--- old/src/hotspot/cpu/x86/assembler_x86.cpp	2020-04-02 18:03:43.216854097 -0700
+++ new/src/hotspot/cpu/x86/assembler_x86.cpp	2020-04-02 18:03:43.044854097 -0700
@@ -984,6 +984,8 @@
     case 0x61: // pcmpestri r, r/a, #8
     case 0x70: // pshufd r, r/a, #8
     case 0x73: // psrldq r, #8
+    case 0x1f: // evpcmpd/evpcmpq
+    case 0x3f: // evpcmpb/evpcmpw
       tail_size = 1;  // the imm8
       break;
     default:
@@ -1209,6 +1211,11 @@
   emit_int8(imm8);
 }
 
+void Assembler::addw(Register dst, Register src) {
+  (void)prefix_and_encode(dst->encoding(), src->encoding());
+  emit_arith(0x03, 0xC0, dst, src);
+}
+
 void Assembler::addw(Address dst, int imm16) {
   InstructionMark im(this);
   emit_int8(0x66);
@@ -1415,6 +1422,11 @@
   emit_int16((unsigned char)0xDD, (0xC0 | encode));
 }
 
+void Assembler::andw(Register dst, Register src) {
+  (void)prefix_and_encode(dst->encoding(), src->encoding());
+  emit_arith(0x23, 0xC0, dst, src);
+}
+
 void Assembler::andl(Address dst, int32_t imm32) {
   InstructionMark im(this);
   prefix(dst);
@@ -1783,6 +1795,13 @@
   emit_int16((unsigned char)0xE6, (0xC0 | encode));
 }
 
+void Assembler::vcvtdq2pd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xE6, (0xC0 | encode));
+}
+
 void Assembler::cvtdq2ps(XMMRegister dst, XMMRegister src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -1790,6 +1809,13 @@
   emit_int16(0x5B, (0xC0 | encode));
 }
 
+void Assembler::vcvtdq2ps(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5B, (0xC0 | encode));
+}
+
 void Assembler::cvtsd2ss(XMMRegister dst, XMMRegister src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   InstructionAttr attributes(AVX_128bit, /* rex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
@@ -1912,18 +1938,18 @@
 }
 
 void Assembler::vpabsb(XMMRegister dst, XMMRegister src, int vector_len) {
-  assert(vector_len == AVX_128bit? VM_Version::supports_avx() :
-  vector_len == AVX_256bit? VM_Version::supports_avx2() :
-  vector_len == AVX_512bit? VM_Version::supports_avx512bw() : 0, "");
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx()      :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2()     :
+         vector_len == AVX_512bit ? VM_Version::supports_avx512bw() : false, "not supported");
   InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
   emit_int16(0x1C, (0xC0 | encode));
 }
 
 void Assembler::vpabsw(XMMRegister dst, XMMRegister src, int vector_len) {
-  assert(vector_len == AVX_128bit? VM_Version::supports_avx() :
-  vector_len == AVX_256bit? VM_Version::supports_avx2() :
-  vector_len == AVX_512bit? VM_Version::supports_avx512bw() : 0, "");
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx()      :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2()     :
+         vector_len == AVX_512bit ? VM_Version::supports_avx512bw() : false, "");
   InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
   emit_int16(0x1D, (0xC0 | encode));
@@ -1946,6 +1972,85 @@
   emit_int16(0x1F, (0xC0 | encode));
 }
 
+void Assembler::vcvtps2pd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5A, (0xC0 | encode));
+}
+
+void Assembler::vcvtpd2ps(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  attributes.set_rex_vex_w_reverted();
+  emit_int16(0x5A, (0xC0 | encode));
+}
+
+void Assembler::evcvtqq2ps(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2 && VM_Version::supports_avx512dq(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5B, (0xC0 | encode));
+}
+
+void Assembler::evcvtqq2pd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2 && VM_Version::supports_avx512dq(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xE6, (0xC0 | encode));
+}
+
+void Assembler::evpmovwb(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2  && VM_Version::supports_avx512bw(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x30, (0xC0 | encode));
+}
+
+void Assembler::evpmovdw(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x33, (0xC0 | encode));
+}
+
+void Assembler::evpmovdb(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x31, (0xC0 | encode));
+}
+
+void Assembler::evpmovqd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x35, (0xC0 | encode));
+}
+
+void Assembler::evpmovqb(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x32, (0xC0 | encode));
+}
+
+void Assembler::evpmovqw(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(src->encoding(), 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x34, (0xC0 | encode));
+}
+
 void Assembler::decl(Address dst) {
   // Don't use it directly. Use MacroAssembler::decrement() instead.
   InstructionMark im(this);
@@ -2543,28 +2648,34 @@
 }
 
 // Move Unaligned EVEX enabled Vector (programmable : 8,16,32,64)
-void Assembler::evmovdqub(XMMRegister dst, XMMRegister src, int vector_len) {
+void Assembler::evmovdqub(XMMRegister dst, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   int prefix = (_legacy_mode_bw) ? VEX_SIMD_F2 : VEX_SIMD_F3;
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), (Assembler::VexSimdPrefix)prefix, VEX_OPCODE_0F, &attributes);
   emit_int16(0x6F, (0xC0 | encode));
 }
 
-void Assembler::evmovdqub(XMMRegister dst, Address src, int vector_len) {
+void Assembler::evmovdqub(XMMRegister dst, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   int prefix = (_legacy_mode_bw) ? VEX_SIMD_F2 : VEX_SIMD_F3;
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(src, 0, dst->encoding(), (Assembler::VexSimdPrefix)prefix, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
-void Assembler::evmovdqub(Address dst, XMMRegister src, int vector_len) {
+void Assembler::evmovdqub(Address dst, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   assert(src != xnoreg, "sanity");
   InstructionMark im(this);
@@ -2572,132 +2683,202 @@
   int prefix = (_legacy_mode_bw) ? VEX_SIMD_F2 : VEX_SIMD_F3;
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(dst, 0, src->encoding(), (Assembler::VexSimdPrefix)prefix, VEX_OPCODE_0F, &attributes);
   emit_int8(0x7F);
   emit_operand(src, dst);
 }
 
-void Assembler::evmovdqub(XMMRegister dst, KRegister mask, Address src, int vector_len) {
+void Assembler::evmovdqub(XMMRegister dst, KRegister mask, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_avx512vlbw(), "");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(src, 0, dst->encoding(), VEX_SIMD_F2, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
-void Assembler::evmovdquw(XMMRegister dst, Address src, int vector_len) {
+void Assembler::evmovdquw(XMMRegister dst, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   int prefix = (_legacy_mode_bw) ? VEX_SIMD_F2 : VEX_SIMD_F3;
   vex_prefix(src, 0, dst->encoding(), (Assembler::VexSimdPrefix)prefix, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
-void Assembler::evmovdquw(XMMRegister dst, KRegister mask, Address src, int vector_len) {
+void Assembler::evmovdquw(XMMRegister dst, KRegister mask, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_avx512vlbw(), "");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(src, 0, dst->encoding(), VEX_SIMD_F2, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
-void Assembler::evmovdquw(Address dst, XMMRegister src, int vector_len) {
+void Assembler::evmovdquw(Address dst, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   assert(src != xnoreg, "sanity");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   int prefix = (_legacy_mode_bw) ? VEX_SIMD_F2 : VEX_SIMD_F3;
   vex_prefix(dst, 0, src->encoding(), (Assembler::VexSimdPrefix)prefix, VEX_OPCODE_0F, &attributes);
   emit_int8(0x7F);
   emit_operand(src, dst);
 }
 
