X86InstrMMX.td 29.6 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
//===-- X86InstrMMX.td - Describe the MMX Instruction Set --*- tablegen -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This file describes the X86 MMX instruction set, defining the instructions,
// and properties of the instructions which are needed for code generation,
// machine code emission, and analysis.
//
// All instructions that use MMX should be in this file, even if they also use
// SSE.
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// MMX Multiclasses
//===----------------------------------------------------------------------===//

// Alias instruction that maps zero vector to pxor mmx.
// This is expanded by ExpandPostRAPseudos to an pxor.
// We set canFoldAsLoad because this can be converted to a constant-pool
// load of an all-zeros value if folding it would be beneficial.
let isReMaterializable = 1, isAsCheapAsAMove = 1, canFoldAsLoad = 1,
    isPseudo = 1, SchedRW = [WriteZero], Predicates = [HasMMX] in {
def MMX_SET0 : I<0, Pseudo, (outs VR64:$dst), (ins), "",
                 [(set VR64:$dst, (x86mmx (MMX_X86movw2d (i32 0))))]>;
}

let Constraints = "$src1 = $dst" in {
  // MMXI_binop_rm_int - Simple MMX binary operator based on intrinsic.
  multiclass MMXI_binop_rm_int<bits<8> opc, string OpcodeStr, Intrinsic IntId,
                               X86FoldableSchedWrite sched, bit Commutable = 0,
                               X86MemOperand OType = i64mem> {
    def irr : MMXI<opc, MRMSrcReg, (outs VR64:$dst),
                 (ins VR64:$src1, VR64:$src2),
                 !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
                 [(set VR64:$dst, (IntId VR64:$src1, VR64:$src2))]>,
              Sched<[sched]> {
      let isCommutable = Commutable;
    }
    def irm : MMXI<opc, MRMSrcMem, (outs VR64:$dst),
                 (ins VR64:$src1, OType:$src2),
                 !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
                 [(set VR64:$dst, (IntId VR64:$src1, (load_mmx addr:$src2)))]>,
                 Sched<[sched.Folded, sched.ReadAfterFold]>;
  }

  multiclass MMXI_binop_rmi_int<bits<8> opc, bits<8> opc2, Format ImmForm,
                                string OpcodeStr, Intrinsic IntId,
                                Intrinsic IntId2, X86FoldableSchedWrite sched,
                                X86FoldableSchedWrite schedImm> {
    def rr : MMXI<opc, MRMSrcReg, (outs VR64:$dst),
                                  (ins VR64:$src1, VR64:$src2),
                  !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
                  [(set VR64:$dst, (IntId VR64:$src1, VR64:$src2))]>,
             Sched<[sched]>;
    def rm : MMXI<opc, MRMSrcMem, (outs VR64:$dst),
                                  (ins VR64:$src1, i64mem:$src2),
                  !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
                  [(set VR64:$dst, (IntId VR64:$src1, (load_mmx addr:$src2)))]>,
                  Sched<[sched.Folded, sched.ReadAfterFold]>;
    def ri : MMXIi8<opc2, ImmForm, (outs VR64:$dst),
                                   (ins VR64:$src1, i32u8imm:$src2),
                    !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
           [(set VR64:$dst, (IntId2 VR64:$src1, timm:$src2))]>,
           Sched<[schedImm]>;
  }
}

/// Unary MMX instructions requiring SSSE3.
multiclass SS3I_unop_rm_int_mm<bits<8> opc, string OpcodeStr,
                               Intrinsic IntId64, X86FoldableSchedWrite sched> {
  def rr : MMXSS38I<opc, MRMSrcReg, (outs VR64:$dst), (ins VR64:$src),
                 !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
                 [(set VR64:$dst, (IntId64 VR64:$src))]>,
           Sched<[sched]>;

  def rm : MMXSS38I<opc, MRMSrcMem, (outs VR64:$dst), (ins i64mem:$src),
                 !strconcat(OpcodeStr, "\t{$src, $dst|$dst, $src}"),
                 [(set VR64:$dst, (IntId64 (load_mmx addr:$src)))]>,
                 Sched<[sched.Folded]>;
}

/// Binary MMX instructions requiring SSSE3.
let ImmT = NoImm, Constraints = "$src1 = $dst" in {
multiclass SS3I_binop_rm_int_mm<bits<8> opc, string OpcodeStr,
                             Intrinsic IntId64, X86FoldableSchedWrite sched,
                             bit Commutable = 0> {
  let isCommutable = Commutable in
  def rr : MMXSS38I<opc, MRMSrcReg, (outs VR64:$dst),
       (ins VR64:$src1, VR64:$src2),
        !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
       [(set VR64:$dst, (IntId64 VR64:$src1, VR64:$src2))]>,
      Sched<[sched]>;
  def rm : MMXSS38I<opc, MRMSrcMem, (outs VR64:$dst),
       (ins VR64:$src1, i64mem:$src2),
        !strconcat(OpcodeStr, "\t{$src2, $dst|$dst, $src2}"),
       [(set VR64:$dst,
         (IntId64 VR64:$src1, (load_mmx addr:$src2)))]>,
      Sched<[sched.Folded, sched.ReadAfterFold]>;
}
}

