forked from OSchip/llvm-project
2812 lines
118 KiB
TableGen
Executable File
2812 lines
118 KiB
TableGen
Executable File
//=- X86SchedBroadwell.td - X86 Broadwell Scheduling ---------*- tablegen -*-=//
|
|
//
|
|
// The LLVM Compiler Infrastructure
|
|
//
|
|
// This file is distributed under the University of Illinois Open Source
|
|
// License. See LICENSE.TXT for details.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// This file defines the machine model for Broadwell to support instruction
|
|
// scheduling and other instruction cost heuristics.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
def BroadwellModel : SchedMachineModel {
|
|
// All x86 instructions are modeled as a single micro-op, and HW can decode 4
|
|
// instructions per cycle.
|
|
let IssueWidth = 4;
|
|
let MicroOpBufferSize = 192; // Based on the reorder buffer.
|
|
let LoadLatency = 5;
|
|
let MispredictPenalty = 16;
|
|
|
|
// Based on the LSD (loop-stream detector) queue size and benchmarking data.
|
|
let LoopMicroOpBufferSize = 50;
|
|
|
|
// This flag is set to allow the scheduler to assign a default model to
|
|
// unrecognized opcodes.
|
|
let CompleteModel = 0;
|
|
}
|
|
|
|
let SchedModel = BroadwellModel in {
|
|
|
|
// Broadwell can issue micro-ops to 8 different ports in one cycle.
|
|
|
|
// Ports 0, 1, 5, and 6 handle all computation.
|
|
// Port 4 gets the data half of stores. Store data can be available later than
|
|
// the store address, but since we don't model the latency of stores, we can
|
|
// ignore that.
|
|
// Ports 2 and 3 are identical. They handle loads and the address half of
|
|
// stores. Port 7 can handle address calculations.
|
|
def BWPort0 : ProcResource<1>;
|
|
def BWPort1 : ProcResource<1>;
|
|
def BWPort2 : ProcResource<1>;
|
|
def BWPort3 : ProcResource<1>;
|
|
def BWPort4 : ProcResource<1>;
|
|
def BWPort5 : ProcResource<1>;
|
|
def BWPort6 : ProcResource<1>;
|
|
def BWPort7 : ProcResource<1>;
|
|
|
|
// Many micro-ops are capable of issuing on multiple ports.
|
|
def BWPort01 : ProcResGroup<[BWPort0, BWPort1]>;
|
|
def BWPort23 : ProcResGroup<[BWPort2, BWPort3]>;
|
|
def BWPort237 : ProcResGroup<[BWPort2, BWPort3, BWPort7]>;
|
|
def BWPort04 : ProcResGroup<[BWPort0, BWPort4]>;
|
|
def BWPort05 : ProcResGroup<[BWPort0, BWPort5]>;
|
|
def BWPort06 : ProcResGroup<[BWPort0, BWPort6]>;
|
|
def BWPort15 : ProcResGroup<[BWPort1, BWPort5]>;
|
|
def BWPort16 : ProcResGroup<[BWPort1, BWPort6]>;
|
|
def BWPort56 : ProcResGroup<[BWPort5, BWPort6]>;
|
|
def BWPort015 : ProcResGroup<[BWPort0, BWPort1, BWPort5]>;
|
|
def BWPort056 : ProcResGroup<[BWPort0, BWPort5, BWPort6]>;
|
|
def BWPort0156: ProcResGroup<[BWPort0, BWPort1, BWPort5, BWPort6]>;
|
|
|
|
// 60 Entry Unified Scheduler
|
|
def BWPortAny : ProcResGroup<[BWPort0, BWPort1, BWPort2, BWPort3, BWPort4,
|
|
BWPort5, BWPort6, BWPort7]> {
|
|
let BufferSize=60;
|
|
}
|
|
|
|
// Integer division issued on port 0.
|
|
def BWDivider : ProcResource<1>;
|
|
// FP division and sqrt on port 0.
|
|
def BWFPDivider : ProcResource<1>;
|
|
|
|
// Loads are 5 cycles, so ReadAfterLd registers needn't be available until 5
|
|
// cycles after the memory operand.
|
|
def : ReadAdvance<ReadAfterLd, 5>;
|
|
|
|
// Many SchedWrites are defined in pairs with and without a folded load.
|
|
// Instructions with folded loads are usually micro-fused, so they only appear
|
|
// as two micro-ops when queued in the reservation station.
|
|
// This multiclass defines the resource usage for variants with and without
|
|
// folded loads.
|
|
multiclass BWWriteResPair<X86FoldableSchedWrite SchedRW,
|
|
list<ProcResourceKind> ExePorts,
|
|
int Lat, list<int> Res = [1], int UOps = 1,
|
|
int LoadLat = 5> {
|
|
// Register variant is using a single cycle on ExePort.
|
|
def : WriteRes<SchedRW, ExePorts> {
|
|
let Latency = Lat;
|
|
let ResourceCycles = Res;
|
|
let NumMicroOps = UOps;
|
|
}
|
|
|
|
// Memory variant also uses a cycle on port 2/3 and adds LoadLat cycles to
|
|
// the latency (default = 5).
|
|
def : WriteRes<SchedRW.Folded, !listconcat([BWPort23], ExePorts)> {
|
|
let Latency = !add(Lat, LoadLat);
|
|
let ResourceCycles = !listconcat([1], Res);
|
|
let NumMicroOps = !add(UOps, 1);
|
|
}
|
|
}
|
|
|
|
// A folded store needs a cycle on port 4 for the store data, and an extra port
|
|
// 2/3/7 cycle to recompute the address.
|
|
def : WriteRes<WriteRMW, [BWPort237,BWPort4]>;
|
|
|
|
// Arithmetic.
|
|
defm : BWWriteResPair<WriteALU, [BWPort0156], 1>; // Simple integer ALU op.
|
|
defm : BWWriteResPair<WriteIMul, [BWPort1], 3>; // Integer multiplication.
|
|
defm : BWWriteResPair<WriteIDiv, [BWPort0, BWDivider], 25, [1, 10]>;
|
|
defm : BWWriteResPair<WriteCRC32, [BWPort1], 3>;
|
|
def : WriteRes<WriteIMulH, []> { let Latency = 3; } // Integer multiplication, high part.
|
|
|
|
def : WriteRes<WriteLEA, [BWPort15]>; // LEA instructions can't fold loads.
|
|
|
|
// Bit counts.
|
|
defm : BWWriteResPair<WriteBitScan, [BWPort1], 3>;
|
|
defm : BWWriteResPair<WriteLZCNT, [BWPort1], 3>;
|
|
defm : BWWriteResPair<WriteTZCNT, [BWPort1], 3>;
|
|
defm : BWWriteResPair<WritePOPCNT, [BWPort1], 3>;
|
|
|
|
// Integer shifts and rotates.
|
|
defm : BWWriteResPair<WriteShift, [BWPort06], 1>;
|
|
|
|
// BMI1 BEXTR, BMI2 BZHI
|
|
defm : BWWriteResPair<WriteBEXTR, [BWPort06,BWPort15], 2, [1,1], 2>;
|
|
defm : BWWriteResPair<WriteBZHI, [BWPort15], 1>;
|
|
|
|
// Loads, stores, and moves, not folded with other operations.
|
|
def : WriteRes<WriteLoad, [BWPort23]> { let Latency = 5; }
|
|
def : WriteRes<WriteStore, [BWPort237, BWPort4]>;
|
|
def : WriteRes<WriteMove, [BWPort0156]>;
|
|
|
|
// Idioms that clear a register, like xorps %xmm0, %xmm0.
|
|
// These can often bypass execution ports completely.
|
|
def : WriteRes<WriteZero, []>;
|
|
|
|
// Treat misc copies as a move.
|
|
def : InstRW<[WriteMove], (instrs COPY)>;
|
|
|
|
// Branches don't produce values, so they have no latency, but they still
|
|
// consume resources. Indirect branches can fold loads.
|
|
defm : BWWriteResPair<WriteJump, [BWPort06], 1>;
|
|
|
|
// Floating point. This covers both scalar and vector operations.
|
|
def : WriteRes<WriteFLoad, [BWPort23]> { let Latency = 5; }
|
|
def : WriteRes<WriteFStore, [BWPort237, BWPort4]>;
|
|
def : WriteRes<WriteFMove, [BWPort5]>;
|
|
|
|
defm : BWWriteResPair<WriteFAdd, [BWPort1], 3>; // Floating point add/sub/compare.
|
|
defm : BWWriteResPair<WriteFMul, [BWPort0], 5>; // Floating point multiplication.
|
|
defm : BWWriteResPair<WriteFDiv, [BWPort0], 12>; // 10-14 cycles. // Floating point division.
|
|
defm : BWWriteResPair<WriteFSqrt, [BWPort0], 15>; // Floating point square root.
|
|
defm : BWWriteResPair<WriteFRcp, [BWPort0], 5>; // Floating point reciprocal estimate.
|
|
defm : BWWriteResPair<WriteFRsqrt, [BWPort0], 5>; // Floating point reciprocal square root estimate.
|
|
defm : BWWriteResPair<WriteFMA, [BWPort01], 5>; // Fused Multiply Add.
|
|
defm : BWWriteResPair<WriteFShuffle, [BWPort5], 1>; // Floating point vector shuffles.
|
|
defm : BWWriteResPair<WriteFBlend, [BWPort015], 1>; // Floating point vector blends.
|
|
defm : BWWriteResPair<WriteFVarBlend, [HWPort5], 2, [2]>; // Fp vector variable blends.
|
|
|
|
// FMA Scheduling helper class.
|
|
// class FMASC { X86FoldableSchedWrite Sched = WriteFAdd; }
|
|
|
|
// Vector integer operations.
|
|
def : WriteRes<WriteVecLoad, [BWPort23]> { let Latency = 5; }
|
|
def : WriteRes<WriteVecStore, [BWPort237, BWPort4]>;
|
|
def : WriteRes<WriteVecMove, [BWPort015]>;
|
|
|
|
defm : BWWriteResPair<WriteVecALU, [BWPort15], 1>; // Vector integer ALU op, no logicals.
|
|
defm : BWWriteResPair<WriteVecShift, [BWPort0], 1>; // Vector integer shifts.
|
|
defm : BWWriteResPair<WriteVecIMul, [BWPort0], 5>; // Vector integer multiply.
|
|
defm : BWWriteResPair<WritePMULLD, [BWPort0], 10, [2], 2, 5>; // PMULLD
|
|
defm : BWWriteResPair<WriteShuffle, [BWPort5], 1>; // Vector shuffles.
|
|
defm : BWWriteResPair<WriteBlend, [BWPort15], 1>; // Vector blends.
|
|
defm : BWWriteResPair<WriteVarBlend, [BWPort5], 2, [2]>; // Vector variable blends.
|
|
defm : BWWriteResPair<WriteMPSAD, [BWPort0, BWPort5], 6, [1, 2]>; // Vector MPSAD.