-void Assembler::evmovdquw(Address dst, KRegister mask, XMMRegister src, int vector_len) {
+void Assembler::evmovdquw(Address dst, KRegister mask, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_avx512vlbw(), "");
   assert(src != xnoreg, "sanity");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
-  attributes.reset_is_clear_context();
   attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(dst, 0, src->encoding(), VEX_SIMD_F2, VEX_OPCODE_0F, &attributes);
   emit_int8(0x7F);
   emit_operand(src, dst);
 }
 
 void Assembler::evmovdqul(XMMRegister dst, XMMRegister src, int vector_len) {
+  // Unmasked instruction
+  evmovdqul(dst, k0, src, /*merge*/ false, vector_len);
+}
+
+void Assembler::evmovdqul(XMMRegister dst, KRegister mask, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int16(0x6F, (0xC0 | encode));
 }
 
 void Assembler::evmovdqul(XMMRegister dst, Address src, int vector_len) {
+  // Unmasked instruction
+  evmovdqul(dst, k0, src, /*merge*/ false, vector_len);
+}
+
+void Assembler::evmovdqul(XMMRegister dst, KRegister mask, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true , /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false , /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(src, 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
 void Assembler::evmovdqul(Address dst, XMMRegister src, int vector_len) {
+  // Unmasked isntruction
+  evmovdqul(dst, k0, src, /*merge*/ true, vector_len);
+}
+
+void Assembler::evmovdqul(Address dst, KRegister mask, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   assert(src != xnoreg, "sanity");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
-  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(dst, 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int8(0x7F);
   emit_operand(src, dst);
 }
 
 void Assembler::evmovdquq(XMMRegister dst, XMMRegister src, int vector_len) {
+  // Unmasked instruction
+  if (dst->encoding() == src->encoding()) return;
+  evmovdquq(dst, k0, src, /*merge*/ false, vector_len);
+}
+
+void Assembler::evmovdquq(XMMRegister dst, KRegister mask, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
-  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int16(0x6F, (0xC0 | encode));
 }
 
 void Assembler::evmovdquq(XMMRegister dst, Address src, int vector_len) {
+  // Unmasked instruction
+  evmovdquq(dst, k0, src, /*merge*/ false, vector_len);
+}
+
+void Assembler::evmovdquq(XMMRegister dst, KRegister mask, Address src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   vex_prefix(src, 0, dst->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int8(0x6F);
   emit_operand(dst, src);
 }
 
 void Assembler::evmovdquq(Address dst, XMMRegister src, int vector_len) {
+  // Unmasked instruction
+  evmovdquq(dst, k0, src, /*merge*/ true, vector_len);
+}
+
+void Assembler::evmovdquq(Address dst, KRegister mask, XMMRegister src, bool merge, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   assert(src != xnoreg, "sanity");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
-  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
   attributes.set_is_evex_instruction();
   vex_prefix(dst, 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
   emit_int8(0x7F);
@@ -2775,6 +2956,29 @@
   emit_operand(src, dst);
 }
 
+void Assembler::movq(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  attributes.set_rex_vex_w_reverted();
+  int encode = simd_prefix_and_encode(src, xnoreg, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xD6, (0xC0 | encode));
+}
+
+void Assembler::movq(Register dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  // swap src/dst to get correct prefix
+  int encode = simd_prefix_and_encode(src, xnoreg, as_XMMRegister(dst->encoding()), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x7E, (0xC0 | encode));
+}
+
+void Assembler::movq(XMMRegister dst, Register src) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, xnoreg, as_XMMRegister(src->encoding()), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x6E, (0xC0 | encode));
+}
+
 void Assembler::movsbl(Register dst, Address src) { // movsxb
   InstructionMark im(this);
   prefix(src, dst);
@@ -3274,6 +3478,11 @@
   emit_int16((unsigned char)0xF7, (0xD0 | encode));
 }
 
+void Assembler::orw(Register dst, Register src) {
+  (void)prefix_and_encode(dst->encoding(), src->encoding());
+  emit_arith(0x0B, 0xC0, dst, src);
+}
+
 void Assembler::orl(Address dst, int32_t imm32) {
   InstructionMark im(this);
   prefix(dst);
@@ -3312,6 +3521,34 @@
   emit_int8(imm8);
 }
 
+void Assembler::packsswb(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x63, (0xC0 | encode));
+}
+
+void Assembler::vpacksswb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 0, "some form of AVX must be enabled");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x63, (0xC0 | encode));
+}
+
+void Assembler::packssdw(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse2(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x6B, (0xC0 | encode));
+}
+
+void Assembler::vpackssdw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 0, "some form of AVX must be enabled");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x6B, (0xC0 | encode));
+}
+
 void Assembler::packuswb(XMMRegister dst, Address src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   assert((UseAVX > 0), "SSE mode requires address alignment 16 bytes");
@@ -3337,21 +3574,74 @@
   emit_int16(0x67, (0xC0 | encode));
 }
 
+void Assembler::packusdw(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x2B, (0xC0 | encode));
+}
+
+void Assembler::vpackusdw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 0, "some form of AVX must be enabled");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x2B, (0xC0 | encode));
+}
+
 void Assembler::vpermq(XMMRegister dst, XMMRegister src, int imm8, int vector_len) {
   assert(VM_Version::supports_avx2(), "");
+  assert(vector_len != AVX_128bit, "");
+  // VEX.256.66.0F3A.W1 00 /r ib
   InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x00, (0xC0 | encode), imm8);
 }
 
 void Assembler::vpermq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
-  assert(UseAVX > 2, "requires AVX512F");
+  assert(vector_len == AVX_256bit ? VM_Version::supports_avx512vl() :
+         vector_len == AVX_512bit ? VM_Version::supports_evex()     : false, "not supported");
   InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_is_evex_instruction();
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
   emit_int16(0x36, (0xC0 | encode));
 }
 