/// PALIGN MMX instructions (require SSSE3).
multiclass ssse3_palign_mm<string asm, Intrinsic IntId,
                           X86FoldableSchedWrite sched> {
  def rri  : MMXSS3AI<0x0F, MRMSrcReg, (outs VR64:$dst),
      (ins VR64:$src1, VR64:$src2, u8imm:$src3),
      !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
      [(set VR64:$dst, (IntId VR64:$src1, VR64:$src2, (i8 timm:$src3)))]>,
      Sched<[sched]>;
  def rmi  : MMXSS3AI<0x0F, MRMSrcMem, (outs VR64:$dst),
      (ins VR64:$src1, i64mem:$src2, u8imm:$src3),
      !strconcat(asm, "\t{$src3, $src2, $dst|$dst, $src2, $src3}"),
      [(set VR64:$dst, (IntId VR64:$src1, (load_mmx addr:$src2),
                                          (i8 timm:$src3)))]>,
      Sched<[sched.Folded, sched.ReadAfterFold]>;
}

multiclass sse12_cvt_pint<bits<8> opc, RegisterClass SrcRC, RegisterClass DstRC,
                         Intrinsic Int, X86MemOperand x86memop, PatFrag ld_frag,
                         string asm, X86FoldableSchedWrite sched, Domain d> {
  def irr : MMXPI<opc, MRMSrcReg, (outs DstRC:$dst), (ins SrcRC:$src), asm,
                  [(set DstRC:$dst, (Int SrcRC:$src))], d>,
            Sched<[sched]>;
  def irm : MMXPI<opc, MRMSrcMem, (outs DstRC:$dst), (ins x86memop:$src), asm,
                  [(set DstRC:$dst, (Int (ld_frag addr:$src)))], d>,
            Sched<[sched.Folded]>;
}

multiclass sse12_cvt_pint_3addr<bits<8> opc, RegisterClass SrcRC,
                    RegisterClass DstRC, Intrinsic Int, X86MemOperand x86memop,
                    PatFrag ld_frag, string asm, Domain d> {
  def irr : MMXPI<opc, MRMSrcReg, (outs DstRC:$dst),
                  (ins DstRC:$src1, SrcRC:$src2), asm,
                  [(set DstRC:$dst, (Int DstRC:$src1, SrcRC:$src2))], d>,
                  Sched<[WriteCvtI2PS]>;
  def irm : MMXPI<opc, MRMSrcMem, (outs DstRC:$dst),
                  (ins DstRC:$src1, x86memop:$src2), asm,
                  [(set DstRC:$dst, (Int DstRC:$src1, (ld_frag addr:$src2)))], d>,
                  Sched<[WriteCvtI2PS.Folded]>;
}

//===----------------------------------------------------------------------===//
// MMX EMMS Instruction
//===----------------------------------------------------------------------===//

let SchedRW = [WriteEMMS],
    Defs = [MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7,
            ST0, ST1, ST2, ST3, ST4, ST5, ST6, ST7] in
def MMX_EMMS  : MMXI<0x77, RawFrm, (outs), (ins), "emms", [(int_x86_mmx_emms)]>;

//===----------------------------------------------------------------------===//
// MMX Scalar Instructions
//===----------------------------------------------------------------------===//

// Data Transfer Instructions
def MMX_MOVD64rr : MMXI<0x6E, MRMSrcReg, (outs VR64:$dst), (ins GR32:$src),
                        "movd\t{$src, $dst|$dst, $src}",
                        [(set VR64:$dst,
                         (x86mmx (MMX_X86movw2d GR32:$src)))]>,
                        Sched<[WriteVecMoveFromGpr]>;
def MMX_MOVD64rm : MMXI<0x6E, MRMSrcMem, (outs VR64:$dst), (ins i32mem:$src),
                        "movd\t{$src, $dst|$dst, $src}",
                        [(set VR64:$dst,
                          (x86mmx (MMX_X86movw2d (loadi32 addr:$src))))]>,
                        Sched<[WriteVecLoad]>;

let mayStore = 1 in
def MMX_MOVD64mr : MMXI<0x7E, MRMDestMem, (outs), (ins i32mem:$dst, VR64:$src),
                        "movd\t{$src, $dst|$dst, $src}", []>,
                   Sched<[WriteVecStore]>;