|
|
|
|
// Vector bitwise operations.
|
|
// These are often used on both floating point and integer vectors.
|
|
defm : BWWriteResPair<WriteVecLogic, [BWPort015], 1>; // Vector and/or/xor.
|
|
|
|
// Conversion between integer and float.
|
|
defm : BWWriteResPair<WriteCvtF2I, [BWPort1], 3>; // Float -> Integer.
|
|
defm : BWWriteResPair<WriteCvtI2F, [BWPort1], 4>; // Integer -> Float.
|
|
defm : BWWriteResPair<WriteCvtF2F, [BWPort1], 3>; // Float -> Float size conversion.
|
|
|
|
// Strings instructions.
|
|
|
|
// Packed Compare Implicit Length Strings, Return Mask
|
|
def : WriteRes<WritePCmpIStrM, [BWPort0]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [3];
|
|
}
|
|
def : WriteRes<WritePCmpIStrMLd, [BWPort0, BWPort23]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [3,1];
|
|
}
|
|
|
|
// Packed Compare Explicit Length Strings, Return Mask
|
|
def : WriteRes<WritePCmpEStrM, [BWPort0, BWPort5, BWPort015, BWPort0156]> {
|
|
let Latency = 19;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [4,3,1,1];
|
|
}
|
|
def : WriteRes<WritePCmpEStrMLd, [BWPort0, BWPort5, BWPort23, BWPort015, BWPort0156]> {
|
|
let Latency = 24;
|
|
let NumMicroOps = 10;
|
|
let ResourceCycles = [4,3,1,1,1];
|
|
}
|
|
|
|
// Packed Compare Implicit Length Strings, Return Index
|
|
def : WriteRes<WritePCmpIStrI, [BWPort0]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [3];
|
|
}
|
|
def : WriteRes<WritePCmpIStrILd, [BWPort0, BWPort23]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [3,1];
|
|
}
|
|
|
|
// Packed Compare Explicit Length Strings, Return Index
|
|
def : WriteRes<WritePCmpEStrI, [BWPort0, BWPort5, BWPort0156]> {
|
|
let Latency = 18;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [4,3,1];
|
|
}
|
|
def : WriteRes<WritePCmpEStrILd, [BWPort0, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 23;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [4,3,1,1];
|
|
}
|
|
|
|
// MOVMSK Instructions.
|
|
def : WriteRes<WriteFMOVMSK, [BWPort0]> { let Latency = 3; }
|
|
def : WriteRes<WriteVecMOVMSK, [BWPort0]> { let Latency = 3; }
|
|
def : WriteRes<WriteMMXMOVMSK, [BWPort0]> { let Latency = 1; }
|
|
|
|
// AES instructions.
|
|
def : WriteRes<WriteAESDecEnc, [BWPort5]> { // Decryption, encryption.
|
|
let Latency = 7;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def : WriteRes<WriteAESDecEncLd, [BWPort5, BWPort23]> {
|
|
let Latency = 12;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
|
|
def : WriteRes<WriteAESIMC, [BWPort5]> { // InvMixColumn.
|
|
let Latency = 14;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [2];
|
|
}
|
|
def : WriteRes<WriteAESIMCLd, [BWPort5, BWPort23]> {
|
|
let Latency = 19;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
|
|
def : WriteRes<WriteAESKeyGen, [BWPort0, BWPort5, BWPort015]> { // Key Generation.
|
|
let Latency = 29;
|
|
let NumMicroOps = 11;
|
|
let ResourceCycles = [2,7,2];
|
|
}
|
|
def : WriteRes<WriteAESKeyGenLd, [BWPort0, BWPort5, BWPort23, BWPort015]> {
|
|
let Latency = 33;
|
|
let NumMicroOps = 11;
|
|
let ResourceCycles = [2,7,1,1];
|
|
}
|
|
|
|
// Carry-less multiplication instructions.
|
|
defm : BWWriteResPair<WriteCLMul, [BWPort0], 5>;
|
|
|
|
// Catch-all for expensive system instructions.
|
|
def : WriteRes<WriteSystem, [BWPort0156]> { let Latency = 100; } // def WriteSystem : SchedWrite;
|
|
|
|
// AVX2.
|
|
defm : BWWriteResPair<WriteFShuffle256, [BWPort5], 3>; // Fp 256-bit width vector shuffles.
|
|
defm : BWWriteResPair<WriteShuffle256, [BWPort5], 3>; // 256-bit width vector shuffles.
|
|
defm : BWWriteResPair<WriteVarVecShift, [BWPort0, BWPort5], 2, [2, 1]>; // Variable vector shifts.
|
|
|
|
// Old microcoded instructions that nobody use.
|
|
def : WriteRes<WriteMicrocoded, [BWPort0156]> { let Latency = 100; } // def WriteMicrocoded : SchedWrite;
|
|
|
|
// Fence instructions.
|
|
def : WriteRes<WriteFence, [BWPort23, BWPort4]>;
|
|
|
|
// Nop, not very useful expect it provides a model for nops!
|
|
def : WriteRes<WriteNop, []>;
|
|
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
// Horizontal add/sub instructions.
|
|
////////////////////////////////////////////////////////////////////////////////
|
|
|
|
defm : BWWriteResPair<WriteFHAdd, [BWPort1], 3>;
|
|
defm : BWWriteResPair<WritePHAdd, [BWPort15], 1>;
|
|
|
|
// Remaining instrs.
|
|
|
|
def BWWriteResGroup1 : SchedWriteRes<[BWPort0]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup1], (instregex "MMX_MOVD64from64rr",
|
|
"MMX_MOVD64grr",
|
|
"MMX_PSLLDri",
|
|
"MMX_PSLLDrr",
|
|
"MMX_PSLLQri",
|
|
"MMX_PSLLQrr",
|
|
"MMX_PSLLWri",
|
|
"MMX_PSLLWrr",
|
|
"MMX_PSRADri",
|
|
"MMX_PSRADrr",
|
|
"MMX_PSRAWri",
|
|
"MMX_PSRAWrr",
|
|
"MMX_PSRLDri",
|
|
"MMX_PSRLDrr",
|
|
"MMX_PSRLQri",
|
|
"MMX_PSRLQrr",
|
|
"MMX_PSRLWri",
|
|
"MMX_PSRLWrr",
|
|
"(V?)MOVPDI2DIrr",
|
|
"(V?)MOVPQIto64rr",
|
|
"(V?)PSLLD(Y?)ri",
|
|
"(V?)PSLLQ(Y?)ri",
|
|
"VPSLLVQ(Y?)rr",
|
|
"(V?)PSLLW(Y?)ri",
|
|
"(V?)PSRAD(Y?)ri",
|
|
"(V?)PSRAW(Y?)ri",
|
|
"(V?)PSRLD(Y?)ri",
|
|
"(V?)PSRLQ(Y?)ri",
|
|
"VPSRLVQ(Y?)rr",
|
|
"(V?)PSRLW(Y?)ri",
|
|
"VTESTPD(Y?)rr",
|
|
"VTESTPS(Y?)rr")>;
|
|
|
|
def BWWriteResGroup2 : SchedWriteRes<[BWPort1]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup2], (instregex "COMP_FST0r",
|
|
"COM_FST0r",
|
|
"UCOM_FPr",
|
|
"UCOM_Fr")>;
|
|
|
|
def BWWriteResGroup3 : SchedWriteRes<[BWPort5]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup3], (instregex "MMX_MOVD64rr",
|
|
"MMX_MOVD64to64rr",
|
|
"MMX_MOVQ2DQrr",
|
|
"MMX_PALIGNRrri",
|
|
"MMX_PSHUFBrr",
|
|
"MMX_PSHUFWri",
|
|
"MMX_PUNPCKHBWirr",
|
|
"MMX_PUNPCKHDQirr",
|
|
"MMX_PUNPCKHWDirr",
|
|
"MMX_PUNPCKLBWirr",
|
|
"MMX_PUNPCKLDQirr",
|
|
"MMX_PUNPCKLWDirr",
|
|
"(V?)ANDNPD(Y?)rr",
|
|
"(V?)ANDNPS(Y?)rr",
|
|
"(V?)ANDPD(Y?)rr",
|
|
"(V?)ANDPS(Y?)rr",
|
|
"VBROADCASTSSrr",
|
|
"(V?)INSERTPSrr",
|
|
"(V?)MOV64toPQIrr",
|
|
"(V?)MOVAPD(Y?)rr",
|
|
"(V?)MOVAPS(Y?)rr",
|
|
"(V?)MOVDDUP(Y?)rr",
|
|
"(V?)MOVDI2PDIrr",
|
|
"(V?)MOVHLPSrr",
|
|
"(V?)MOVLHPSrr",
|
|
"(V?)MOVSDrr",
|
|
"(V?)MOVSHDUP(Y?)rr",
|
|
"(V?)MOVSLDUP(Y?)rr",
|
|
"(V?)MOVSSrr",
|
|
"(V?)MOVUPD(Y?)rr",
|
|
"(V?)MOVUPS(Y?)rr",
|
|
"(V?)ORPD(Y?)rr",
|
|
"(V?)ORPS(Y?)rr",
|
|
"(V?)PACKSSDW(Y?)rr",
|
|
"(V?)PACKSSWB(Y?)rr",
|
|
"(V?)PACKUSDW(Y?)rr",
|
|
"(V?)PACKUSWB(Y?)rr",
|
|
"(V?)PALIGNR(Y?)rri",
|
|
"(V?)PBLENDW(Y?)rri",
|
|
"VPBROADCASTDrr",
|
|
"VPBROADCASTQrr",
|
|
"VPERMILPD(Y?)ri",
|
|
"VPERMILPD(Y?)rr",
|
|
"VPERMILPS(Y?)ri",
|
|
"VPERMILPS(Y?)rr",
|
|
"(V?)PMOVSXBDrr",
|
|
"(V?)PMOVSXBQrr",
|
|
"(V?)PMOVSXBWrr",
|
|
"(V?)PMOVSXDQrr",
|
|
"(V?)PMOVSXWDrr",
|
|
"(V?)PMOVSXWQrr",
|
|
"(V?)PMOVZXBDrr",
|
|
"(V?)PMOVZXBQrr",
|
|
"(V?)PMOVZXBWrr",
|
|
"(V?)PMOVZXDQrr",
|
|
"(V?)PMOVZXWDrr",
|
|
