:$src3),
+      !if(Is2Addr,
+        !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
+        !strconcat(asm,
+                  "\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}")),
+      []>;
+
+  def R128rr : SS3AI<0x0F, MRMSrcReg, (outs VR128:$dst),
+      (ins VR128:$src1, VR128:$src2, i8imm:$src3),
+      !if(Is2Addr,
+        !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
+        !strconcat(asm,
+                  "\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}")),
+      []>, OpSize;
+  def R128rm : SS3AI<0x0F, MRMSrcMem, (outs VR128:$dst),
+      (ins VR128:$src1, i128mem:$src2, i8imm:$src3),
+      !if(Is2Addr,
+        !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
+        !strconcat(asm,
+                  "\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}")),
+      []>, OpSize;
 }
 
+let isAsmParserOnly = 1, Predicates = [HasAVX, HasSSE3] in
+  defm VPALIGN : sse3_palign<"vpalignr", 0>, VEX_4V;
+let Constraints = "$src1 = $dst" in
+  defm PALIGN : sse3_palign<"palignr">;
+
 let AddedComplexity = 5 in {
 
 def : Pat<(v1i64 (palign:$src3 VR64:$src1, VR64:$src2)),

diff --git a/lib/Target/X86/X86MCCodeEmitter.cpp b/lib/Target/X86/X86MCCodeEmitter.cpp
index 9f36aaa28c7c9060128e22fbeff5c502296e45de..f60e73968dc01f008a6fcc9b7b72a2f965a0439a 100644 (file)
--- a/lib/Target/X86/X86MCCodeEmitter.cpp
+++ b/lib/Target/X86/X86MCCodeEmitter.cpp
@@ -543,7 +543,7 @@ void X86MCCodeEmitter::EmitVEXOpcodePrefix(uint64_t TSFlags, unsigned &CurByte,
   //
   unsigned char LastByte = VEX_PP | (VEX_L << 2) | (VEX_4V << 3);
 
-  if (VEX_B && VEX_X && !VEX_W) { // 2 byte VEX prefix
+  if (VEX_B && VEX_X && !VEX_W && (VEX_5M == 1)) { // 2 byte VEX prefix
     EmitByte(0xC5, CurByte, OS);
     EmitByte(LastByte | (VEX_R << 7), CurByte, OS);
     return;

diff --git a/test/MC/AsmParser/X86/x86_32-encoding.s b/test/MC/AsmParser/X86/x86_32-encoding.s
index b2789cda3f17e4e01c4d4b527826de09e3faa43a..6854d40f230178e1259e5ed92bf75b08578875b2 100644 (file)
--- a/test/MC/AsmParser/X86/x86_32-encoding.s
+++ b/test/MC/AsmParser/X86/x86_32-encoding.s
@@ -11614,3 +11614,131 @@
 // CHECK: encoding: [0xc5,0xe9,0x7d,0x18]
           vhsubpd  (%eax), %xmm2, %xmm3
 
+// CHECK: vpabsb  %xmm1, %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1c,0xd1]
+          vpabsb  %xmm1, %xmm2
+
+// CHECK: vpabsb  (%eax), %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1c,0x10]
+          vpabsb  (%eax), %xmm2
+
+// CHECK: vpabsw  %xmm1, %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1d,0xd1]
+          vpabsw  %xmm1, %xmm2
+
+// CHECK: vpabsw  (%eax), %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1d,0x10]
+          vpabsw  (%eax), %xmm2
+
+// CHECK: vpabsd  %xmm1, %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1e,0xd1]
+          vpabsd  %xmm1, %xmm2
+
+// CHECK: vpabsd  (%eax), %xmm2
+// CHECK: encoding: [0xc4,0xe2,0x79,0x1e,0x10]
+          vpabsd  (%eax), %xmm2
+
+// CHECK: vphaddw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x01,0xd9]
+          vphaddw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphaddw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x01,0x18]
+          vphaddw  (%eax), %xmm2, %xmm3
+
+// CHECK: vphaddd  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x02,0xd9]
+          vphaddd  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphaddd  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x02,0x18]
+          vphaddd  (%eax), %xmm2, %xmm3
+
+// CHECK: vphaddsw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x03,0xd9]
+          vphaddsw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphaddsw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x03,0x18]
+          vphaddsw  (%eax), %xmm2, %xmm3
+
+// CHECK: vphsubw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x05,0xd9]
+          vphsubw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphsubw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x05,0x18]
+          vphsubw  (%eax), %xmm2, %xmm3
+
+// CHECK: vphsubd  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x06,0xd9]
+          vphsubd  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphsubd  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x06,0x18]
+          vphsubd  (%eax), %xmm2, %xmm3
+
+// CHECK: vphsubsw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x07,0xd9]
+          vphsubsw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vphsubsw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x07,0x18]
+          vphsubsw  (%eax), %xmm2, %xmm3
+
+// CHECK: vpmaddubsw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x04,0xd9]
+          vpmaddubsw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpmaddubsw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x04,0x18]
+          vpmaddubsw  (%eax), %xmm2, %xmm3
+
+// CHECK: vpshufb  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x00,0xd9]
+          vpshufb  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpshufb  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x00,0x18]
+          vpshufb  (%eax), %xmm2, %xmm3
+
+// CHECK: vpsignb  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x08,0xd9]
+          vpsignb  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpsignb  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x08,0x18]
+          vpsignb  (%eax), %xmm2, %xmm3
+
+// CHECK: vpsignw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x09,0xd9]
+          vpsignw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpsignw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x09,0x18]
+          vpsignw  (%eax), %xmm2, %xmm3
+
+// CHECK: vpsignd  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x0a,0xd9]
+          vpsignd  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpsignd  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x0a,0x18]
+          vpsignd  (%eax), %xmm2, %xmm3
+
+// CHECK: vpmulhrsw  %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x0b,0xd9]
+          vpmulhrsw  %xmm1, %xmm2, %xmm3
+
+// CHECK: vpmulhrsw  (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe2,0x69,0x0b,0x18]
+          vpmulhrsw  (%eax), %xmm2, %xmm3
+
+// CHECK: vpalignr  $7, %xmm1, %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe3,0x69,0x0f,0xd9,0x07]
+          vpalignr  $7, %xmm1, %xmm2, %xmm3
+
+// CHECK: vpalignr  $7, (%eax), %xmm2, %xmm3
+// CHECK: encoding: [0xc4,0xe3,0x69,0x0f,0x18,0x07]
+          vpalignr  $7, (%eax), %xmm2, %xmm3
+