def MMX_MOVD64grr : MMXI<0x7E, MRMDestReg, (outs GR32:$dst), (ins VR64:$src),
                         "movd\t{$src, $dst|$dst, $src}",
                         [(set GR32:$dst,
                          (MMX_X86movd2w (x86mmx VR64:$src)))]>,
                         Sched<[WriteVecMoveToGpr]>, FoldGenData<"MMX_MOVD64rr">;

let isBitcast = 1 in
def MMX_MOVD64to64rr : MMXRI<0x6E, MRMSrcReg, (outs VR64:$dst), (ins GR64:$src),
                             "movq\t{$src, $dst|$dst, $src}",
                             [(set VR64:$dst, (bitconvert GR64:$src))]>,
                             Sched<[WriteVecMoveFromGpr]>;

let isCodeGenOnly = 1, ForceDisassemble = 1, hasSideEffects = 0, mayLoad = 1 in
def MMX_MOVD64to64rm : MMXRI<0x6E, MRMSrcMem, (outs VR64:$dst),
                             (ins i64mem:$src), "movq\t{$src, $dst|$dst, $src}",
                             []>, Sched<[SchedWriteVecMoveLS.MMX.RM]>;

let isBitcast = 1 in {
def MMX_MOVD64from64rr : MMXRI<0x7E, MRMDestReg,
                               (outs GR64:$dst), (ins VR64:$src),
                               "movq\t{$src, $dst|$dst, $src}",
                               [(set GR64:$dst, (bitconvert VR64:$src))]>,
                               Sched<[WriteVecMoveToGpr]>;
let SchedRW = [WriteVecMove], hasSideEffects = 0, isMoveReg = 1 in {
def MMX_MOVQ64rr : MMXI<0x6F, MRMSrcReg, (outs VR64:$dst), (ins VR64:$src),
                        "movq\t{$src, $dst|$dst, $src}", []>;
let isCodeGenOnly = 1, ForceDisassemble = 1 in
def MMX_MOVQ64rr_REV : MMXI<0x7F, MRMDestReg, (outs VR64:$dst), (ins VR64:$src),
                            "movq\t{$src, $dst|$dst, $src}", []>,
                            FoldGenData<"MMX_MOVQ64rr">;
} // SchedRW, hasSideEffects, isMoveReg
} // isBitcast

def : InstAlias<"movq.s\t{$src, $dst|$dst, $src}",
                (MMX_MOVQ64rr_REV VR64:$dst, VR64:$src), 0>;

let isCodeGenOnly = 1, ForceDisassemble = 1, hasSideEffects = 0, mayStore = 1 in
def MMX_MOVD64from64rm : MMXRI<0x7E, MRMDestMem,
                               (outs), (ins i64mem:$dst, VR64:$src),
                               "movq\t{$src, $dst|$dst, $src}", []>,
                               Sched<[SchedWriteVecMoveLS.MMX.MR]>;

let SchedRW = [SchedWriteVecMoveLS.MMX.RM] in {
let canFoldAsLoad = 1 in
def MMX_MOVQ64rm : MMXI<0x6F, MRMSrcMem, (outs VR64:$dst), (ins i64mem:$src),
                        "movq\t{$src, $dst|$dst, $src}",
                        [(set VR64:$dst, (load_mmx addr:$src))]>;
} // SchedRW

let SchedRW = [SchedWriteVecMoveLS.MMX.MR] in
def MMX_MOVQ64mr : MMXI<0x7F, MRMDestMem, (outs), (ins i64mem:$dst, VR64:$src),
                        "movq\t{$src, $dst|$dst, $src}",
                        [(store (x86mmx VR64:$src), addr:$dst)]>;

def MMX_X86movdq2q : SDNode<"X86ISD::MOVDQ2Q", SDTypeProfile<1, 1,
                            [SDTCisVT<0, x86mmx>, SDTCisVT<1, v2i64>]>>;
def MMX_X86movq2dq : SDNode<"X86ISD::MOVQ2DQ", SDTypeProfile<1, 1,
                            [SDTCisVT<0, v2i64>, SDTCisVT<1, x86mmx>]>>;

let SchedRW = [SchedWriteVecMoveLS.XMM.RR] in {
def MMX_MOVDQ2Qrr : MMXSDIi8<0xD6, MRMSrcReg, (outs VR64:$dst),
                             (ins VR128:$src), "movdq2q\t{$src, $dst|$dst, $src}",
                             [(set VR64:$dst,
                               (x86mmx (MMX_X86movdq2q VR128:$src)))]>;