+void Assembler::vpermb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_avx512_vbmi(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16((unsigned char)0x8D, (0xC0 | encode));
+}
+
+void Assembler::vpermw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx512vlbw() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx512vlbw() :
+         vector_len == AVX_512bit ? VM_Version::supports_avx512bw()   : false, "not supported");
+  InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16((unsigned char)0x8D, (0xC0 | encode));
+}
+
+void Assembler::vpermd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_evex(), "");
+  // VEX.NDS.256.66.0F38.W0 36 /r
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x36, (0xC0 | encode));
+}
+
+void Assembler::vpermd(XMMRegister dst, XMMRegister nds, Address src, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_evex(), "");
+  // VEX.NDS.256.66.0F38.W0 36 /r
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  vex_prefix(src, nds->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8(0x36);
+  emit_operand(dst, src);
+}
+
 void Assembler::vperm2i128(XMMRegister dst,  XMMRegister nds, XMMRegister src, int imm8) {
   assert(VM_Version::supports_avx2(), "");
   InstructionAttr attributes(AVX_256bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
@@ -3366,6 +3656,28 @@
   emit_int24(0x06, (0xC0 | encode), imm8);
 }
 
+void Assembler::vpermilps(XMMRegister dst, XMMRegister src, int imm8, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x04, (0xC0 | encode), imm8);
+}
+
+void Assembler::vpermilpd(XMMRegister dst, XMMRegister src, int imm8, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ VM_Version::supports_evex(),/* legacy_mode */ false,/* no_mask_reg */ true, /* uses_vl */ false);
+  attributes.set_rex_vex_w_reverted();
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x05, (0xC0 | encode), imm8);
+}
+
+void Assembler::vpermpd(XMMRegister dst, XMMRegister src, int imm8, int vector_len) {
+  assert(vector_len <= AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ true, /* legacy_mode */false, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x01, (0xC0 | encode), imm8);
+}
+
 void Assembler::evpermi2q(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -3374,7 +3686,6 @@
   emit_int16(0x76, (0xC0 | encode));
 }
 
-
 void Assembler::pause() {
   emit_int16((unsigned char)0xF3, (unsigned char)0x90);
 }
@@ -3408,9 +3719,18 @@
   emit_int16(0x74, (0xC0 | encode));
 }
 
+void Assembler::vpcmpCCbwd(XMMRegister dst, XMMRegister nds, XMMRegister src, int cond_encoding, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(cond_encoding, (0xC0 | encode));
+}
+
 // In this context, the dst vector contains the components that are equal, non equal components are zeroed in dst
 void Assembler::vpcmpeqb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
-  assert(VM_Version::supports_avx(), "");
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
   InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int16(0x74, (0xC0 | encode));
@@ -3497,7 +3817,7 @@
 void Assembler::evpcmpeqb(KRegister kdst, KRegister mask, XMMRegister nds, Address src, int vector_len) {
   assert(VM_Version::supports_avx512vlbw(), "");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_reg_mask */ false, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.reset_is_clear_context();
   attributes.set_embedded_opmask_register_specifier(mask);
@@ -3517,7 +3837,8 @@
 
 // In this context, the dst vector contains the components that are equal, non equal components are zeroed in dst
 void Assembler::vpcmpeqw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
-  assert(VM_Version::supports_avx(), "");
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
   InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int16(0x75, (0xC0 | encode));
@@ -3554,29 +3875,32 @@
 
 // In this context, the dst vector contains the components that are equal, non equal components are zeroed in dst
 void Assembler::vpcmpeqd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
-  assert(VM_Version::supports_avx(), "");
-  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int16(0x76, (0xC0 | encode));
 }
 
 // In this context, kdst is written the mask used to process the equal components
-void Assembler::evpcmpeqd(KRegister kdst, XMMRegister nds, XMMRegister src, int vector_len) {
+void Assembler::evpcmpeqd(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src, int vector_len) {
   assert(VM_Version::supports_evex(), "");
-  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_is_evex_instruction();
   attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
   int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int16(0x76, (0xC0 | encode));
 }
 
-void Assembler::evpcmpeqd(KRegister kdst, XMMRegister nds, Address src, int vector_len) {
+void Assembler::evpcmpeqd(KRegister kdst, KRegister mask, XMMRegister nds, Address src, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_FV, /* input_size_in_bits */ EVEX_32bit);
-  attributes.reset_is_clear_context();
   attributes.set_is_evex_instruction();
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
   int dst_enc = kdst->encoding();
   vex_prefix(src, nds->encoding(), dst_enc, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int8(0x76);
@@ -3591,6 +3915,13 @@
   emit_int16(0x29, (0xC0 | encode));
 }
 
+void Assembler::vpcmpCCq(XMMRegister dst, XMMRegister nds, XMMRegister src, int cond_encoding, int vector_len) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(cond_encoding, (0xC0 | encode));
+}
+
 // In this context, the dst vector contains the components that are equal, non equal components are zeroed in dst
 void Assembler::vpcmpeqq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
   assert(VM_Version::supports_avx(), "");
@@ -3623,6 +3954,31 @@
   emit_operand(as_Register(dst_enc), src);
 }
 
+void Assembler::evpmovd2m(KRegister kdst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2  && VM_Version::supports_avx512dq(), "");
+  assert(vector_len == AVX_512bit || VM_Version::supports_avx512vl(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(kdst->encoding(), 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x39, (0xC0 | encode));
+}
+
+void Assembler::evpmovq2m(KRegister kdst, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2  && VM_Version::supports_avx512dq(), "");
+  assert(vector_len == AVX_512bit || VM_Version::supports_avx512vl(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(kdst->encoding(), 0, src->encoding(), VEX_SIMD_F3, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x39, (0xC0 | encode));
+}
+
+void Assembler::pcmpgtq(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x37, (0xC0 | encode));
+}
+
 void Assembler::pmovmskb(Register dst, XMMRegister src) {
   assert(VM_Version::supports_sse2(), "");
   InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
@@ -3639,14 +3995,14 @@
 
 void Assembler::pextrd(Register dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(src, xnoreg, as_XMMRegister(dst->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x16, (0xC0 | encode), imm8);
 }
 
 void Assembler::pextrd(Address dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
   simd_prefix(src, xnoreg, dst, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x16);
@@ -3656,14 +4012,14 @@
 
 void Assembler::pextrq(Register dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(src, xnoreg, as_XMMRegister(dst->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x16, (0xC0 | encode), imm8);
 }
 
 void Assembler::pextrq(Address dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_64bit);
   simd_prefix(src, xnoreg, dst, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x16);
@@ -3673,14 +4029,14 @@
 
 void Assembler::pextrw(Register dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse2(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(as_XMMRegister(dst->encoding()), xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int24((unsigned char)0xC5, (0xC0 | encode), imm8);
 }
 
 void Assembler::pextrw(Address dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_16bit);
   simd_prefix(src, xnoreg, dst, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x15);
@@ -3688,9 +4044,16 @@
   emit_int8(imm8);
 }
 