"(V?)PMOVZXWQrr",
|
|
"(V?)PSHUFB(Y?)rr",
|
|
"(V?)PSHUFD(Y?)ri",
|
|
"(V?)PSHUFHW(Y?)ri",
|
|
"(V?)PSHUFLW(Y?)ri",
|
|
"(V?)PSLLDQ(Y?)ri",
|
|
"(V?)PSRLDQ(Y?)ri",
|
|
"(V?)PUNPCKHBW(Y?)rr",
|
|
"(V?)PUNPCKHDQ(Y?)rr",
|
|
"(V?)PUNPCKHQDQ(Y?)rr",
|
|
"(V?)PUNPCKHWD(Y?)rr",
|
|
"(V?)PUNPCKLBW(Y?)rr",
|
|
"(V?)PUNPCKLDQ(Y?)rr",
|
|
"(V?)PUNPCKLQDQ(Y?)rr",
|
|
"(V?)PUNPCKLWD(Y?)rr",
|
|
"(V?)SHUFPD(Y?)rri",
|
|
"(V?)SHUFPS(Y?)rri",
|
|
"(V?)UNPCKHPD(Y?)rr",
|
|
"(V?)UNPCKHPS(Y?)rr",
|
|
"(V?)UNPCKLPD(Y?)rr",
|
|
"(V?)UNPCKLPS(Y?)rr",
|
|
"(V?)XORPD(Y?)rr",
|
|
"(V?)XORPS(Y?)rr")>;
|
|
|
|
def BWWriteResGroup4 : SchedWriteRes<[BWPort6]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup4], (instregex "JMP(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup5 : SchedWriteRes<[BWPort01]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup5], (instregex "FINCSTP",
|
|
"FNOP")>;
|
|
|
|
def BWWriteResGroup6 : SchedWriteRes<[BWPort06]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup6], (instrs CDQ, CQO)>;
|
|
def: InstRW<[BWWriteResGroup6], (instregex "ADC(16|32|64)ri",
|
|
"ADC(16|32|64)i",
|
|
"ADC(8|16|32|64)rr",
|
|
"ADCX(32|64)rr",
|
|
"ADOX(32|64)rr",
|
|
"BT(16|32|64)ri8",
|
|
"BT(16|32|64)rr",
|
|
"BTC(16|32|64)ri8",
|
|
"BTC(16|32|64)rr",
|
|
"BTR(16|32|64)ri8",
|
|
"BTR(16|32|64)rr",
|
|
"BTS(16|32|64)ri8",
|
|
"BTS(16|32|64)rr",
|
|
"CMOV(AE|B|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)(16|32|64)rr",
|
|
"J(A|AE|B|BE|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)_1",
|
|
"J(A|AE|B|BE|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)_4",
|
|
"JMP_1",
|
|
"JMP_4",
|
|
"RORX(32|64)ri",
|
|
"SAR(8|16|32|64)r1",
|
|
"SAR(8|16|32|64)ri",
|
|
"SARX(32|64)rr",
|
|
"SBB(16|32|64)ri",
|
|
"SBB(16|32|64)i",
|
|
"SBB(8|16|32|64)rr",
|
|
"SET(AE|B|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)r",
|
|
"SHL(8|16|32|64)r1",
|
|
"SHL(8|16|32|64)ri",
|
|
"SHLX(32|64)rr",
|
|
"SHR(8|16|32|64)r1",
|
|
"SHR(8|16|32|64)ri",
|
|
"SHRX(32|64)rr")>;
|
|
|
|
def BWWriteResGroup7 : SchedWriteRes<[BWPort15]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup7], (instregex "ANDN(32|64)rr",
|
|
"BLSI(32|64)rr",
|
|
"BLSMSK(32|64)rr",
|
|
"BLSR(32|64)rr",
|
|
"LEA(16|32|64)(_32)?r",
|
|
"MMX_PABSBrr",
|
|
"MMX_PABSDrr",
|
|
"MMX_PABSWrr",
|
|
"MMX_PADDBirr",
|
|
"MMX_PADDDirr",
|
|
"MMX_PADDQirr",
|
|
"MMX_PADDSBirr",
|
|
"MMX_PADDSWirr",
|
|
"MMX_PADDUSBirr",
|
|
"MMX_PADDUSWirr",
|
|
"MMX_PADDWirr",
|
|
"MMX_PAVGBirr",
|
|
"MMX_PAVGWirr",
|
|
"MMX_PCMPEQBirr",
|
|
"MMX_PCMPEQDirr",
|
|
"MMX_PCMPEQWirr",
|
|
"MMX_PCMPGTBirr",
|
|
"MMX_PCMPGTDirr",
|
|
"MMX_PCMPGTWirr",
|
|
"MMX_PMAXSWirr",
|
|
"MMX_PMAXUBirr",
|
|
"MMX_PMINSWirr",
|
|
"MMX_PMINUBirr",
|
|
"MMX_PSIGNBrr",
|
|
"MMX_PSIGNDrr",
|
|
"MMX_PSIGNWrr",
|
|
"MMX_PSUBBirr",
|
|
"MMX_PSUBDirr",
|
|
"MMX_PSUBQirr",
|
|
"MMX_PSUBSBirr",
|
|
"MMX_PSUBSWirr",
|
|
"MMX_PSUBUSBirr",
|
|
"MMX_PSUBUSWirr",
|
|
"MMX_PSUBWirr",
|
|
"(V?)PABSB(Y?)rr",
|
|
"(V?)PABSD(Y?)rr",
|
|
"(V?)PABSW(Y?)rr",
|
|
"(V?)PADDB(Y?)rr",
|
|
"(V?)PADDD(Y?)rr",
|
|
"(V?)PADDQ(Y?)rr",
|
|
"(V?)PADDSB(Y?)rr",
|
|
"(V?)PADDSW(Y?)rr",
|
|
"(V?)PADDUSB(Y?)rr",
|
|
"(V?)PADDUSW(Y?)rr",
|
|
"(V?)PADDW(Y?)rr",
|
|
"(V?)PAVGB(Y?)rr",
|
|
"(V?)PAVGW(Y?)rr",
|
|
"(V?)PCMPEQB(Y?)rr",
|
|
"(V?)PCMPEQD(Y?)rr",
|
|
"(V?)PCMPEQQ(Y?)rr",
|
|
"(V?)PCMPEQW(Y?)rr",
|
|
"(V?)PCMPGTB(Y?)rr",
|
|
"(V?)PCMPGTD(Y?)rr",
|
|
"(V?)PCMPGTW(Y?)rr",
|
|
"(V?)PMAXSB(Y?)rr",
|
|
"(V?)PMAXSD(Y?)rr",
|
|
"(V?)PMAXSW(Y?)rr",
|
|
"(V?)PMAXUB(Y?)rr",
|
|
"(V?)PMAXUD(Y?)rr",
|
|
"(V?)PMAXUW(Y?)rr",
|
|
"(V?)PMINSB(Y?)rr",
|
|
"(V?)PMINSD(Y?)rr",
|
|
"(V?)PMINSW(Y?)rr",
|
|
"(V?)PMINUB(Y?)rr",
|
|
"(V?)PMINUD(Y?)rr",
|
|
"(V?)PMINUW(Y?)rr",
|
|
"(V?)PSIGNB(Y?)rr",
|
|
"(V?)PSIGND(Y?)rr",
|
|
"(V?)PSIGNW(Y?)rr",
|
|
"(V?)PSUBB(Y?)rr",
|
|
"(V?)PSUBD(Y?)rr",
|
|
"(V?)PSUBQ(Y?)rr",
|
|
"(V?)PSUBSB(Y?)rr",
|
|
"(V?)PSUBSW(Y?)rr",
|
|
"(V?)PSUBUSB(Y?)rr",
|
|
"(V?)PSUBUSW(Y?)rr",
|
|
"(V?)PSUBW(Y?)rr")>;
|
|
|
|
def BWWriteResGroup8 : SchedWriteRes<[BWPort015]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup8], (instregex "MMX_MOVQ64rr",
|
|
"MMX_PANDNirr",
|
|
"MMX_PANDirr",
|
|
"MMX_PORirr",
|
|
"MMX_PXORirr",
|
|
"(V?)BLENDPD(Y?)rri",
|
|
"(V?)BLENDPS(Y?)rri",
|
|
"(V?)MOVDQA(Y?)rr",
|
|
"(V?)MOVDQU(Y?)rr",
|
|
"(V?)MOVPQI2QIrr",
|
|
"VMOVZPQILo2PQIrr",
|
|
"(V?)PANDN(Y?)rr",
|
|
"(V?)PAND(Y?)rr",
|
|
"VPBLENDD(Y?)rri",
|
|
"(V?)POR(Y?)rr",
|
|
"(V?)PXOR(Y?)rr")>;
|
|
|
|
def BWWriteResGroup9 : SchedWriteRes<[BWPort0156]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup9], (instrs CBW, CWDE, CDQE)>;
|
|
def: InstRW<[BWWriteResGroup9], (instregex "CLC",
|
|
"CMC",
|
|
"LAHF",
|
|
"NOOP",
|
|
"SAHF",
|
|
"SGDT64m",
|
|
"SIDT64m",
|
|
"SLDT64m",
|
|
"SMSW16m",
|
|
"STC",
|
|
"STRm",
|
|
"SYSCALL",
|
|
"XCHG(16|32|64)rr")>;
|
|
|
|
def BWWriteResGroup10 : SchedWriteRes<[BWPort4,BWPort237]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup10], (instregex "FBSTPm",
|
|
"MMX_MOVD64from64rm",
|
|
"MMX_MOVD64mr",
|
|
"MMX_MOVNTQmr",
|
|
"MMX_MOVQ64mr",
|
|
"MOVNTI_64mr",
|
|
"MOVNTImr",
|
|
"ST_FP32m",
|
|
"ST_FP64m",
|
|
"ST_FP80m",
|
|
"VEXTRACTF128mr",
|
|
"VEXTRACTI128mr",
|
|
"(V?)MOVAPD(Y?)mr",
|
|
"(V?)MOVAPS(Y?)mr",
|
|
"(V?)MOVDQA(Y?)mr",
|
|
"(V?)MOVDQU(Y?)mr",
|
|
"(V?)MOVHPDmr",
|
|
"(V?)MOVHPSmr",
|
|
"(V?)MOVLPDmr",
|
|
"(V?)MOVLPSmr",
|
|
"(V?)MOVNTDQ(V?)mr",
|
|
"(V?)MOVNTPD(V?)mr",
|
|
"(V?)MOVNTPS(V?)mr",
|
|
"(V?)MOVPDI2DImr",
|
|
"(V?)MOVPQI2QImr",
|
|
"(V?)MOVPQIto64mr",
|
|
"(V?)MOVSDmr",
|
|
"(V?)MOVSSmr",
|
|
"(V?)MOVUPD(Y?)mr",
|
|
"(V?)MOVUPS(Y?)mr")>;
|
|
|
|
def BWWriteResGroup11 : SchedWriteRes<[BWPort5]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup11], (instregex "BLENDVPDrr0",
|
|
"BLENDVPSrr0",
|
|
"MMX_PINSRWrr",
|
|
"PBLENDVBrr0",
|
|
"VBLENDVPD(Y?)rr",
|
|
"VBLENDVPS(Y?)rr",
|
|
"VPBLENDVB(Y?)rr",
|
|
"(V?)PINSRBrr",
|
|
"(V?)PINSRDrr",
|
|
"(V?)PINSRQrr",
|
|
"(V?)PINSRWrr")>;
|
|
|
|
def BWWriteResGroup12 : SchedWriteRes<[BWPort01]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup12], (instregex "FDECSTP")>;
|
|
|
|
def BWWriteResGroup13 : SchedWriteRes<[BWPort06]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup13], (instregex "ROL(8|16|32|64)r1",
|
|
"ROL(8|16|32|64)ri",
|
|
"ROR(8|16|32|64)r1",
|
|
"ROR(8|16|32|64)ri")>;
|
|
|
|
def BWWriteResGroup14 : SchedWriteRes<[BWPort0156]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup14], (instregex "LFENCE",
|
|
"MFENCE",
|
|
"WAIT",
|
|
"XGETBV")>;
|
|
|
|
def BWWriteResGroup15 : SchedWriteRes<[BWPort0,BWPort5]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup15], (instregex "MMX_PEXTRWrr",
|
|
"VCVTPH2PS(Y?)rr",
|
|
"(V?)CVTPS2PDrr",
|
|
"(V?)CVTSS2SDrr",