def MMX_MOVQ2DQrr : MMXS2SIi8<0xD6, MRMSrcReg, (outs VR128:$dst),
                              (ins VR64:$src), "movq2dq\t{$src, $dst|$dst, $src}",
                              [(set VR128:$dst,
                                (v2i64 (MMX_X86movq2dq VR64:$src)))]>;

let isCodeGenOnly = 1, hasSideEffects = 1 in {
def MMX_MOVQ2FR64rr: MMXS2SIi8<0xD6, MRMSrcReg, (outs FR64:$dst),
                               (ins VR64:$src), "movq2dq\t{$src, $dst|$dst, $src}",
                               []>;

def MMX_MOVFR642Qrr: MMXSDIi8<0xD6, MRMSrcReg, (outs VR64:$dst),
                              (ins FR64:$src), "movdq2q\t{$src, $dst|$dst, $src}",
                              []>;
}
} // SchedRW

let Predicates = [HasMMX, HasSSE1] in
def MMX_MOVNTQmr  : MMXI<0xE7, MRMDestMem, (outs), (ins i64mem:$dst, VR64:$src),
                         "movntq\t{$src, $dst|$dst, $src}",
                         [(int_x86_mmx_movnt_dq addr:$dst, VR64:$src)]>,
                         Sched<[SchedWriteVecMoveLSNT.MMX.MR]>;

// Arithmetic Instructions
defm MMX_PABSB : SS3I_unop_rm_int_mm<0x1C, "pabsb", int_x86_ssse3_pabs_b,
                                     SchedWriteVecALU.MMX>;
defm MMX_PABSW : SS3I_unop_rm_int_mm<0x1D, "pabsw", int_x86_ssse3_pabs_w,
                                     SchedWriteVecALU.MMX>;
defm MMX_PABSD : SS3I_unop_rm_int_mm<0x1E, "pabsd", int_x86_ssse3_pabs_d,
                                     SchedWriteVecALU.MMX>;
// -- Addition
defm MMX_PADDB : MMXI_binop_rm_int<0xFC, "paddb", int_x86_mmx_padd_b,
                                   SchedWriteVecALU.MMX, 1>;
defm MMX_PADDW : MMXI_binop_rm_int<0xFD, "paddw", int_x86_mmx_padd_w,
                                   SchedWriteVecALU.MMX, 1>;
defm MMX_PADDD : MMXI_binop_rm_int<0xFE, "paddd", int_x86_mmx_padd_d,
                                   SchedWriteVecALU.MMX, 1>;
let Predicates = [HasMMX, HasSSE2] in
defm MMX_PADDQ : MMXI_binop_rm_int<0xD4, "paddq", int_x86_mmx_padd_q,
                                   SchedWriteVecALU.MMX, 1>;
defm MMX_PADDSB  : MMXI_binop_rm_int<0xEC, "paddsb" , int_x86_mmx_padds_b,
                                     SchedWriteVecALU.MMX, 1>;
defm MMX_PADDSW  : MMXI_binop_rm_int<0xED, "paddsw" , int_x86_mmx_padds_w,
                                     SchedWriteVecALU.MMX, 1>;

defm MMX_PADDUSB : MMXI_binop_rm_int<0xDC, "paddusb", int_x86_mmx_paddus_b,
                                   SchedWriteVecALU.MMX, 1>;
defm MMX_PADDUSW : MMXI_binop_rm_int<0xDD, "paddusw", int_x86_mmx_paddus_w,
                                   SchedWriteVecALU.MMX, 1>;

defm MMX_PHADDW  : SS3I_binop_rm_int_mm<0x01, "phaddw", int_x86_ssse3_phadd_w,
                                        SchedWritePHAdd.MMX>;
defm MMX_PHADDD  : SS3I_binop_rm_int_mm<0x02, "phaddd", int_x86_ssse3_phadd_d,
                                        SchedWritePHAdd.MMX>;
defm MMX_PHADDSW : SS3I_binop_rm_int_mm<0x03, "phaddsw",int_x86_ssse3_phadd_sw,
                                        SchedWritePHAdd.MMX>;