+void Assembler::pextrb(Register dst, XMMRegister src, int imm8) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(src, xnoreg, as_XMMRegister(dst->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x14, (0xC0 | encode), imm8);
+}
+
 void Assembler::pextrb(Address dst, XMMRegister src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_8bit);
   simd_prefix(src, xnoreg, dst, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x14);
@@ -3700,14 +4063,14 @@
 
 void Assembler::pinsrd(XMMRegister dst, Register src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(dst, dst, as_XMMRegister(src->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x22, (0xC0 | encode), imm8);
 }
 
 void Assembler::pinsrd(XMMRegister dst, Address src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
   simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x22);
@@ -3715,16 +4078,23 @@
   emit_int8(imm8);
 }
 
+void Assembler::vpinsrd(XMMRegister dst, XMMRegister nds, Register src, int imm8) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x22, (0xC0 | encode), imm8);
+}
+
 void Assembler::pinsrq(XMMRegister dst, Register src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(dst, dst, as_XMMRegister(src->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x22, (0xC0 | encode), imm8);
 }
 
 void Assembler::pinsrq(XMMRegister dst, Address src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_64bit);
   simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x22);
@@ -3732,16 +4102,23 @@
   emit_int8(imm8);
 }
 
+void Assembler::vpinsrq(XMMRegister dst, XMMRegister nds, Register src, int imm8) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ true, /* legacy_mode */ _legacy_mode_dq, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x22, (0xC0 | encode), imm8);
+}
+
 void Assembler::pinsrw(XMMRegister dst, Register src, int imm8) {
   assert(VM_Version::supports_sse2(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = simd_prefix_and_encode(dst, dst, as_XMMRegister(src->encoding()), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int24((unsigned char)0xC4, (0xC0 | encode), imm8);
 }
 
 void Assembler::pinsrw(XMMRegister dst, Address src, int imm8) {
   assert(VM_Version::supports_sse2(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_16bit);
   simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
   emit_int8((unsigned char)0xC4);
@@ -3749,9 +4126,16 @@
   emit_int8(imm8);
 }
 
+void Assembler::vpinsrw(XMMRegister dst, XMMRegister nds, Register src, int imm8) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC4, (0xC0 | encode), imm8);
+}
+
 void Assembler::pinsrb(XMMRegister dst, Address src, int imm8) {
   assert(VM_Version::supports_sse4_1(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
   attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_8bit);
   simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int8(0x20);
@@ -3759,6 +4143,34 @@
   emit_int8(imm8);
 }
 
+void Assembler::pinsrb(XMMRegister dst, Register src, int imm8) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, dst, as_XMMRegister(src->encoding()), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x20, (0xC0 | encode), imm8);
+}
+
+void Assembler::vpinsrb(XMMRegister dst, XMMRegister nds, Register src, int imm8) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x20, (0xC0 | encode), imm8);
+}
+
+void Assembler::insertps(XMMRegister dst, XMMRegister src, int imm8) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x21, (0xC0 | encode), imm8);
+}
+
+void Assembler::vinsertps(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x21, (0xC0 | encode), imm8);
+}
+
 void Assembler::pmovzxbw(XMMRegister dst, Address src) {
   assert(VM_Version::supports_sse4_1(), "");
   InstructionMark im(this);
@@ -3783,6 +4195,41 @@
   emit_int16(0x20, (0xC0 | encode));
 }
 
+void Assembler::pmovzxdq(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x35, (0xC0 | encode));
+}
+
+void Assembler::pmovsxbd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x21, (0xC0 | encode));
+}
+
+void Assembler::pmovzxbd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x31, (0xC0 | encode));
+}
+
+void Assembler::pmovsxbq(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x22, (0xC0 | encode));
+}
+
+void Assembler::pmovsxwd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x23, (0xC0 | encode));
+}
+
 void Assembler::vpmovzxbw(XMMRegister dst, Address src, int vector_len) {
   assert(VM_Version::supports_avx(), "");
   InstructionMark im(this);
@@ -3816,7 +4263,7 @@
   assert(VM_Version::supports_avx512vlbw(), "");
   assert(dst != xnoreg, "sanity");
   InstructionMark im(this);
-  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
   attributes.set_address_attributes(/* tuple_type */ EVEX_HVM, /* input_size_in_bits */ EVEX_NObit);
   attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
@@ -3824,6 +4271,86 @@
   emit_int8(0x30);
   emit_operand(dst, src);
 }
+
+void Assembler::evpandd(XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F.W0 DB /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xDB, (0xC0 | encode));
+}
+
+void Assembler::vpmovzxdq(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len > AVX_128bit ? VM_Version::supports_avx2() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x35, (0xC0 | encode));
+}
+
+void Assembler::vpmovzxbd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len > AVX_128bit ? VM_Version::supports_avx2() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x31, (0xC0 | encode));
+}
+
+void Assembler::vpmovzxbq(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len > AVX_128bit ? VM_Version::supports_avx2() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x32, (0xC0 | encode));
+}
+
+void Assembler::vpmovsxbd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2() :
+             VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x21, (0xC0 | encode));
+}
+
+void Assembler::vpmovsxbq(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2() :
+             VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x22, (0xC0 | encode));
+}
+
+void Assembler::vpmovsxwd(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2() :
+             VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x23, (0xC0 | encode));
+}
+
+void Assembler::vpmovsxwq(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2() :
+             VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x24, (0xC0 | encode));
+}
+
+void Assembler::vpmovsxdq(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+         vector_len == AVX_256bit ? VM_Version::supports_avx2() :
+             VM_Version::supports_evex(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x25, (0xC0 | encode));
+}
+
 void Assembler::evpmovwb(Address dst, XMMRegister src, int vector_len) {
   assert(VM_Version::supports_avx512vlbw(), "");
   assert(src != xnoreg, "sanity");
@@ -4050,6 +4577,14 @@
   emit_int8(mode & 0xFF);
 }
 
+void Assembler::pshufhw(XMMRegister dst, XMMRegister src, int mode) {
+  assert(isByte(mode), "invalid value");
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_F3, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x70, (0xC0 | encode), mode & 0xFF);
+}
+
 void Assembler::pshuflw(XMMRegister dst, XMMRegister src, int mode) {
   assert(isByte(mode), "invalid value");
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
@@ -4080,6 +4615,35 @@
   emit_int24(0x43, (0xC0 | encode), imm8 & 0xFF);
 }
 
+void Assembler::pshufpd(XMMRegister dst, XMMRegister src, int imm8) {
+  assert(isByte(imm8), "invalid value");
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC6, (0xC0 | encode), imm8 & 0xFF);
+}
+
+void Assembler::vpshufpd(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8, int vector_len) {
+  InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_rex_vex_w_reverted();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC6, (0xC0 | encode), imm8 & 0xFF);
+}
+
+void Assembler::pshufps(XMMRegister dst, XMMRegister src, int imm8) {
+  assert(isByte(imm8), "invalid value");
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC6, (0xC0 | encode), imm8 & 0xFF);
+}
+
+void Assembler::vpshufps(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8, int vector_len) {
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC6, (0xC0 | encode), imm8 & 0xFF);
+}
+
 void Assembler::psrldq(XMMRegister dst, int shift) {
   // Shift left 128 bit value in dst XMMRegister by shift number of bytes.
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
@@ -4151,6 +4715,13 @@
   emit_int16(0x17, (0xC0 | encode));
 }
 