|
|
"(V?)EXTRACTPSrr",
|
|
"(V?)PEXTRBrr",
|
|
"(V?)PEXTRDrr",
|
|
"(V?)PEXTRQrr",
|
|
"(V?)PEXTRWrr",
|
|
"(V?)PSLLDrr",
|
|
"(V?)PSLLQrr",
|
|
"(V?)PSLLWrr",
|
|
"(V?)PSRADrr",
|
|
"(V?)PSRAWrr",
|
|
"(V?)PSRLDrr",
|
|
"(V?)PSRLQrr",
|
|
"(V?)PSRLWrr",
|
|
"(V?)PTESTrr")>;
|
|
|
|
def BWWriteResGroup16 : SchedWriteRes<[BWPort6,BWPort0156]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup16], (instregex "CLFLUSH")>;
|
|
|
|
def BWWriteResGroup17 : SchedWriteRes<[BWPort01,BWPort015]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup17], (instregex "MMX_MOVDQ2Qrr")>;
|
|
|
|
def BWWriteResGroup18 : SchedWriteRes<[BWPort237,BWPort0156]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup18], (instregex "SFENCE")>;
|
|
|
|
def BWWriteResGroup19 : SchedWriteRes<[BWPort06,BWPort15]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup19], (instrs BSWAP64r)>;
|
|
|
|
def BWWriteResGroup19_1 : SchedWriteRes<[BWPort15]> {
|
|
let Latency = 1;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup19_1], (instrs BSWAP32r)>;
|
|
|
|
def BWWriteResGroup20 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup20], (instrs CWD)>;
|
|
def: InstRW<[BWWriteResGroup20], (instrs JCXZ, JECXZ, JRCXZ)>;
|
|
def: InstRW<[BWWriteResGroup20], (instregex "ADC8i8",
|
|
"ADC8ri",
|
|
"CMOV(A|BE)(16|32|64)rr",
|
|
"SBB8i8",
|
|
"SBB8ri",
|
|
"SET(A|BE)r")>;
|
|
|
|
def BWWriteResGroup21 : SchedWriteRes<[BWPort4,BWPort5,BWPort237]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup21], (instregex "(V?)EXTRACTPSmr",
|
|
"(V?)PEXTRBmr",
|
|
"(V?)PEXTRDmr",
|
|
"(V?)PEXTRQmr",
|
|
"(V?)PEXTRWmr",
|
|
"(V?)STMXCSR")>;
|
|
|
|
def BWWriteResGroup22 : SchedWriteRes<[BWPort4,BWPort6,BWPort237]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup22], (instregex "FNSTCW16m")>;
|
|
|
|
def BWWriteResGroup23 : SchedWriteRes<[BWPort4,BWPort237,BWPort06]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup23], (instregex "SET(AE|B|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)m")>;
|
|
|
|
def BWWriteResGroup24 : SchedWriteRes<[BWPort4,BWPort237,BWPort15]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup24], (instregex "MOVBE(16|32|64)mr")>;
|
|
|
|
def BWWriteResGroup25 : SchedWriteRes<[BWPort4,BWPort237,BWPort0156]> {
|
|
let Latency = 2;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup25], (instrs PUSH16r, PUSH32r, PUSH64r)>;
|
|
def: InstRW<[BWWriteResGroup25], (instregex "PUSH(16|32|64)rmr",
|
|
"PUSH64i8",
|
|
"STOSB",
|
|
"STOSL",
|
|
"STOSQ",
|
|
"STOSW")>;
|
|
|
|
def BWWriteResGroup27 : SchedWriteRes<[BWPort1]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup27], (instrs IMUL16rr, IMUL32rr, IMUL32rri, IMUL32rri8, IMUL64rr, IMUL64rri32, IMUL64rri8)>;
|
|
def: InstRW<[BWWriteResGroup27], (instrs IMUL8r, MUL8r)>;
|
|
def: InstRW<[BWWriteResGroup27], (instregex "ADD_FPrST0",
|
|
"ADD_FST0r",
|
|
"ADD_FrST0",
|
|
"MMX_CVTPI2PSirr",
|
|
"PDEP(32|64)rr",
|
|
"PEXT(32|64)rr",
|
|
"SHLD(16|32|64)rri8",
|
|
"SHRD(16|32|64)rri8",
|
|
"SUBR_FPrST0",
|
|
"SUBR_FST0r",
|
|
"SUBR_FrST0",
|
|
"SUB_FPrST0",
|
|
"SUB_FST0r",
|
|
"SUB_FrST0",
|
|
"(V?)ADDPD(Y?)rr",
|
|
"(V?)ADDPS(Y?)rr",
|
|
"(V?)ADDSDrr",
|
|
"(V?)ADDSSrr",
|
|
"(V?)ADDSUBPD(Y?)rr",
|
|
"(V?)ADDSUBPS(Y?)rr",
|
|
"(V?)CMPPD(Y?)rri",
|
|
"(V?)CMPPS(Y?)rri",
|
|
"(V?)CMPSDrr",
|
|
"(V?)CMPSSrr",
|
|
"(V?)COMISDrr",
|
|
"(V?)COMISSrr",
|
|
"(V?)CVTDQ2PS(Y?)rr",
|
|
"(V?)CVTPS2DQ(Y?)rr",
|
|
"(V?)CVTTPS2DQ(Y?)rr",
|
|
"(V?)MAX(C?)PD(Y?)rr",
|
|
"(V?)MAX(C?)PS(Y?)rr",
|
|
"(V?)MAX(C?)SDrr",
|
|
"(V?)MAX(C?)SSrr",
|
|
"(V?)MIN(C?)PD(Y?)rr",
|
|
"(V?)MIN(C?)PS(Y?)rr",
|
|
"(V?)MIN(C?)SDrr",
|
|
"(V?)MIN(C?)SSrr",
|
|
"(V?)SUBPD(Y?)rr",
|
|
"(V?)SUBPS(Y?)rr",
|
|
"(V?)SUBSDrr",
|
|
"(V?)SUBSSrr",
|
|
"(V?)UCOMISDrr",
|
|
"(V?)UCOMISSrr")>;
|
|
|
|
def BWWriteResGroup27_16 : SchedWriteRes<[BWPort1, BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup27_16], (instrs IMUL16rri, IMUL16rri8)>;
|
|
|
|
def BWWriteResGroup28 : SchedWriteRes<[BWPort5]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup28], (instregex "VBROADCASTSDYrr",
|
|
"VBROADCASTSSYrr",
|
|
"VEXTRACTF128rr",
|
|
"VEXTRACTI128rr",
|
|
"VINSERTF128rr",
|
|
"VINSERTI128rr",
|
|
"VPBROADCASTB(Y?)rr",
|
|
"VPBROADCASTDYrr",
|
|
"VPBROADCASTQYrr",
|
|
"VPBROADCASTW(Y?)rr",
|
|
"VPERM2F128rr",
|
|
"VPERM2I128rr",
|
|
"VPERMDYrr",
|
|
"VPERMPDYri",
|
|
"VPERMPSYrr",
|
|
"VPERMQYri",
|
|
"VPMOVSXBDYrr",
|
|
"VPMOVSXBQYrr",
|
|
"VPMOVSXBWYrr",
|
|
"VPMOVSXDQYrr",
|
|
"VPMOVSXWDYrr",
|
|
"VPMOVSXWQYrr",
|
|
"VPMOVZXBDYrr",
|
|
"VPMOVZXBQYrr",
|
|
"VPMOVZXBWYrr",
|
|
"VPMOVZXDQYrr",
|
|
"VPMOVZXWDYrr",
|
|
"VPMOVZXWQYrr")>;
|
|
|
|
def BWWriteResGroup29 : SchedWriteRes<[BWPort01]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup29], (instregex "(V?)MULPD(Y?)rr",
|
|
"(V?)MULPS(Y?)rr",
|
|
"(V?)MULSDrr",
|
|
"(V?)MULSSrr")>;
|
|
|
|
def BWWriteResGroup30 : SchedWriteRes<[BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup30], (instregex "XADD(8|16|32|64)rr",
|
|
"XCHG8rr")>;
|
|
|
|
def BWWriteResGroup31 : SchedWriteRes<[BWPort0,BWPort5]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup31], (instregex "VPSLLVD(Y?)rr",
|
|
"VPSRAVD(Y?)rr",
|
|
"VPSRLVD(Y?)rr")>;
|
|
|
|
def BWWriteResGroup32 : SchedWriteRes<[BWPort5,BWPort15]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup32], (instregex "MMX_PHADDDrr",
|
|
"MMX_PHADDSWrr",
|
|
"MMX_PHADDWrr",
|
|
"MMX_PHSUBDrr",
|
|
"MMX_PHSUBSWrr",
|
|
"MMX_PHSUBWrr",
|
|
"(V?)PHADDD(Y?)rr",
|
|
"(V?)PHADDSW(Y?)rr",
|
|
"(V?)PHADDW(Y?)rr",
|
|
"(V?)PHSUBD(Y?)rr",
|
|
"(V?)PHSUBSW(Y?)rr",
|
|
"(V?)PHSUBW(Y?)rr")>;
|
|
|
|
def BWWriteResGroup33 : SchedWriteRes<[BWPort5,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup33], (instregex "MMX_PACKSSDWirr",
|
|
"MMX_PACKSSWBirr",
|
|
"MMX_PACKUSWBirr")>;
|
|
|
|
def BWWriteResGroup34 : SchedWriteRes<[BWPort6,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup34], (instregex "CLD")>;
|
|
|
|
def BWWriteResGroup35 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup35], (instregex "RCL(8|16|32|64)r1",
|
|
"RCL(8|16|32|64)ri",
|
|
"RCR(8|16|32|64)r1",
|
|
"RCR(8|16|32|64)ri")>;
|
|
|
|
def BWWriteResGroup36 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup36], (instregex "ROL(8|16|32|64)rCL",
|
|
"ROR(8|16|32|64)rCL",
|
|
"SAR(8|16|32|64)rCL",
|
|
"SHL(8|16|32|64)rCL",
|
|
"SHR(8|16|32|64)rCL")>;
|
|
|
|
def BWWriteResGroup37 : SchedWriteRes<[BWPort4,BWPort6,BWPort237,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup37], (instregex "CALL(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup38 : SchedWriteRes<[BWPort4,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 3;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup38], (instregex "CALL64pcrel32",