diff --git a/test/MC/AsmParser/X86/x86_64-encoding.s b/test/MC/AsmParser/X86/x86_64-encoding.s
index 8be9d67d99281b945902d3f0d1ce0616c654c2f0..af33f74ed8a424185e61b559ab6fd7f08c6fbf42 100644 (file)
--- a/test/MC/AsmParser/X86/x86_64-encoding.s
+++ b/test/MC/AsmParser/X86/x86_64-encoding.s
@@ -1662,3 +1662,131 @@ pshufb  CPI1_0(%rip), %xmm1
 // CHECK: encoding: [0xc5,0x19,0x7d,0x28]
           vhsubpd  (%rax), %xmm12, %xmm13
 
+// CHECK: vpabsb  %xmm11, %xmm12
+// CHECK: encoding: [0xc4,0x42,0x79,0x1c,0xe3]
+          vpabsb  %xmm11, %xmm12
+
+// CHECK: vpabsb  (%rax), %xmm12
+// CHECK: encoding: [0xc4,0x62,0x79,0x1c,0x20]
+          vpabsb  (%rax), %xmm12
+
+// CHECK: vpabsw  %xmm11, %xmm12
+// CHECK: encoding: [0xc4,0x42,0x79,0x1d,0xe3]
+          vpabsw  %xmm11, %xmm12
+
+// CHECK: vpabsw  (%rax), %xmm12
+// CHECK: encoding: [0xc4,0x62,0x79,0x1d,0x20]
+          vpabsw  (%rax), %xmm12
+
+// CHECK: vpabsd  %xmm11, %xmm12
+// CHECK: encoding: [0xc4,0x42,0x79,0x1e,0xe3]
+          vpabsd  %xmm11, %xmm12
+
+// CHECK: vpabsd  (%rax), %xmm12
+// CHECK: encoding: [0xc4,0x62,0x79,0x1e,0x20]
+          vpabsd  (%rax), %xmm12
+
+// CHECK: vphaddw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x01,0xeb]
+          vphaddw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphaddw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x01,0x28]
+          vphaddw  (%rax), %xmm12, %xmm13
+
+// CHECK: vphaddd  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x02,0xeb]
+          vphaddd  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphaddd  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x02,0x28]
+          vphaddd  (%rax), %xmm12, %xmm13
+
+// CHECK: vphaddsw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x03,0xeb]
+          vphaddsw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphaddsw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x03,0x28]
+          vphaddsw  (%rax), %xmm12, %xmm13
+
+// CHECK: vphsubw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x05,0xeb]
+          vphsubw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphsubw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x05,0x28]
+          vphsubw  (%rax), %xmm12, %xmm13
+
+// CHECK: vphsubd  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x06,0xeb]
+          vphsubd  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphsubd  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x06,0x28]
+          vphsubd  (%rax), %xmm12, %xmm13
+
+// CHECK: vphsubsw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x07,0xeb]
+          vphsubsw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vphsubsw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x07,0x28]
+          vphsubsw  (%rax), %xmm12, %xmm13
+
+// CHECK: vpmaddubsw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x04,0xeb]
+          vpmaddubsw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpmaddubsw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x04,0x28]
+          vpmaddubsw  (%rax), %xmm12, %xmm13
+
+// CHECK: vpshufb  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x00,0xeb]
+          vpshufb  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpshufb  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x00,0x28]
+          vpshufb  (%rax), %xmm12, %xmm13
+
+// CHECK: vpsignb  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x08,0xeb]
+          vpsignb  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpsignb  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x08,0x28]
+          vpsignb  (%rax), %xmm12, %xmm13
+
+// CHECK: vpsignw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x09,0xeb]
+          vpsignw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpsignw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x09,0x28]
+          vpsignw  (%rax), %xmm12, %xmm13
+
+// CHECK: vpsignd  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x0a,0xeb]
+          vpsignd  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpsignd  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x0a,0x28]
+          vpsignd  (%rax), %xmm12, %xmm13
+
+// CHECK: vpmulhrsw  %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x42,0x19,0x0b,0xeb]
+          vpmulhrsw  %xmm11, %xmm12, %xmm13
+
+// CHECK: vpmulhrsw  (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x62,0x19,0x0b,0x28]
+          vpmulhrsw  (%rax), %xmm12, %xmm13
+
+// CHECK: vpalignr  $7, %xmm11, %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x43,0x19,0x0f,0xeb,0x07]
+          vpalignr  $7, %xmm11, %xmm12, %xmm13
+
+// CHECK: vpalignr  $7, (%rax), %xmm12, %xmm13
+// CHECK: encoding: [0xc4,0x63,0x19,0x0f,0x28,0x07]
+          vpalignr  $7, (%rax), %xmm12, %xmm13
+