+void Assembler::vptest(XMMRegister dst, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x17, (0xC0 | encode));
+}
+
 void Assembler::punpcklbw(XMMRegister dst, Address src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   assert((UseAVX > 0), "SSE mode requires address alignment 16 bytes");
@@ -4819,6 +5390,11 @@
   emit_operand(dst, src);
 }
 
+void Assembler::xorw(Register dst, Register src) {
+  (void)prefix_and_encode(dst->encoding(), src->encoding());
+  emit_arith(0x33, 0xC0, dst, src);
+}
+
 // AVX 3-operands scalar float-point arithmetic instructions
 
 void Assembler::vaddsd(XMMRegister dst, XMMRegister nds, Address src) {
@@ -5732,6 +6308,13 @@
   emit_int16(0x40, (0xC0 | encode));
 }
 
+void Assembler::pmuludq(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse2(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF4, (0xC0 | encode));
+}
+
 void Assembler::vpmullw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
   assert(UseAVX > 0, "requires some form of AVX");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -5754,6 +6337,13 @@
   emit_int16(0x40, (0xC0 | encode));
 }
 
+void Assembler::vpmuludq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF4, (0xC0 | encode));
+}
+
 void Assembler::vpmullw(XMMRegister dst, XMMRegister nds, Address src, int vector_len) {
   assert(UseAVX > 0, "requires some form of AVX");
   InstructionMark im(this);
@@ -5785,95 +6375,256 @@
   emit_operand(dst, src);
 }
 
-// Shift packed integers left by specified number of bits.
-void Assembler::psllw(XMMRegister dst, int shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+// Min, max
+void Assembler::pminsb(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
   InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
-  // XMM6 is for /6 encoding: 66 0F 71 /6 ib
-  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x71, (0xC0 | encode), shift & 0xFF);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x38, (0xC0 | encode));
 }
 
-void Assembler::pslld(XMMRegister dst, int shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  // XMM6 is for /6 encoding: 66 0F 72 /6 ib
-  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x72, (0xC0 | encode), shift & 0xFF);
+void Assembler::vpminsb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_avx512bw()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x38, (0xC0 | encode));
 }
 
-void Assembler::psllq(XMMRegister dst, int shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  // XMM6 is for /6 encoding: 66 0F 73 /6 ib
-  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x73, (0xC0 | encode), shift & 0xFF);
+void Assembler::pminsw(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse2(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEA, (0xC0 | encode));
 }
 
-void Assembler::psllw(XMMRegister dst, XMMRegister shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int16((unsigned char)0xF1, (0xC0 | encode));
+void Assembler::vpminsw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_avx512bw()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEA, (0xC0 | encode));
 }
 
-void Assembler::pslld(XMMRegister dst, XMMRegister shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+void Assembler::pminsd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
   InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int16((unsigned char)0xF2, (0xC0 | encode));
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x39, (0xC0 | encode));
 }
 
-void Assembler::psllq(XMMRegister dst, XMMRegister shift) {
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  attributes.set_rex_vex_w_reverted();
-  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int16((unsigned char)0xF3, (0xC0 | encode));
+void Assembler::vpminsd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_evex()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x39, (0xC0 | encode));
 }
 
-void Assembler::vpsllw(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
-  // XMM6 is for /6 encoding: 66 0F 71 /6 ib
-  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x71, (0xC0 | encode), shift & 0xFF);
+void Assembler::vpminsq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "requires AVX512F");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x39, (0xC0 | encode));
 }
 
-void Assembler::vpslld(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
-  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+void Assembler::minps(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5D, (0xC0 | encode));
+}
+void Assembler::vminps(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len >= AVX_512bit ? VM_Version::supports_evex() : VM_Version::supports_avx(), "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  // XMM6 is for /6 encoding: 66 0F 72 /6 ib
-  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x72, (0xC0 | encode), shift & 0xFF);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5D, (0xC0 | encode));
 }
 
-void Assembler::vpsllq(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
-  InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  attributes.set_rex_vex_w_reverted();
-  // XMM6 is for /6 encoding: 66 0F 73 /6 ib
-  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int24(0x73, (0xC0 | encode), shift & 0xFF);
+void Assembler::minpd(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5D, (0xC0 | encode));
 }
-
-void Assembler::vpsllw(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int16((unsigned char)0xF1, (0xC0 | encode));
+void Assembler::vminpd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len >= AVX_512bit ? VM_Version::supports_evex() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5D, (0xC0 | encode));
 }
 
-void Assembler::vpslld(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int16((unsigned char)0xF2, (0xC0 | encode));
+void Assembler::pmaxsb(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x3C, (0xC0 | encode));
 }
 
-void Assembler::vpsllq(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
-  assert(UseAVX > 0, "requires some form of AVX");
+void Assembler::vpmaxsb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_avx512bw()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x3C, (0xC0 | encode));
+}
+
+void Assembler::pmaxsw(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse2(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEE, (0xC0 | encode));
+}
+
+void Assembler::vpmaxsw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_avx512bw()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEE, (0xC0 | encode));
+}
+
+void Assembler::pmaxsd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x3D, (0xC0 | encode));
+}
+
+void Assembler::vpmaxsd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() :
+        (vector_len == AVX_256bit ? VM_Version::supports_avx2() : VM_Version::supports_evex()), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x3D, (0xC0 | encode));
+}
+
+void Assembler::vpmaxsq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "requires AVX512F");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x3D, (0xC0 | encode));
+}
+
+void Assembler::maxps(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5F, (0xC0 | encode));
+}
+
+void Assembler::vmaxps(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len >= AVX_512bit ? VM_Version::supports_evex() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5F, (0xC0 | encode));
+}
+
+void Assembler::maxpd(XMMRegister dst, XMMRegister src) {
+  NOT_LP64(assert(VM_Version::supports_sse(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5F, (0xC0 | encode));
+}
+
+void Assembler::vmaxpd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len >= AVX_512bit ? VM_Version::supports_evex() : VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x5F, (0xC0 | encode));
+}
+
+// Shift packed integers left by specified number of bits.
+void Assembler::psllw(XMMRegister dst, int shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  // XMM6 is for /6 encoding: 66 0F 71 /6 ib
+  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x71, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::pslld(XMMRegister dst, int shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  // XMM6 is for /6 encoding: 66 0F 72 /6 ib
+  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x72, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::psllq(XMMRegister dst, int shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  // XMM6 is for /6 encoding: 66 0F 73 /6 ib
+  int encode = simd_prefix_and_encode(xmm6, dst, dst, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x73, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::psllw(XMMRegister dst, XMMRegister shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF1, (0xC0 | encode));
+}
+
+void Assembler::pslld(XMMRegister dst, XMMRegister shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF2, (0xC0 | encode));
+}
+
+void Assembler::psllq(XMMRegister dst, XMMRegister shift) {
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_rex_vex_w_reverted();
+  int encode = simd_prefix_and_encode(dst, dst, shift, VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF3, (0xC0 | encode));
+}
+
+void Assembler::vpsllw(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  // XMM6 is for /6 encoding: 66 0F 71 /6 ib
+  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x71, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::vpslld(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  NOT_LP64(assert(VM_Version::supports_sse2(), ""));
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  // XMM6 is for /6 encoding: 66 0F 72 /6 ib
+  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x72, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::vpsllq(XMMRegister dst, XMMRegister src, int shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_rex_vex_w_reverted();
+  // XMM6 is for /6 encoding: 66 0F 73 /6 ib
+  int encode = vex_prefix_and_encode(xmm6->encoding(), dst->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24(0x73, (0xC0 | encode), shift & 0xFF);
+}
+
+void Assembler::vpsllw(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF1, (0xC0 | encode));
+}
+
+void Assembler::vpslld(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xF2, (0xC0 | encode));
+}
+
+void Assembler::vpsllq(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 0, "requires some form of AVX");
   InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_rex_vex_w_reverted();
   int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
@@ -6106,13 +6857,67 @@
   emit_int16((unsigned char)0xDB, (0xC0 | encode));
 }
 