|
|
"SET(A|BE)m")>;
|
|
|
|
def BWWriteResGroup39 : SchedWriteRes<[BWPort0,BWPort1]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup39], (instregex "(V?)CVTSD2SI64rr",
|
|
"(V?)CVTSD2SIrr",
|
|
"(V?)CVTSS2SI64rr",
|
|
"(V?)CVTSS2SIrr",
|
|
"(V?)CVTTSD2SI64rr",
|
|
"(V?)CVTTSD2SIrr",
|
|
"(V?)CVTTSS2SI64rr",
|
|
"(V?)CVTTSS2SIrr")>;
|
|
|
|
def BWWriteResGroup40 : SchedWriteRes<[BWPort0,BWPort5]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup40], (instregex "VCVTPS2PDYrr",
|
|
"VPSLLDYrr",
|
|
"VPSLLQYrr",
|
|
"VPSLLWYrr",
|
|
"VPSRADYrr",
|
|
"VPSRAWYrr",
|
|
"VPSRLDYrr",
|
|
"VPSRLQYrr",
|
|
"VPSRLWYrr",
|
|
"VPTESTYrr")>;
|
|
|
|
def BWWriteResGroup41 : SchedWriteRes<[BWPort0,BWPort0156]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup41], (instregex "FNSTSW16r")>;
|
|
|
|
def BWWriteResGroup42 : SchedWriteRes<[BWPort1,BWPort5]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup42], (instrs IMUL64r, MUL64r, MULX64rr)>;
|
|
def: InstRW<[BWWriteResGroup42], (instregex "MMX_CVTPD2PIirr",
|
|
"MMX_CVTPI2PDirr",
|
|
"MMX_CVTPS2PIirr",
|
|
"MMX_CVTTPD2PIirr",
|
|
"MMX_CVTTPS2PIirr",
|
|
"(V?)CVTDQ2PDrr",
|
|
"(V?)CVTPD2DQrr",
|
|
"(V?)CVTPD2PSrr",
|
|
"VCVTPS2PHrr",
|
|
"(V?)CVTSD2SSrr",
|
|
"(V?)CVTSI642SDrr",
|
|
"(V?)CVTSI2SDrr",
|
|
"(V?)CVTSI2SSrr",
|
|
"(V?)CVTTPD2DQrr")>;
|
|
|
|
def BWWriteResGroup42_16 : SchedWriteRes<[BWPort1,BWPort06,BWPort0156]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 4;
|
|
}
|
|
def: InstRW<[BWWriteResGroup42_16], (instrs IMUL16r, MUL16r)>;
|
|
|
|
def BWWriteResGroup43 : SchedWriteRes<[BWPort0,BWPort4,BWPort237]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup43], (instregex "FNSTSWm")>;
|
|
|
|
def BWWriteResGroup44 : SchedWriteRes<[BWPort1,BWPort4,BWPort237]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup44], (instregex "ISTT_FP16m",
|
|
"ISTT_FP32m",
|
|
"ISTT_FP64m",
|
|
"IST_F16m",
|
|
"IST_F32m",
|
|
"IST_FP16m",
|
|
"IST_FP32m",
|
|
"IST_FP64m",
|
|
"VCVTPS2PH(Y?)mr")>;
|
|
|
|
def BWWriteResGroup45 : SchedWriteRes<[BWPort0156]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup45], (instregex "FNCLEX")>;
|
|
|
|
def BWWriteResGroup46 : SchedWriteRes<[BWPort015,BWPort0156]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup46], (instregex "VZEROUPPER")>;
|
|
|
|
def BWWriteResGroup47 : SchedWriteRes<[BWPort0]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup47], (instregex "MMX_PMADDUBSWrr",
|
|
"MMX_PMADDWDirr",
|
|
"MMX_PMULHRSWrr",
|
|
"MMX_PMULHUWirr",
|
|
"MMX_PMULHWirr",
|
|
"MMX_PMULLWirr",
|
|
"MMX_PMULUDQirr",
|
|
"MMX_PSADBWirr",
|
|
"MUL_FPrST0",
|
|
"MUL_FST0r",
|
|
"MUL_FrST0",
|
|
"(V?)PCMPGTQ(Y?)rr",
|
|
"(V?)PHMINPOSUWrr",
|
|
"(V?)PMADDUBSW(Y?)rr",
|
|
"(V?)PMADDWD(Y?)rr",
|
|
"(V?)PMULDQ(Y?)rr",
|
|
"(V?)PMULHRSW(Y?)rr",
|
|
"(V?)PMULHUW(Y?)rr",
|
|
"(V?)PMULHW(Y?)rr",
|
|
"(V?)PMULLW(Y?)rr",
|
|
"(V?)PMULUDQ(Y?)rr",
|
|
"(V?)PSADBW(Y?)rr",
|
|
"(V?)RCPPSr",
|
|
"(V?)RCPSSr",
|
|
"(V?)RSQRTPSr",
|
|
"(V?)RSQRTSSr")>;
|
|
|
|
def BWWriteResGroup48 : SchedWriteRes<[BWPort01]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup48],
|
|
(instregex "VF(N)?M(ADD|SUB|ADDSUB|SUBADD)(132|213|231)P(D|S)(Y)?r",
|
|
"VF(N)?M(ADD|SUB)(132|213|231)S(D|S)r")>;
|
|
|
|
def BWWriteResGroup49 : SchedWriteRes<[BWPort23]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup49], (instregex "MMX_MOVD64rm",
|
|
"MMX_MOVD64to64rm",
|
|
"MMX_MOVQ64rm",
|
|
"MOV(16|32|64)rm",
|
|
"MOVSX(16|32|64)rm16",
|
|
"MOVSX(16|32|64)rm32",
|
|
"MOVSX(16|32|64)rm8",
|
|
"MOVZX(16|32|64)rm16",
|
|
"MOVZX(16|32|64)rm8",
|
|
"PREFETCHNTA",
|
|
"PREFETCHT0",
|
|
"PREFETCHT1",
|
|
"PREFETCHT2",
|
|
"VBROADCASTSSrm",
|
|
"(V?)LDDQUrm",
|
|
"(V?)MOV64toPQIrm",
|
|
"(V?)MOVAPDrm",
|
|
"(V?)MOVAPSrm",
|
|
"(V?)MOVDDUPrm",
|
|
"(V?)MOVDI2PDIrm",
|
|
"(V?)MOVDQArm",
|
|
"(V?)MOVDQUrm",
|
|
"(V?)MOVNTDQArm",
|
|
"(V?)MOVQI2PQIrm",
|
|
"(V?)MOVSDrm",
|
|
"(V?)MOVSHDUPrm",
|
|
"(V?)MOVSLDUPrm",
|
|
"(V?)MOVSSrm",
|
|
"(V?)MOVUPDrm",
|
|
"(V?)MOVUPSrm",
|
|
"VPBROADCASTDrm",
|
|
"VPBROADCASTQrm")>;
|
|
|
|
def BWWriteResGroup50 : SchedWriteRes<[BWPort1,BWPort5]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup50], (instregex "(V?)CVTSI642SSrr",
|
|
"(V?)HADDPD(Y?)rr",
|
|
"(V?)HADDPS(Y?)rr",
|
|
"(V?)HSUBPD(Y?)rr",
|
|
"(V?)HSUBPS(Y?)rr")>;
|
|
|
|
def BWWriteResGroup51 : SchedWriteRes<[BWPort1,BWPort6,BWPort06]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup51], (instregex "STR(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup52 : SchedWriteRes<[BWPort1,BWPort06,BWPort0156]> {
|
|
let Latency = 4;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup52], (instrs IMUL32r, MUL32r, MULX32rr)>;
|
|
|
|
def BWWriteResGroup53 : SchedWriteRes<[BWPort0,BWPort4,BWPort237,BWPort15]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup53], (instregex "VMASKMOVPD(Y?)mr",
|
|
"VMASKMOVPS(Y?)mr",
|
|
"VPMASKMOVD(Y?)mr",
|
|
"VPMASKMOVQ(Y?)mr")>;
|
|
|
|
def BWWriteResGroup54 : SchedWriteRes<[BWPort6,BWPort0156]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup54], (instregex "PAUSE")>;
|
|
|
|
def BWWriteResGroup55 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup55], (instregex "XSETBV")>;
|
|
|
|
def BWWriteResGroup56 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [2,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup56], (instregex "CMPXCHG(8|16|32|64)rr")>;
|
|
|
|
def BWWriteResGroup57 : SchedWriteRes<[BWPort4,BWPort237,BWPort0156]> {
|
|
let Latency = 5;
|
|
let NumMicroOps = 6;
|
|
let ResourceCycles = [1,1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup57], (instregex "PUSHF16", "PUSHF64")>;
|
|
|
|
def BWWriteResGroup58 : SchedWriteRes<[BWPort23]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup58], (instregex "LD_F32m",
|
|
"LD_F64m",
|
|
"LD_F80m",
|
|
"VBROADCASTF128",
|
|
"VBROADCASTI128",
|
|
"VBROADCASTSDYrm",
|
|
"VBROADCASTSSYrm",
|
|
"VLDDQUYrm",
|
|
"VMOVAPDYrm",
|
|
"VMOVAPSYrm",
|
|
"VMOVDDUPYrm",
|
|
"VMOVDQAYrm",
|
|
"VMOVDQUYrm",
|
|
"VMOVNTDQAYrm",
|
|
"VMOVSHDUPYrm",
|
|
"VMOVSLDUPYrm",
|
|
"VMOVUPDYrm",
|
|
"VMOVUPSYrm",
|
|
"VPBROADCASTDYrm",
|
|
"VPBROADCASTQYrm",
|
|
"(V?)ROUNDPD(Y?)r",
|
|
"(V?)ROUNDPS(Y?)r",
|
|
"(V?)ROUNDSDr",
|
|
"(V?)ROUNDSSr")>;
|
|
|
|
def BWWriteResGroup59 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup59], (instregex "MMX_PSLLDrm",
|
|
"MMX_PSLLQrm",
|
|
"MMX_PSLLWrm",
|
|
"MMX_PSRADrm",
|
|
"MMX_PSRAWrm",
|
|
"MMX_PSRLDrm",
|
|
"MMX_PSRLQrm",
|
|
"MMX_PSRLWrm",
|
|
"VCVTPH2PS(Y?)rm",
|
|
"(V?)CVTPS2PDrm",
|
|
"(V?)CVTSS2SDrm",
|
|
"VPSLLVQrm",
|
|
"VPSRLVQrm",
|
|
"VTESTPDrm",
|
|
"VTESTPSrm")>;
|
|
|
|
def BWWriteResGroup60 : SchedWriteRes<[BWPort1,BWPort5]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup60], (instregex "VCVTDQ2PDYrr",