// -- Subtraction
defm MMX_PSUBB : MMXI_binop_rm_int<0xF8, "psubb", int_x86_mmx_psub_b,
                                   SchedWriteVecALU.MMX>;
defm MMX_PSUBW : MMXI_binop_rm_int<0xF9, "psubw", int_x86_mmx_psub_w,
                                   SchedWriteVecALU.MMX>;
defm MMX_PSUBD : MMXI_binop_rm_int<0xFA, "psubd", int_x86_mmx_psub_d,
                                   SchedWriteVecALU.MMX>;
let Predicates = [HasMMX, HasSSE2] in
defm MMX_PSUBQ : MMXI_binop_rm_int<0xFB, "psubq", int_x86_mmx_psub_q,
                                   SchedWriteVecALU.MMX>;

defm MMX_PSUBSB  : MMXI_binop_rm_int<0xE8, "psubsb" , int_x86_mmx_psubs_b,
                                   SchedWriteVecALU.MMX>;
defm MMX_PSUBSW  : MMXI_binop_rm_int<0xE9, "psubsw" , int_x86_mmx_psubs_w,
                                   SchedWriteVecALU.MMX>;

defm MMX_PSUBUSB : MMXI_binop_rm_int<0xD8, "psubusb", int_x86_mmx_psubus_b,
                                   SchedWriteVecALU.MMX>;
defm MMX_PSUBUSW : MMXI_binop_rm_int<0xD9, "psubusw", int_x86_mmx_psubus_w,
                                   SchedWriteVecALU.MMX>;

defm MMX_PHSUBW  : SS3I_binop_rm_int_mm<0x05, "phsubw", int_x86_ssse3_phsub_w,
                                        SchedWritePHAdd.MMX>;
defm MMX_PHSUBD  : SS3I_binop_rm_int_mm<0x06, "phsubd", int_x86_ssse3_phsub_d,
                                        SchedWritePHAdd.MMX>;
defm MMX_PHSUBSW : SS3I_binop_rm_int_mm<0x07, "phsubsw",int_x86_ssse3_phsub_sw,
                                        SchedWritePHAdd.MMX>;

// -- Multiplication
defm MMX_PMULLW  : MMXI_binop_rm_int<0xD5, "pmullw", int_x86_mmx_pmull_w,
                                     SchedWriteVecIMul.MMX, 1>;

defm MMX_PMULHW  : MMXI_binop_rm_int<0xE5, "pmulhw",  int_x86_mmx_pmulh_w,
                                     SchedWriteVecIMul.MMX, 1>;
let Predicates = [HasMMX, HasSSE1] in
defm MMX_PMULHUW : MMXI_binop_rm_int<0xE4, "pmulhuw", int_x86_mmx_pmulhu_w,
                                     SchedWriteVecIMul.MMX, 1>;
let Predicates = [HasMMX, HasSSE2] in
defm MMX_PMULUDQ : MMXI_binop_rm_int<0xF4, "pmuludq", int_x86_mmx_pmulu_dq,
                                     SchedWriteVecIMul.MMX, 1>;
defm MMX_PMULHRSW : SS3I_binop_rm_int_mm<0x0B, "pmulhrsw",
                                     int_x86_ssse3_pmul_hr_sw,
                                     SchedWriteVecIMul.MMX, 1>;

// -- Miscellanea
defm MMX_PMADDWD : MMXI_binop_rm_int<0xF5, "pmaddwd", int_x86_mmx_pmadd_wd,
                                     SchedWriteVecIMul.MMX, 1>;

defm MMX_PMADDUBSW : SS3I_binop_rm_int_mm<0x04, "pmaddubsw",
                                          int_x86_ssse3_pmadd_ub_sw,
                                          SchedWriteVecIMul.MMX>;
let Predicates = [HasMMX, HasSSE1] in {
defm MMX_PAVGB   : MMXI_binop_rm_int<0xE0, "pavgb", int_x86_mmx_pavg_b,
                                     SchedWriteVecALU.MMX, 1>;
defm MMX_PAVGW   : MMXI_binop_rm_int<0xE3, "pavgw", int_x86_mmx_pavg_w,
                                     SchedWriteVecALU.MMX, 1>;

defm MMX_PMINUB  : MMXI_binop_rm_int<0xDA, "pminub", int_x86_mmx_pminu_b,
                                     SchedWriteVecALU.MMX, 1>;
defm MMX_PMINSW  : MMXI_binop_rm_int<0xEA, "pminsw", int_x86_mmx_pmins_w,
                                     SchedWriteVecALU.MMX, 1>;

defm MMX_PMAXUB  : MMXI_binop_rm_int<0xDE, "pmaxub", int_x86_mmx_pmaxu_b,
                                     SchedWriteVecALU.MMX, 1>;
defm MMX_PMAXSW  : MMXI_binop_rm_int<0xEE, "pmaxsw", int_x86_mmx_pmaxs_w,
                                     SchedWriteVecALU.MMX, 1>;

defm MMX_PSADBW  : MMXI_binop_rm_int<0xF6, "psadbw", int_x86_mmx_psad_bw,
                                     SchedWritePSADBW.MMX, 1>;
}

defm MMX_PSIGNB :  SS3I_binop_rm_int_mm<0x08, "psignb", int_x86_ssse3_psign_b,
                                        SchedWriteVecALU.MMX>;
defm MMX_PSIGNW :  SS3I_binop_rm_int_mm<0x09, "psignw", int_x86_ssse3_psign_w,
                                        SchedWriteVecALU.MMX>;
defm MMX_PSIGND :  SS3I_binop_rm_int_mm<0x0A, "psignd", int_x86_ssse3_psign_d,
                                        SchedWriteVecALU.MMX>;
let Constraints = "$src1 = $dst" in
  defm MMX_PALIGNR : ssse3_palign_mm<"palignr", int_x86_mmx_palignr_b,
                                     SchedWriteShuffle.MMX>;