+//Variable Shift packed integers logically left.
+void Assembler::vpsllvd(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 1, "requires AVX2");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x47, (0xC0 | encode));
+}
+
+void Assembler::vpsllvq(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 1, "requires AVX2");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x47, (0xC0 | encode));
+}
+
+//Variable Shift packed integers logically right.
+void Assembler::vpsrlvd(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 1, "requires AVX2");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x45, (0xC0 | encode));
+}
+
+void Assembler::vpsrlvq(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 1, "requires AVX2");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x45, (0xC0 | encode));
+}
+
+//Variable right Shift arithmetic packed integers .
+void Assembler::vpsravd(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 1, "requires AVX2");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x46, (0xC0 | encode));
+}
+
+void Assembler::evpsravw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_avx512bw(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x11, (0xC0 | encode));
+}
+
+void Assembler::evpsravq(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
+  assert(UseAVX > 2, "requires AVX512");
+  assert(vector_len == Assembler::AVX_512bit || VM_Version::supports_avx512vl(), "requires AVX512VL");
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x46, (0xC0 | encode));
+}
+
 void Assembler::vpshldvd(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
   assert(VM_Version::supports_avx512_vbmi2(), "requires vbmi2");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_is_evex_instruction();
   int encode = vex_prefix_and_encode(dst->encoding(), src->encoding(), shift->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
-  emit_int8(0x71);
-  emit_int8((0xC0 | encode));
+  emit_int16(0x71, (0xC0 | encode));
 }
 
 void Assembler::vpshrdvd(XMMRegister dst, XMMRegister src, XMMRegister shift, int vector_len) {
@@ -6138,7 +6943,6 @@
   emit_int16((unsigned char)0xDF, (0xC0 | encode));
 }
 
-
 void Assembler::por(XMMRegister dst, XMMRegister src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -6171,6 +6975,35 @@
 }
 
 
+void Assembler::evpord(XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F.W0 EB /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEB, (0xC0 | encode));
+}
+
+void Assembler::evpord(XMMRegister dst, KRegister mask, XMMRegister nds, Address src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F.W0 EB /r
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_FV, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  vex_prefix(src, nds->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int8((unsigned char)0xEB);
+  emit_operand(dst, src);
+}
+
 void Assembler::pxor(XMMRegister dst, XMMRegister src) {
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
   InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -6195,13 +7028,33 @@
   emit_operand(dst, src);
 }
 
+void Assembler::vpxorq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(UseAVX > 2, "requires some form of EVEX");
+  InstructionAttr attributes(vector_len, /* vex_w */ VM_Version::supports_evex(), /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
+  attributes.set_rex_vex_w_reverted();
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEF, (0xC0 | encode));
+}
+
+void Assembler::evpxord(XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F.W0 EF /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16((unsigned char)0xEF, (0xC0 | encode));
+}
+
 void Assembler::evpxorq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
   assert(VM_Version::supports_evex(), "requires EVEX support");
   InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ true, /* uses_vl */ true);
   attributes.set_is_evex_instruction();
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
-  emit_int8((unsigned char)0xEF);
-  emit_int8((0xC0 | encode));
+  emit_int16((unsigned char)0xEF, (0xC0 | encode));
 }
 
 void Assembler::evpxorq(XMMRegister dst, XMMRegister nds, Address src, int vector_len) {
@@ -6216,7 +7069,6 @@
   emit_operand(dst, src);
 }
 
-
 // vinserti forms
 
 void Assembler::vinserti128(XMMRegister dst, XMMRegister nds, XMMRegister src, uint8_t imm8) {
@@ -6776,12 +7628,67 @@
   int encode = vex_prefix_and_encode(dst->encoding(), 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
   emit_int16(0x7C, (0xC0 | encode));
 }
+
+void Assembler::vpgatherdd(XMMRegister dst, Address src, XMMRegister mask, int vector_len) {
+  assert(VM_Version::supports_avx2(), "");
+  assert(vector_len == Assembler::AVX_128bit || vector_len == Assembler::AVX_256bit, "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  vex_prefix(src, mask->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x90);
+  emit_operand(dst, src);
+}
+
+void Assembler::vpgatherdq(XMMRegister dst, Address src, XMMRegister mask, int vector_len) {
+  assert(VM_Version::supports_avx2(), "");
+  assert(vector_len == Assembler::AVX_128bit || vector_len == Assembler::AVX_256bit, "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  vex_prefix(src, mask->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x90);
+  emit_operand(dst, src);
+}
+
+void Assembler::vgatherdpd(XMMRegister dst, Address src, XMMRegister mask, int vector_len) {
+  assert(VM_Version::supports_avx2(), "");
+  assert(vector_len == Assembler::AVX_128bit || vector_len == Assembler::AVX_256bit, "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  vex_prefix(src, mask->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x92);
+  emit_operand(dst, src);
+}
+
+void Assembler::vgatherdps(XMMRegister dst, Address src, XMMRegister mask, int vector_len) {
+  assert(VM_Version::supports_avx2(), "");
+  assert(vector_len == Assembler::AVX_128bit || vector_len == Assembler::AVX_256bit, "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ false, /* uses_vl */ true);
+  vex_prefix(src, mask->encoding(), dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x92);
+  emit_operand(dst, src);
+}
 void Assembler::evpgatherdd(XMMRegister dst, KRegister mask, Address src, int vector_len) {
   assert(VM_Version::supports_evex(), "");
   assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
   InstructionMark im(this);
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
-  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_64bit);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
   attributes.reset_is_clear_context();
   attributes.set_embedded_opmask_register_specifier(mask);
   attributes.set_is_evex_instruction();
@@ -6790,19 +7697,129 @@
   emit_int8((unsigned char)0x90);
   emit_operand(dst, src);
 }
-// Carry-Less Multiplication Quadword
-void Assembler::pclmulqdq(XMMRegister dst, XMMRegister src, int mask) {
-  assert(VM_Version::supports_clmul(), "");
-  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
-  emit_int24(0x44, (0xC0 | encode), (unsigned char)mask);
-}
 