|
|
"VCVTPD2DQYrr",
|
|
"VCVTPD2PSYrr",
|
|
"VCVTPS2PHYrr",
|
|
"VCVTTPD2DQYrr")>;
|
|
|
|
def BWWriteResGroup61 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup61], (instregex "MMX_PALIGNRrmi",
|
|
"MMX_PINSRWrm",
|
|
"MMX_PSHUFBrm",
|
|
"MMX_PSHUFWmi",
|
|
"MMX_PUNPCKHBWirm",
|
|
"MMX_PUNPCKHDQirm",
|
|
"MMX_PUNPCKHWDirm",
|
|
"MMX_PUNPCKLBWirm",
|
|
"MMX_PUNPCKLDQirm",
|
|
"MMX_PUNPCKLWDirm",
|
|
"(V?)ANDNPDrm",
|
|
"(V?)ANDNPSrm",
|
|
"(V?)ANDPDrm",
|
|
"(V?)ANDPSrm",
|
|
"(V?)INSERTPSrm",
|
|
"(V?)MOVHPDrm",
|
|
"(V?)MOVHPSrm",
|
|
"(V?)MOVLPDrm",
|
|
"(V?)MOVLPSrm",
|
|
"(V?)ORPDrm",
|
|
"(V?)ORPSrm",
|
|
"(V?)PACKSSDWrm",
|
|
"(V?)PACKSSWBrm",
|
|
"(V?)PACKUSDWrm",
|
|
"(V?)PACKUSWBrm",
|
|
"(V?)PALIGNRrmi",
|
|
"(V?)PBLENDWrmi",
|
|
"VPERMILPDmi",
|
|
"VPERMILPDrm",
|
|
"VPERMILPSmi",
|
|
"VPERMILPSrm",
|
|
"(V?)PINSRBrm",
|
|
"(V?)PINSRDrm",
|
|
"(V?)PINSRQrm",
|
|
"(V?)PINSRWrm",
|
|
"(V?)PMOVSXBDrm",
|
|
"(V?)PMOVSXBQrm",
|
|
"(V?)PMOVSXBWrm",
|
|
"(V?)PMOVSXDQrm",
|
|
"(V?)PMOVSXWDrm",
|
|
"(V?)PMOVSXWQrm",
|
|
"(V?)PMOVZXBDrm",
|
|
"(V?)PMOVZXBQrm",
|
|
"(V?)PMOVZXBWrm",
|
|
"(V?)PMOVZXDQrm",
|
|
"(V?)PMOVZXWDrm",
|
|
"(V?)PMOVZXWQrm",
|
|
"(V?)PSHUFBrm",
|
|
"(V?)PSHUFDmi",
|
|
"(V?)PSHUFHWmi",
|
|
"(V?)PSHUFLWmi",
|
|
"(V?)PUNPCKHBWrm",
|
|
"(V?)PUNPCKHDQrm",
|
|
"(V?)PUNPCKHQDQrm",
|
|
"(V?)PUNPCKHWDrm",
|
|
"(V?)PUNPCKLBWrm",
|
|
"(V?)PUNPCKLDQrm",
|
|
"(V?)PUNPCKLQDQrm",
|
|
"(V?)PUNPCKLWDrm",
|
|
"(V?)SHUFPDrmi",
|
|
"(V?)SHUFPSrmi",
|
|
"(V?)UNPCKHPDrm",
|
|
"(V?)UNPCKHPSrm",
|
|
"(V?)UNPCKLPDrm",
|
|
"(V?)UNPCKLPSrm",
|
|
"(V?)XORPDrm",
|
|
"(V?)XORPSrm")>;
|
|
|
|
def BWWriteResGroup62 : SchedWriteRes<[BWPort6,BWPort23]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup62], (instregex "FARJMP64",
|
|
"JMP(16|32|64)m")>;
|
|
|
|
def BWWriteResGroup63 : SchedWriteRes<[BWPort23,BWPort06]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup63], (instregex "BT(16|32|64)mi8",
|
|
"CMOV(AE|B|E|G|GE|L|LE|NE|NO|NP|NS|O|P|S)(16|32|64)rm",
|
|
"RORX(32|64)mi",
|
|
"SARX(32|64)rm",
|
|
"SHLX(32|64)rm",
|
|
"SHRX(32|64)rm")>;
|
|
def: InstRW<[BWWriteResGroup63, ReadAfterLd], (instrs ADC8rm, ADC16rm, ADC32rm, ADC64rm,
|
|
ADCX32rm, ADCX64rm,
|
|
ADOX32rm, ADOX64rm,
|
|
SBB8rm, SBB16rm, SBB32rm, SBB64rm)>;
|
|
|
|
def BWWriteResGroup64 : SchedWriteRes<[BWPort23,BWPort15]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup64], (instregex "ANDN(32|64)rm",
|
|
"BLSI(32|64)rm",
|
|
"BLSMSK(32|64)rm",
|
|
"BLSR(32|64)rm",
|
|
"MMX_PABSBrm",
|
|
"MMX_PABSDrm",
|
|
"MMX_PABSWrm",
|
|
"MMX_PADDBirm",
|
|
"MMX_PADDDirm",
|
|
"MMX_PADDQirm",
|
|
"MMX_PADDSBirm",
|
|
"MMX_PADDSWirm",
|
|
"MMX_PADDUSBirm",
|
|
"MMX_PADDUSWirm",
|
|
"MMX_PADDWirm",
|
|
"MMX_PAVGBirm",
|
|
"MMX_PAVGWirm",
|
|
"MMX_PCMPEQBirm",
|
|
"MMX_PCMPEQDirm",
|
|
"MMX_PCMPEQWirm",
|
|
"MMX_PCMPGTBirm",
|
|
"MMX_PCMPGTDirm",
|
|
"MMX_PCMPGTWirm",
|
|
"MMX_PMAXSWirm",
|
|
"MMX_PMAXUBirm",
|
|
"MMX_PMINSWirm",
|
|
"MMX_PMINUBirm",
|
|
"MMX_PSIGNBrm",
|
|
"MMX_PSIGNDrm",
|
|
"MMX_PSIGNWrm",
|
|
"MMX_PSUBBirm",
|
|
"MMX_PSUBDirm",
|
|
"MMX_PSUBQirm",
|
|
"MMX_PSUBSBirm",
|
|
"MMX_PSUBSWirm",
|
|
"MMX_PSUBUSBirm",
|
|
"MMX_PSUBUSWirm",
|
|
"MMX_PSUBWirm",
|
|
"MOVBE(16|32|64)rm",
|
|
"(V?)PABSBrm",
|
|
"(V?)PABSDrm",
|
|
"(V?)PABSWrm",
|
|
"(V?)PADDBrm",
|
|
"(V?)PADDDrm",
|
|
"(V?)PADDQrm",
|
|
"(V?)PADDSBrm",
|
|
"(V?)PADDSWrm",
|
|
"(V?)PADDUSBrm",
|
|
"(V?)PADDUSWrm",
|
|
"(V?)PADDWrm",
|
|
"(V?)PAVGBrm",
|
|
"(V?)PAVGWrm",
|
|
"(V?)PCMPEQBrm",
|
|
"(V?)PCMPEQDrm",
|
|
"(V?)PCMPEQQrm",
|
|
"(V?)PCMPEQWrm",
|
|
"(V?)PCMPGTBrm",
|
|
"(V?)PCMPGTDrm",
|
|
"(V?)PCMPGTWrm",
|
|
"(V?)PMAXSBrm",
|
|
"(V?)PMAXSDrm",
|
|
"(V?)PMAXSWrm",
|
|
"(V?)PMAXUBrm",
|
|
"(V?)PMAXUDrm",
|
|
"(V?)PMAXUWrm",
|
|
"(V?)PMINSBrm",
|
|
"(V?)PMINSDrm",
|
|
"(V?)PMINSWrm",
|
|
"(V?)PMINUBrm",
|
|
"(V?)PMINUDrm",
|
|
"(V?)PMINUWrm",
|
|
"(V?)PSIGNBrm",
|
|
"(V?)PSIGNDrm",
|
|
"(V?)PSIGNWrm",
|
|
"(V?)PSUBBrm",
|
|
"(V?)PSUBDrm",
|
|
"(V?)PSUBQrm",
|
|
"(V?)PSUBSBrm",
|
|
"(V?)PSUBSWrm",
|
|
"(V?)PSUBUSBrm",
|
|
"(V?)PSUBUSWrm",
|
|
"(V?)PSUBWrm")>;
|
|
|
|
def BWWriteResGroup65 : SchedWriteRes<[BWPort23,BWPort015]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup65], (instregex "MMX_PANDNirm",
|
|
"MMX_PANDirm",
|
|
"MMX_PORirm",
|
|
"MMX_PXORirm",
|
|
"(V?)BLENDPDrmi",
|
|
"(V?)BLENDPSrmi",
|
|
"VINSERTF128rm",
|
|
"VINSERTI128rm",
|
|
"(V?)PANDNrm",
|
|
"(V?)PANDrm",
|
|
"VPBLENDDrmi",
|
|
"(V?)PORrm",
|
|
"(V?)PXORrm")>;
|
|
|
|
def BWWriteResGroup66 : SchedWriteRes<[BWPort23,BWPort0156]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup66], (instrs POP16r, POP32r, POP64r)>;
|
|
def: InstRW<[BWWriteResGroup66], (instregex "POP(16|32|64)rmr")>;
|
|
|
|
def BWWriteResGroup67 : SchedWriteRes<[BWPort1,BWPort06,BWPort0156]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup67], (instregex "SHLD(16|32|64)rrCL",
|
|
"SHRD(16|32|64)rrCL")>;
|
|
|
|
def BWWriteResGroup68 : SchedWriteRes<[BWPort1,BWPort6,BWPort06,BWPort0156]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup68], (instregex "SLDT(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup69 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort06]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup69], (instregex "BTC(16|32|64)mi8",
|
|
"BTR(16|32|64)mi8",
|
|
"BTS(16|32|64)mi8",
|
|
"SAR(8|16|32|64)m1",
|
|
"SAR(8|16|32|64)mi",
|
|
"SHL(8|16|32|64)m1",
|
|
"SHL(8|16|32|64)mi",
|
|
"SHR(8|16|32|64)m1",
|
|
"SHR(8|16|32|64)mi")>;
|
|
|
|
def BWWriteResGroup70 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup70], (instregex "POP(16|32|64)rmm",
|
|
"PUSH(16|32|64)rmm")>;
|
|
|
|
def BWWriteResGroup71 : SchedWriteRes<[BWPort6,BWPort0156]> {
|
|
let Latency = 6;
|
|
let NumMicroOps = 6;
|
|
let ResourceCycles = [1,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup71], (instregex "STD")>;
|
|
|
|
def BWWriteResGroup73 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup73], (instregex "VPSLLDYrm",
|
|
"VPSLLQYrm",
|
|
"VPSLLVQYrm",
|
|
"VPSLLWYrm",
|
|
"VPSRADYrm",
|
|
"VPSRAWYrm",
|
|
"VPSRLDYrm",
|
|
"VPSRLQYrm",
|
|
"VPSRLVQYrm",
|
|
"VPSRLWYrm",
|
|
"VTESTPDYrm",
|
|
"VTESTPSYrm")>;
|
|
|
|
def BWWriteResGroup74 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup74], (instregex "FCOM32m",
|
|
"FCOM64m",
|
|
"FCOMP32m",
|
|
"FCOMP64m")>;
|
|
|
|
def BWWriteResGroup75 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup75], (instregex "VANDNPDYrm",
|
|
"VANDNPSYrm",
|
|
"VANDPDYrm",
|
|
"VANDPSYrm",
|
|
"VORPDYrm",
|
|
"VORPSYrm",
|
|
"VPACKSSDWYrm",
|
|
"VPACKSSWBYrm",
|
|
"VPACKUSDWYrm",
|
|
"VPACKUSWBYrm",
|
|
"VPALIGNRYrmi",
|
|
"VPBLENDWYrmi",
|
|
"VPERMILPDYmi",
|
|
"VPERMILPDYrm",
|
|