// Logical Instructions
defm MMX_PAND : MMXI_binop_rm_int<0xDB, "pand", int_x86_mmx_pand,
                                  SchedWriteVecLogic.MMX, 1>;
defm MMX_POR  : MMXI_binop_rm_int<0xEB, "por" , int_x86_mmx_por,
                                  SchedWriteVecLogic.MMX, 1>;
defm MMX_PXOR : MMXI_binop_rm_int<0xEF, "pxor", int_x86_mmx_pxor,
                                  SchedWriteVecLogic.MMX, 1>;
defm MMX_PANDN : MMXI_binop_rm_int<0xDF, "pandn", int_x86_mmx_pandn,
                                   SchedWriteVecLogic.MMX>;

// Shift Instructions
defm MMX_PSRLW : MMXI_binop_rmi_int<0xD1, 0x71, MRM2r, "psrlw",
                                    int_x86_mmx_psrl_w, int_x86_mmx_psrli_w,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;
defm MMX_PSRLD : MMXI_binop_rmi_int<0xD2, 0x72, MRM2r, "psrld",
                                    int_x86_mmx_psrl_d, int_x86_mmx_psrli_d,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;
defm MMX_PSRLQ : MMXI_binop_rmi_int<0xD3, 0x73, MRM2r, "psrlq",
                                    int_x86_mmx_psrl_q, int_x86_mmx_psrli_q,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;

defm MMX_PSLLW : MMXI_binop_rmi_int<0xF1, 0x71, MRM6r, "psllw",
                                    int_x86_mmx_psll_w, int_x86_mmx_pslli_w,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;
defm MMX_PSLLD : MMXI_binop_rmi_int<0xF2, 0x72, MRM6r, "pslld",
                                    int_x86_mmx_psll_d, int_x86_mmx_pslli_d,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;
defm MMX_PSLLQ : MMXI_binop_rmi_int<0xF3, 0x73, MRM6r, "psllq",
                                    int_x86_mmx_psll_q, int_x86_mmx_pslli_q,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;

defm MMX_PSRAW : MMXI_binop_rmi_int<0xE1, 0x71, MRM4r, "psraw",
                                    int_x86_mmx_psra_w, int_x86_mmx_psrai_w,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;
defm MMX_PSRAD : MMXI_binop_rmi_int<0xE2, 0x72, MRM4r, "psrad",
                                    int_x86_mmx_psra_d, int_x86_mmx_psrai_d,
                                    SchedWriteVecShift.MMX,
                                    SchedWriteVecShiftImm.MMX>;

// Comparison Instructions
defm MMX_PCMPEQB : MMXI_binop_rm_int<0x74, "pcmpeqb", int_x86_mmx_pcmpeq_b,
                                     SchedWriteVecALU.MMX>;
defm MMX_PCMPEQW : MMXI_binop_rm_int<0x75, "pcmpeqw", int_x86_mmx_pcmpeq_w,
                                     SchedWriteVecALU.MMX>;
defm MMX_PCMPEQD : MMXI_binop_rm_int<0x76, "pcmpeqd", int_x86_mmx_pcmpeq_d,
                                     SchedWriteVecALU.MMX>;

defm MMX_PCMPGTB : MMXI_binop_rm_int<0x64, "pcmpgtb", int_x86_mmx_pcmpgt_b,
                                     SchedWriteVecALU.MMX>;
defm MMX_PCMPGTW : MMXI_binop_rm_int<0x65, "pcmpgtw", int_x86_mmx_pcmpgt_w,
                                     SchedWriteVecALU.MMX>;
defm MMX_PCMPGTD : MMXI_binop_rm_int<0x66, "pcmpgtd", int_x86_mmx_pcmpgt_d,
                                     SchedWriteVecALU.MMX>;