-// Carry-Less Multiplication Quadword
-void Assembler::vpclmulqdq(XMMRegister dst, XMMRegister nds, XMMRegister src, int mask) {
-  assert(VM_Version::supports_avx() && VM_Version::supports_clmul(), "");
-  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+void Assembler::evpgatherdq(XMMRegister dst, KRegister mask, Address src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  // swap src<->dst for encoding
+  vex_prefix(src, 0, dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x90);
+  emit_operand(dst, src);
+}
+
+void Assembler::evgatherdpd(XMMRegister dst, KRegister mask, Address src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  // swap src<->dst for encoding
+  vex_prefix(src, 0, dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x92);
+  emit_operand(dst, src);
+}
+
+void Assembler::evgatherdps(XMMRegister dst, KRegister mask, Address src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(dst != xnoreg, "sanity");
+  assert(src.isxmmindex(),"expected to be xmm index");
+  assert(dst != src.xmmindex(), "instruction will #UD if dst and index are the same");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  // swap src<->dst for encoding
+  vex_prefix(src, 0, dst->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0x92);
+  emit_operand(dst, src);
+}
+
+void Assembler::evpscatterdd(Address dst, KRegister mask, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  vex_prefix(dst, 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0xA0);
+  emit_operand(src, dst);
+}
+
+void Assembler::evpscatterdq(Address dst, KRegister mask, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  vex_prefix(dst, 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0xA0);
+  emit_operand(src, dst);
+}
+
+void Assembler::evscatterdps(Address dst, KRegister mask, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  vex_prefix(dst, 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0xA2);
+  emit_operand(src, dst);
+}
+
+void Assembler::evscatterdpd(Address dst, KRegister mask, XMMRegister src, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(mask != k0, "instruction will #UD if mask is in k0");
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_T1S, /* input_size_in_bits */ EVEX_32bit);
+  attributes.reset_is_clear_context();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.set_is_evex_instruction();
+  vex_prefix(dst, 0, src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int8((unsigned char)0xA2);
+  emit_operand(src, dst);
+}
+// Carry-Less Multiplication Quadword
+void Assembler::pclmulqdq(XMMRegister dst, XMMRegister src, int mask) {
+  assert(VM_Version::supports_clmul(), "");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x44, (0xC0 | encode), (unsigned char)mask);
+}
+
+// Carry-Less Multiplication Quadword
+void Assembler::vpclmulqdq(XMMRegister dst, XMMRegister nds, XMMRegister src, int mask) {
+  assert(VM_Version::supports_avx() && VM_Version::supports_clmul(), "");
+  InstructionAttr attributes(AVX_128bit, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   emit_int24(0x44, (0xC0 | encode), (unsigned char)mask);
 }
 
@@ -7387,7 +8404,8 @@
   // fourth EVEX.L'L for vector length : 0 is 128, 1 is 256, 2 is 512, currently we do not support 1024
   byte4 |= ((_attributes->get_vector_len())& 0x3) << 5;
   // last is EVEX.z for zero/merge actions
-  if (_attributes->is_no_reg_mask() == false) {
+  if (_attributes->is_no_reg_mask() == false &&
+      _attributes->get_embedded_opmask_register_specifier() != 0) {
     byte4 |= (_attributes->is_clear_context() ? EVEX_Z : 0);
   }
 
@@ -7555,7 +8573,7 @@
   emit_int16(0x5D, (0xC0 | encode));
 }
 
-void Assembler::cmppd(XMMRegister dst, XMMRegister nds, XMMRegister src, int cop, int vector_len) {
+void Assembler::vcmppd(XMMRegister dst, XMMRegister nds, XMMRegister src, int cop, int vector_len) {
   assert(VM_Version::supports_avx(), "");
   assert(vector_len <= AVX_256bit, "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
@@ -7563,8 +8581,8 @@
   emit_int24((unsigned char)0xC2, (0xC0 | encode), (0xF & cop));
 }
 
-void Assembler::blendvpd(XMMRegister dst, XMMRegister nds, XMMRegister src1, XMMRegister src2, int vector_len) {
-  assert(VM_Version::supports_avx(), "");
+void Assembler::vblendvpd(XMMRegister dst, XMMRegister nds, XMMRegister src1, XMMRegister src2, int vector_len) {
+  assert(UseAVX > 0 && (vector_len == AVX_128bit || vector_len == AVX_256bit), "");
   assert(vector_len <= AVX_256bit, "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src1->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
@@ -7572,28 +8590,330 @@
   emit_int24(0x4B, (0xC0 | encode), (0xF0 & src2_enc << 4));
 }
 
-void Assembler::cmpps(XMMRegister dst, XMMRegister nds, XMMRegister src, int cop, int vector_len) {
-  assert(VM_Version::supports_avx(), "");
+void Assembler::vpblendd(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8, int vector_len) {
+  assert(VM_Version::supports_avx2(), "");
   assert(vector_len <= AVX_256bit, "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
-  int encode = simd_prefix_and_encode(dst, nds, src, VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
-  emit_int24((unsigned char)0xC2, (0xC0 | encode), (0xF & cop));
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x02, (0xC0 | encode), (unsigned char)imm8);
 }
 
-void Assembler::blendvps(XMMRegister dst, XMMRegister nds, XMMRegister src1, XMMRegister src2, int vector_len) {
+void Assembler::vcmpps(XMMRegister dst, XMMRegister nds, XMMRegister src, int comparison, int vector_len) {
   assert(VM_Version::supports_avx(), "");
   assert(vector_len <= AVX_256bit, "");
   InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC2, (0xC0 | encode), (unsigned char)comparison);
+}
+
+void Assembler::evcmpps(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        ComparisonPredicateFP comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.0F.W0 C2 /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_NONE, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC2, (0xC0 | encode), comparison);
+}
+
+void Assembler::evcmppd(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        ComparisonPredicateFP comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F.W1 C2 /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int24((unsigned char)0xC2, (0xC0 | encode), comparison);
+}
+
+void Assembler::blendvps(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  assert(UseAVX <= 0, "sse encoding is inconsistent with avx encoding");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x14, (0xC0 | encode));
+}
+
+void Assembler::blendvpd(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  assert(UseAVX <= 0, "sse encoding is inconsistent with avx encoding");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x15, (0xC0 | encode));
+}
+
+void Assembler::pblendvb(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_sse4_1(), "");
+  assert(UseAVX <= 0, "sse encoding is inconsistent with avx encoding");
+  InstructionAttr attributes(AVX_128bit, /* rex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = simd_prefix_and_encode(dst, xnoreg, src, VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x10, (0xC0 | encode));
+}
+
+void Assembler::vblendvps(XMMRegister dst, XMMRegister nds, XMMRegister src1, XMMRegister src2, int vector_len) {
+  assert(UseAVX > 0 && (vector_len == AVX_128bit || vector_len == AVX_256bit), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src1->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
   int src2_enc = src2->encoding();
   emit_int24(0x4A, (0xC0 | encode), (0xF0 & src2_enc << 4));
 }
 