"VPERMILPSYmi",
|
|
"VPERMILPSYrm",
|
|
"VPSHUFBYrm",
|
|
"VPSHUFDYmi",
|
|
"VPSHUFHWYmi",
|
|
"VPSHUFLWYmi",
|
|
"VPUNPCKHBWYrm",
|
|
"VPUNPCKHDQYrm",
|
|
"VPUNPCKHQDQYrm",
|
|
"VPUNPCKHWDYrm",
|
|
"VPUNPCKLBWYrm",
|
|
"VPUNPCKLDQYrm",
|
|
"VPUNPCKLQDQYrm",
|
|
"VPUNPCKLWDYrm",
|
|
"VSHUFPDYrmi",
|
|
"VSHUFPSYrmi",
|
|
"VUNPCKHPDYrm",
|
|
"VUNPCKHPSYrm",
|
|
"VUNPCKLPDYrm",
|
|
"VUNPCKLPSYrm",
|
|
"VXORPDYrm",
|
|
"VXORPSYrm")>;
|
|
|
|
def BWWriteResGroup76 : SchedWriteRes<[BWPort23,BWPort15]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup76], (instregex "VPABSBYrm",
|
|
"VPABSDYrm",
|
|
"VPABSWYrm",
|
|
"VPADDBYrm",
|
|
"VPADDDYrm",
|
|
"VPADDQYrm",
|
|
"VPADDSBYrm",
|
|
"VPADDSWYrm",
|
|
"VPADDUSBYrm",
|
|
"VPADDUSWYrm",
|
|
"VPADDWYrm",
|
|
"VPAVGBYrm",
|
|
"VPAVGWYrm",
|
|
"VPCMPEQBYrm",
|
|
"VPCMPEQDYrm",
|
|
"VPCMPEQQYrm",
|
|
"VPCMPEQWYrm",
|
|
"VPCMPGTBYrm",
|
|
"VPCMPGTDYrm",
|
|
"VPCMPGTWYrm",
|
|
"VPMAXSBYrm",
|
|
"VPMAXSDYrm",
|
|
"VPMAXSWYrm",
|
|
"VPMAXUBYrm",
|
|
"VPMAXUDYrm",
|
|
"VPMAXUWYrm",
|
|
"VPMINSBYrm",
|
|
"VPMINSDYrm",
|
|
"VPMINSWYrm",
|
|
"VPMINUBYrm",
|
|
"VPMINUDYrm",
|
|
"VPMINUWYrm",
|
|
"VPSIGNBYrm",
|
|
"VPSIGNDYrm",
|
|
"VPSIGNWYrm",
|
|
"VPSUBBYrm",
|
|
"VPSUBDYrm",
|
|
"VPSUBQYrm",
|
|
"VPSUBSBYrm",
|
|
"VPSUBSWYrm",
|
|
"VPSUBUSBYrm",
|
|
"VPSUBUSWYrm",
|
|
"VPSUBWYrm")>;
|
|
|
|
def BWWriteResGroup77 : SchedWriteRes<[BWPort23,BWPort015]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup77], (instregex "VBLENDPDYrmi",
|
|
"VBLENDPSYrmi",
|
|
"VPANDNYrm",
|
|
"VPANDYrm",
|
|
"VPBLENDDYrmi",
|
|
"VPORYrm",
|
|
"VPXORYrm")>;
|
|
|
|
def BWWriteResGroup78 : SchedWriteRes<[BWPort0,BWPort5]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup78], (instregex "(V?)MPSADBW(Y?)rri")>;
|
|
|
|
def BWWriteResGroup79 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup79], (instregex "BLENDVPDrm0",
|
|
"BLENDVPSrm0",
|
|
"MMX_PACKSSDWirm",
|
|
"MMX_PACKSSWBirm",
|
|
"MMX_PACKUSWBirm",
|
|
"PBLENDVBrm0",
|
|
"VBLENDVPDrm",
|
|
"VBLENDVPSrm",
|
|
"VMASKMOVPDrm",
|
|
"VMASKMOVPSrm",
|
|
"VPBLENDVBrm",
|
|
"VPMASKMOVDrm",
|
|
"VPMASKMOVQrm")>;
|
|
|
|
def BWWriteResGroup80 : SchedWriteRes<[BWPort23,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup80], (instrs LEAVE, LEAVE64,
|
|
SCASB, SCASL, SCASQ, SCASW)>;
|
|
|
|
def BWWriteResGroup81 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup81], (instregex "PSLLDrm",
|
|
"PSLLQrm",
|
|
"PSLLWrm",
|
|
"PSRADrm",
|
|
"PSRAWrm",
|
|
"PSRLDrm",
|
|
"PSRLQrm",
|
|
"PSRLWrm",
|
|
"PTESTrm",
|
|
"VPSLLDrm",
|
|
"VPSLLQrm",
|
|
"VPSLLWrm",
|
|
"VPSRADrm",
|
|
"VPSRAWrm",
|
|
"VPSRLDrm",
|
|
"VPSRLQrm",
|
|
"VPSRLWrm",
|
|
"VPTESTrm")>;
|
|
|
|
def BWWriteResGroup82 : SchedWriteRes<[BWPort0,BWPort01,BWPort23]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup82], (instregex "FLDCW16m")>;
|
|
|
|
def BWWriteResGroup83 : SchedWriteRes<[BWPort0,BWPort23,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup83], (instregex "(V?)LDMXCSR")>;
|
|
|
|
def BWWriteResGroup84 : SchedWriteRes<[BWPort6,BWPort23,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup84], (instregex "LRETQ",
|
|
"RETQ")>;
|
|
|
|
def BWWriteResGroup86 : SchedWriteRes<[BWPort23,BWPort06,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup86], (instregex "CMOV(A|BE)(16|32|64)rm")>;
|
|
|
|
def BWWriteResGroup87 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort06]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup87], (instregex "ROL(8|16|32|64)m1",
|
|
"ROL(8|16|32|64)mi",
|
|
"ROR(8|16|32|64)m1",
|
|
"ROR(8|16|32|64)mi")>;
|
|
|
|
def BWWriteResGroup88 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup88], (instregex "XADD(8|16|32|64)rm")>;
|
|
|
|
def BWWriteResGroup89 : SchedWriteRes<[BWPort4,BWPort6,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup89], (instregex "CALL(16|32|64)m",
|
|
"FARCALL64")>;
|
|
|
|
def BWWriteResGroup90 : SchedWriteRes<[BWPort6,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 7;
|
|
let NumMicroOps = 7;
|
|
let ResourceCycles = [2,2,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup90], (instrs LOOP)>;
|
|
|
|
def BWWriteResGroup91 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup91], (instrs IMUL32rm, IMUL32rmi, IMUL32rmi8, IMUL64rm, IMUL64rmi8, IMUL64rmi32)>;
|
|
def: InstRW<[BWWriteResGroup91], (instrs IMUL8m, MUL8m)>;
|
|
def: InstRW<[BWWriteResGroup91], (instregex "MMX_CVTPI2PSirm",
|
|
"MMX_CVTPS2PIirm",
|
|
"MMX_CVTTPS2PIirm",
|
|
"PDEP(32|64)rm",
|
|
"PEXT(32|64)rm",
|
|
"(V?)ADDPDrm",
|
|
"(V?)ADDPSrm",
|
|
"(V?)ADDSDrm",
|
|
"(V?)ADDSSrm",
|
|
"(V?)ADDSUBPDrm",
|
|
"(V?)ADDSUBPSrm",
|
|
"(V?)CMPPDrmi",
|
|
"(V?)CMPPSrmi",
|
|
"(V?)CMPSDrm",
|
|
"(V?)CMPSSrm",
|
|
"(V?)COMISDrm",
|
|
"(V?)COMISSrm",
|
|
"(V?)CVTDQ2PSrm",
|
|
"(V?)CVTPS2DQrm",
|
|
"(V?)CVTTPS2DQrm",
|
|
"(V?)MAX(C?)PDrm",
|
|
"(V?)MAX(C?)PSrm",
|
|
"(V?)MAX(C?)SDrm",
|
|
"(V?)MAX(C?)SSrm",
|
|
"(V?)MIN(C?)PDrm",
|
|
"(V?)MIN(C?)PSrm",
|
|
"(V?)MIN(C?)SDrm",
|
|
"(V?)MIN(C?)SSrm",
|
|
"(V?)SUBPDrm",
|
|
"(V?)SUBPSrm",
|
|
"(V?)SUBSDrm",
|
|
"(V?)SUBSSrm",
|
|
"(V?)UCOMISDrm",
|
|
"(V?)UCOMISSrm")>;
|
|
|
|
def BWWriteResGroup91_16 : SchedWriteRes<[BWPort1, BWPort0156, BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup91_16], (instrs IMUL16rm, IMUL16rmi, IMUL16rmi8)>;
|
|
|
|
def BWWriteResGroup91_16_2 : SchedWriteRes<[BWPort1, BWPort0156, BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 5;
|
|
}
|
|
def: InstRW<[BWWriteResGroup91_16_2], (instrs IMUL16m, MUL16m)>;
|
|
|
|
def BWWriteResGroup92 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup92], (instregex "VPMOVSXBDYrm",
|
|
"VPMOVSXBQYrm",
|
|
"VPMOVSXBWYrm",
|
|
"VPMOVSXDQYrm",
|
|
"VPMOVSXWDYrm",
|
|
"VPMOVSXWQYrm",
|
|
"VPMOVZXWDYrm")>;
|
|
|
|
def BWWriteResGroup93 : SchedWriteRes<[BWPort01,BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup93], (instregex "(V?)MULPDrm",
|
|
"(V?)MULPSrm",
|
|
"(V?)MULSDrm",
|
|
"(V?)MULSSrm")>;
|
|
|
|
def BWWriteResGroup94 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup94], (instregex "VBLENDVPDYrm",
|
|
"VBLENDVPSYrm",
|
|
"VMASKMOVPDYrm",
|
|
"VMASKMOVPSYrm",
|
|
"VPBLENDVBYrm",
|
|
"VPMASKMOVDYrm",
|
|
"VPMASKMOVQYrm")>;
|
|
|
|
def BWWriteResGroup95 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup95], (instregex "VPSLLVDrm",
|
|
"VPSRAVDrm",
|
|
"VPSRLVDrm")>;
|
|
|
|
def BWWriteResGroup96 : SchedWriteRes<[BWPort5,BWPort23,BWPort15]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup96], (instregex "MMX_PHADDDrm",
|
|
"MMX_PHADDSWrm",
|
|
"MMX_PHADDWrm",
|
|
"MMX_PHSUBDrm",
|
|
"MMX_PHSUBSWrm",
|
|
"MMX_PHSUBWrm",
|
|
"(V?)PHADDDrm",
|
|
"(V?)PHADDSWrm",
|
|
"(V?)PHADDWrm",
|
|
"(V?)PHSUBDrm",
|
|
"(V?)PHSUBSWrm",
|
|
"(V?)PHSUBWrm")>;
|
|
|
|