// -- Unpack Instructions
defm MMX_PUNPCKHBW : MMXI_binop_rm_int<0x68, "punpckhbw",
                                       int_x86_mmx_punpckhbw,
                                       SchedWriteShuffle.MMX>;
defm MMX_PUNPCKHWD : MMXI_binop_rm_int<0x69, "punpckhwd",
                                       int_x86_mmx_punpckhwd,
                                       SchedWriteShuffle.MMX>;
defm MMX_PUNPCKHDQ : MMXI_binop_rm_int<0x6A, "punpckhdq",
                                       int_x86_mmx_punpckhdq,
                                       SchedWriteShuffle.MMX>;
defm MMX_PUNPCKLBW : MMXI_binop_rm_int<0x60, "punpcklbw",
                                       int_x86_mmx_punpcklbw,
                                       SchedWriteShuffle.MMX,
                                       0, i32mem>;
defm MMX_PUNPCKLWD : MMXI_binop_rm_int<0x61, "punpcklwd",
                                       int_x86_mmx_punpcklwd,
                                       SchedWriteShuffle.MMX,
                                       0, i32mem>;
defm MMX_PUNPCKLDQ : MMXI_binop_rm_int<0x62, "punpckldq",
                                       int_x86_mmx_punpckldq,
                                       SchedWriteShuffle.MMX,
                                       0, i32mem>;

// -- Pack Instructions
defm MMX_PACKSSWB : MMXI_binop_rm_int<0x63, "packsswb", int_x86_mmx_packsswb,
                                      SchedWriteShuffle.MMX>;
defm MMX_PACKSSDW : MMXI_binop_rm_int<0x6B, "packssdw", int_x86_mmx_packssdw,
                                      SchedWriteShuffle.MMX>;
defm MMX_PACKUSWB : MMXI_binop_rm_int<0x67, "packuswb", int_x86_mmx_packuswb,
                                      SchedWriteShuffle.MMX>;

// -- Shuffle Instructions
defm MMX_PSHUFB : SS3I_binop_rm_int_mm<0x00, "pshufb", int_x86_ssse3_pshuf_b,
                                       SchedWriteVarShuffle.MMX>;

let Predicates = [HasMMX, HasSSE1] in {
def MMX_PSHUFWri : MMXIi8<0x70, MRMSrcReg,
                          (outs VR64:$dst), (ins VR64:$src1, u8imm:$src2),
                          "pshufw\t{$src2, $src1, $dst|$dst, $src1, $src2}",
                          [(set VR64:$dst,
                             (int_x86_sse_pshuf_w VR64:$src1, timm:$src2))]>,
                          Sched<[SchedWriteShuffle.MMX]>;
def MMX_PSHUFWmi : MMXIi8<0x70, MRMSrcMem,
                          (outs VR64:$dst), (ins i64mem:$src1, u8imm:$src2),
                          "pshufw\t{$src2, $src1, $dst|$dst, $src1, $src2}",
                          [(set VR64:$dst,
                             (int_x86_sse_pshuf_w (load_mmx addr:$src1),
                                                   timm:$src2))]>,
                          Sched<[SchedWriteShuffle.MMX.Folded]>;
}

// -- Conversion Instructions
defm MMX_CVTPS2PI : sse12_cvt_pint<0x2D, VR128, VR64, int_x86_sse_cvtps2pi,
                      f64mem, load, "cvtps2pi\t{$src, $dst|$dst, $src}",
                      WriteCvtPS2I, SSEPackedSingle>, PS, SIMD_EXC;
defm MMX_CVTPD2PI : sse12_cvt_pint<0x2D, VR128, VR64, int_x86_sse_cvtpd2pi,
                      f128mem, memop, "cvtpd2pi\t{$src, $dst|$dst, $src}",
                      WriteCvtPD2I, SSEPackedDouble>, PD, SIMD_EXC;
defm MMX_CVTTPS2PI : sse12_cvt_pint<0x2C, VR128, VR64, int_x86_sse_cvttps2pi,
                       f64mem, load, "cvttps2pi\t{$src, $dst|$dst, $src}",
                       WriteCvtPS2I, SSEPackedSingle>, PS, SIMD_EXC;
defm MMX_CVTTPD2PI : sse12_cvt_pint<0x2C, VR128, VR64, int_x86_sse_cvttpd2pi,
                       f128mem, memop, "cvttpd2pi\t{$src, $dst|$dst, $src}",
                       WriteCvtPD2I, SSEPackedDouble>, PD, SIMD_EXC;
defm MMX_CVTPI2PD : sse12_cvt_pint<0x2A, VR64, VR128, int_x86_sse_cvtpi2pd,
                         i64mem, load, "cvtpi2pd\t{$src, $dst|$dst, $src}",
                         WriteCvtI2PD, SSEPackedDouble>, PD;
let Constraints = "$src1 = $dst" in {
  defm MMX_CVTPI2PS : sse12_cvt_pint_3addr<0x2A, VR64, VR128,
                         int_x86_sse_cvtpi2ps,
                         i64mem, load, "cvtpi2ps\t{$src2, $dst|$dst, $src2}",
                         SSEPackedSingle>, PS, SIMD_EXC;
}