-void Assembler::vpblendd(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8, int vector_len) {
-  assert(VM_Version::supports_avx2(), "");
-  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ true);
+void Assembler::vblendps(XMMRegister dst, XMMRegister nds, XMMRegister src, int imm8, int vector_len) {
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
   int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
-  emit_int24(0x02, (0xC0 | encode), (unsigned char)imm8);
+  emit_int24(0x0C, (0xC0 | encode), imm8);
+}
+
+void Assembler::vpcmpgtb(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x64, (0xC0 | encode));
+}
+
+void Assembler::vpcmpgtw(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x65, (0xC0 | encode));
+}
+
+void Assembler::vpcmpgtd(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F, &attributes);
+  emit_int16(0x66, (0xC0 | encode));
+}
+
+void Assembler::vpcmpgtq(XMMRegister dst, XMMRegister nds, XMMRegister src, int vector_len) {
+  assert(vector_len == AVX_128bit ? VM_Version::supports_avx() : VM_Version::supports_avx2(), "");
+  assert(vector_len <= AVX_256bit, "evex encoding is different - has k register as dest");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x37, (0xC0 | encode));
+}
+
+void Assembler::evpcmpd(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W0 1F /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x1F, (0xC0 | encode), comparison);
+}
+
+void Assembler::evpcmpd(KRegister kdst, KRegister mask, XMMRegister nds, Address src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W0 1F /r ib
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_FV, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int dst_enc = kdst->encoding();
+  vex_prefix(src, nds->encoding(), dst_enc, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int8((unsigned char)0x1F);
+  emit_operand(as_Register(dst_enc), src);
+  emit_int8((unsigned char)comparison);
+}
+
+void Assembler::evpcmpq(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W1 1F /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x1F, (0xC0 | encode), comparison);
+}
+
+void Assembler::evpcmpq(KRegister kdst, KRegister mask, XMMRegister nds, Address src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W1 1F /r ib
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_FV, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int dst_enc = kdst->encoding();
+  vex_prefix(src, nds->encoding(), dst_enc, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int8((unsigned char)0x1F);
+  emit_operand(as_Register(dst_enc), src);
+  emit_int8((unsigned char)comparison);
+}
+
+void Assembler::evpcmpb(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W0 3F /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x3F, (0xC0 | encode), comparison);
+}
+
+void Assembler::evpcmpb(KRegister kdst, KRegister mask, XMMRegister nds, Address src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W0 3F /r ib
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int dst_enc = kdst->encoding();
+  vex_prefix(src, nds->encoding(), dst_enc, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int8((unsigned char)0x3F);
+  emit_operand(as_Register(dst_enc), src);
+  emit_int8((unsigned char)comparison);
+}
+
+void Assembler::evpcmpw(KRegister kdst, KRegister mask, XMMRegister nds, XMMRegister src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W1 3F /r ib
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int encode = vex_prefix_and_encode(kdst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int24(0x3F, (0xC0 | encode), comparison);
+}
+
+void Assembler::evpcmpw(KRegister kdst, KRegister mask, XMMRegister nds, Address src,
+                        int comparison, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  assert(comparison >= Assembler::eq && comparison <= Assembler::_true, "");
+  // Encoding: EVEX.NDS.XXX.66.0F3A.W1 3F /r ib
+  InstructionMark im(this);
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_address_attributes(/* tuple_type */ EVEX_FVM, /* input_size_in_bits */ EVEX_NObit);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  attributes.reset_is_clear_context();
+  int dst_enc = kdst->encoding();
+  vex_prefix(src, nds->encoding(), dst_enc, VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  emit_int8((unsigned char)0x3F);
+  emit_operand(as_Register(dst_enc), src);
+  emit_int8((unsigned char)comparison);
+}
+
+void Assembler::vpblendvb(XMMRegister dst, XMMRegister nds, XMMRegister src, XMMRegister mask, int vector_len) {
+  assert(VM_Version::supports_avx(), "");
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ true, /* no_mask_reg */ true, /* uses_vl */ false);
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_3A, &attributes);
+  int mask_enc = mask->encoding();
+  emit_int24(0x4C, (0xC0 | encode), 0xF0 & mask_enc << 4);
+}
+
+void Assembler::evblendmpd(XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F38.W1 65 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x65, (0xC0 | encode));
+}
+
+void Assembler::evblendmps(XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  // Encoding: EVEX.NDS.XXX.66.0F38.W0 65 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x65, (0xC0 | encode));
+}
+
+void Assembler::evpblendmb (XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  // Encoding: EVEX.NDS.512.66.0F38.W0 66 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x66, (0xC0 | encode));
+}
+
+void Assembler::evpblendmw (XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  assert(VM_Version::supports_avx512bw(), "");
+  // Encoding: EVEX.NDS.512.66.0F38.W1 66 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ _legacy_mode_bw, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x66, (0xC0 | encode));
+}
+
+void Assembler::evpblendmd (XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  //Encoding: EVEX.NDS.512.66.0F38.W0 64 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ false, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x64, (0xC0 | encode));
+}
+
+void Assembler::evpblendmq (XMMRegister dst, KRegister mask, XMMRegister nds, XMMRegister src, bool merge, int vector_len) {
+  assert(VM_Version::supports_evex(), "");
+  //Encoding: EVEX.NDS.512.66.0F38.W1 64 /r
+  InstructionAttr attributes(vector_len, /* vex_w */ true, /* legacy_mode */ false, /* no_mask_reg */ false, /* uses_vl */ true);
+  attributes.set_is_evex_instruction();
+  attributes.set_embedded_opmask_register_specifier(mask);
+  if (merge) {
+    attributes.reset_is_clear_context();
+  }
+  int encode = vex_prefix_and_encode(dst->encoding(), nds->encoding(), src->encoding(), VEX_SIMD_66, VEX_OPCODE_0F_38, &attributes);
+  emit_int16(0x64, (0xC0 | encode));
 }
 
 void Assembler::shlxl(Register dst, Register src1, Register src2) {