// Extract / Insert
let Predicates = [HasMMX, HasSSE1] in
def MMX_PEXTRWrr: MMXIi8<0xC5, MRMSrcReg,
                     (outs GR32orGR64:$dst), (ins VR64:$src1, i32u8imm:$src2),
                     "pextrw\t{$src2, $src1, $dst|$dst, $src1, $src2}",
                     [(set GR32orGR64:$dst, (int_x86_mmx_pextr_w VR64:$src1,
                                             timm:$src2))]>,
                     Sched<[WriteVecExtract]>;
let Constraints = "$src1 = $dst" in {
let Predicates = [HasMMX, HasSSE1] in {
  def MMX_PINSRWrr : MMXIi8<0xC4, MRMSrcReg,
                    (outs VR64:$dst),
                    (ins VR64:$src1, GR32orGR64:$src2, i32u8imm:$src3),
                    "pinsrw\t{$src3, $src2, $dst|$dst, $src2, $src3}",
                    [(set VR64:$dst, (int_x86_mmx_pinsr_w VR64:$src1,
                                      GR32orGR64:$src2, timm:$src3))]>,
                    Sched<[WriteVecInsert, ReadDefault, ReadInt2Fpu]>;

  def MMX_PINSRWrm : MMXIi8<0xC4, MRMSrcMem,
                   (outs VR64:$dst),
                   (ins VR64:$src1, i16mem:$src2, i32u8imm:$src3),
                   "pinsrw\t{$src3, $src2, $dst|$dst, $src2, $src3}",
                   [(set VR64:$dst, (int_x86_mmx_pinsr_w VR64:$src1,
                                       (i32 (anyext (loadi16 addr:$src2))),
                                     timm:$src3))]>,
                   Sched<[WriteVecInsert.Folded, WriteVecInsert.ReadAfterFold]>;
}
}

// Mask creation
let Predicates = [HasMMX, HasSSE1] in
def MMX_PMOVMSKBrr : MMXI<0xD7, MRMSrcReg, (outs GR32orGR64:$dst),
                          (ins VR64:$src),
                          "pmovmskb\t{$src, $dst|$dst, $src}",
                          [(set GR32orGR64:$dst,
                                (int_x86_mmx_pmovmskb VR64:$src))]>,
                          Sched<[WriteMMXMOVMSK]>;

// Misc.
let SchedRW = [SchedWriteShuffle.MMX] in {
let Uses = [EDI], Predicates = [HasMMX, HasSSE1,Not64BitMode] in
def MMX_MASKMOVQ : MMXI32<0xF7, MRMSrcReg, (outs), (ins VR64:$src, VR64:$mask),
                          "maskmovq\t{$mask, $src|$src, $mask}",
                          [(int_x86_mmx_maskmovq VR64:$src, VR64:$mask, EDI)]>;
let Uses = [RDI], Predicates = [HasMMX, HasSSE1,In64BitMode] in
def MMX_MASKMOVQ64: MMXI64<0xF7, MRMSrcReg, (outs), (ins VR64:$src, VR64:$mask),
                           "maskmovq\t{$mask, $src|$src, $mask}",
                           [(int_x86_mmx_maskmovq VR64:$src, VR64:$mask, RDI)]>;
}

// 64-bit bit convert.
let Predicates = [HasMMX, HasSSE2] in {
def : Pat<(f64 (bitconvert (x86mmx VR64:$src))),
          (MMX_MOVQ2FR64rr VR64:$src)>;
def : Pat<(x86mmx (bitconvert (f64 FR64:$src))),
          (MMX_MOVFR642Qrr FR64:$src)>;
def : Pat<(x86mmx (MMX_X86movdq2q
                   (bc_v2i64 (v4i32 (X86cvtp2Int (v4f32 VR128:$src)))))),
          (MMX_CVTPS2PIirr VR128:$src)>;
def : Pat<(x86mmx (MMX_X86movdq2q
                   (bc_v2i64 (v4i32 (X86cvttp2si (v4f32 VR128:$src)))))),
          (MMX_CVTTPS2PIirr VR128:$src)>;
def : Pat<(x86mmx (MMX_X86movdq2q
                   (bc_v2i64 (v4i32 (X86cvtp2Int (v2f64 VR128:$src)))))),
          (MMX_CVTPD2PIirr VR128:$src)>;
def : Pat<(x86mmx (MMX_X86movdq2q
                   (bc_v2i64 (v4i32 (X86cvttp2si (v2f64 VR128:$src)))))),
          (MMX_CVTTPD2PIirr VR128:$src)>;
}