def BWWriteResGroup97 : SchedWriteRes<[BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup97], (instregex "RCL(8|16|32|64)m1",
|
|
"RCL(8|16|32|64)mi",
|
|
"RCR(8|16|32|64)m1",
|
|
"RCR(8|16|32|64)mi")>;
|
|
|
|
def BWWriteResGroup98 : SchedWriteRes<[BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup98], (instregex "ROR(8|16|32|64)mCL")>;
|
|
|
|
def BWWriteResGroup99 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 6;
|
|
let ResourceCycles = [1,1,1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup99], (instregex "XCHG(8|16|32|64)rm")>;
|
|
|
|
def BWWriteResGroup100 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 8;
|
|
let NumMicroOps = 6;
|
|
let ResourceCycles = [1,1,1,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup100], (instregex "ADC(8|16|32|64)mi",
|
|
"CMPXCHG(8|16|32|64)rm",
|
|
"ROL(8|16|32|64)mCL",
|
|
"SAR(8|16|32|64)mCL",
|
|
"SBB(8|16|32|64)mi",
|
|
"SHL(8|16|32|64)mCL",
|
|
"SHR(8|16|32|64)mCL")>;
|
|
def: InstRW<[BWWriteResGroup100, ReadAfterLd], (instrs ADC8mr, ADC16mr, ADC32mr, ADC64mr,
|
|
SBB8mr, SBB16mr, SBB32mr, SBB64mr)>;
|
|
|
|
def BWWriteResGroup101 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup101], (instregex "ADD_F32m",
|
|
"ADD_F64m",
|
|
"ILD_F16m",
|
|
"ILD_F32m",
|
|
"ILD_F64m",
|
|
"SUBR_F32m",
|
|
"SUBR_F64m",
|
|
"SUB_F32m",
|
|
"SUB_F64m",
|
|
"VADDPDYrm",
|
|
"VADDPSYrm",
|
|
"VADDSUBPDYrm",
|
|
"VADDSUBPSYrm",
|
|
"VCMPPDYrmi",
|
|
"VCMPPSYrmi",
|
|
"VCVTDQ2PSYrm",
|
|
"VCVTPS2DQYrm",
|
|
"VCVTTPS2DQYrm",
|
|
"VMAX(C?)PDYrm",
|
|
"VMAX(C?)PSYrm",
|
|
"VMIN(C?)PDYrm",
|
|
"VMIN(C?)PSYrm",
|
|
"VSUBPDYrm",
|
|
"VSUBPSYrm")>;
|
|
|
|
def BWWriteResGroup102 : SchedWriteRes<[BWPort5,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup102], (instregex "VPERM2F128rm",
|
|
"VPERM2I128rm",
|
|
"VPERMDYrm",
|
|
"VPERMPDYmi",
|
|
"VPERMPSYrm",
|
|
"VPERMQYmi",
|
|
"VPMOVZXBDYrm",
|
|
"VPMOVZXBQYrm",
|
|
"VPMOVZXBWYrm",
|
|
"VPMOVZXDQYrm",
|
|
"VPMOVZXWQYrm")>;
|
|
|
|
def BWWriteResGroup103 : SchedWriteRes<[BWPort01,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup103], (instregex "VMULPDYrm",
|
|
"VMULPSYrm")>;
|
|
|
|
def BWWriteResGroup104 : SchedWriteRes<[BWPort0,BWPort1,BWPort5]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup104], (instregex "(V?)DPPDrri")>;
|
|
|
|
def BWWriteResGroup105 : SchedWriteRes<[BWPort0,BWPort1,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup105], (instregex "(V?)CVTSD2SI64rm",
|
|
"(V?)CVTSD2SIrm",
|
|
"(V?)CVTSS2SI64rm",
|
|
"(V?)CVTSS2SIrm",
|
|
"(V?)CVTTSD2SI64rm",
|
|
"(V?)CVTTSD2SIrm",
|
|
"VCVTTSS2SI64rm",
|
|
"(V?)CVTTSS2SIrm")>;
|
|
|
|
def BWWriteResGroup106 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup106], (instregex "VCVTPS2PDYrm")>;
|
|
|
|
def BWWriteResGroup107 : SchedWriteRes<[BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup107], (instrs IMUL64m, MUL64m, MULX64rm)>;
|
|
def: InstRW<[BWWriteResGroup107], (instregex "CVTPD2DQrm",
|
|
"CVTPD2PSrm",
|
|
"CVTTPD2DQrm",
|
|
"MMX_CVTPD2PIirm",
|
|
"MMX_CVTPI2PDirm",
|
|
"MMX_CVTTPD2PIirm",
|
|
"(V?)CVTDQ2PDrm",
|
|
"(V?)CVTSD2SSrm")>;
|
|
|
|
def BWWriteResGroup108 : SchedWriteRes<[BWPort5,BWPort23,BWPort015]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup108], (instregex "VPBROADCASTB(Y?)rm",
|
|
"VPBROADCASTW(Y?)rm")>;
|
|
|
|
def BWWriteResGroup109 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup109], (instregex "VPSLLVDYrm",
|
|
"VPSRAVDYrm",
|
|
"VPSRLVDYrm")>;
|
|
|
|
def BWWriteResGroup110 : SchedWriteRes<[BWPort5,BWPort23,BWPort15]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup110], (instregex "VPHADDDYrm",
|
|
"VPHADDSWYrm",
|
|
"VPHADDWYrm",
|
|
"VPHSUBDYrm",
|
|
"VPHSUBSWYrm",
|
|
"VPHSUBWYrm")>;
|
|
|
|
def BWWriteResGroup111 : SchedWriteRes<[BWPort1,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup111], (instregex "SHLD(16|32|64)mri8",
|
|
"SHRD(16|32|64)mri8")>;
|
|
|
|
def BWWriteResGroup112 : SchedWriteRes<[BWPort23,BWPort06,BWPort0156]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup112], (instregex "RDRAND(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup113 : SchedWriteRes<[BWPort1,BWPort6,BWPort23,BWPort0156]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [1,2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup113], (instregex "LAR(16|32|64)rm",
|
|
"LSL(16|32|64)rm")>;
|
|
|
|
def BWWriteResGroup115 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup115], (instregex "MMX_PMADDUBSWrm",
|
|
"MMX_PMADDWDirm",
|
|
"MMX_PMULHRSWrm",
|
|
"MMX_PMULHUWirm",
|
|
"MMX_PMULHWirm",
|
|
"MMX_PMULLWirm",
|
|
"MMX_PMULUDQirm",
|
|
"MMX_PSADBWirm",
|
|
"(V?)PCMPGTQrm",
|
|
"(V?)PHMINPOSUWrm",
|
|
"(V?)PMADDUBSWrm",
|
|
"(V?)PMADDWDrm",
|
|
"(V?)PMULDQrm",
|
|
"(V?)PMULHRSWrm",
|
|
"(V?)PMULHUWrm",
|
|
"(V?)PMULHWrm",
|
|
"(V?)PMULLWrm",
|
|
"(V?)PMULUDQrm",
|
|
"(V?)PSADBWrm",
|
|
"(V?)RCPPSm",
|
|
"(V?)RCPSSm",
|
|
"(V?)RSQRTPSm",
|
|
"(V?)RSQRTSSm")>;
|
|
|
|
def BWWriteResGroup116 : SchedWriteRes<[BWPort01,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup116],
|
|
(instregex "VF(N)?M(ADD|SUB|ADDSUB|SUBADD)(132|213|231)P(D|S)m",
|
|
"VF(N)?M(ADD|SUB)(132|213|231)S(D|S)m")>;
|
|
|
|
def BWWriteResGroup117 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup117], (instregex "FICOM16m",
|
|
"FICOM32m",
|
|
"FICOMP16m",
|
|
"FICOMP32m")>;
|
|
|
|
def BWWriteResGroup118 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup118], (instregex "VPTESTYrm")>;
|
|
|
|
def BWWriteResGroup119 : SchedWriteRes<[BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup119], (instregex "(V?)HADDPDrm",
|
|
"(V?)HADDPSrm",
|
|
"(V?)HSUBPDrm",
|
|
"(V?)HSUBPSrm")>;
|
|
|
|
def BWWriteResGroup120 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 10;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup120], (instregex "CVTTSS2SI64rm")>;
|
|
|
|
def BWWriteResGroup121 : SchedWriteRes<[BWPort1,BWPort23,BWPort06,BWPort0156]> {
|
|
let Latency = 9;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup121], (instrs IMUL32m, MUL32m, MULX32rm)>;
|
|
|
|
def BWWriteResGroup122 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup122], (instregex "(V?)DIVPSrr")>;
|
|
|
|
def BWWriteResGroup122_1 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,3]; // Really 2.5 cycle throughput
|
|
}
|
|
def: InstRW<[BWWriteResGroup122_1], (instregex "(V?)DIVSSrr")>;
|
|
|
|
def BWWriteResGroup123 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup123], (instregex "MUL_F32m",
|
|
"MUL_F64m",
|
|
"VPCMPGTQYrm",
|
|
"VPMADDUBSWYrm",
|
|
"VPMADDWDYrm",
|
|
"VPMULDQYrm",
|
|
"VPMULHRSWYrm",
|
|
"VPMULHUWYrm",
|
|
"VPMULHWYrm",
|
|
"VPMULLWYrm",
|
|
"VPMULUDQYrm",
|
|
"VPSADBWYrm")>;
|
|
|
|
def BWWriteResGroup124 : SchedWriteRes<[BWPort01,BWPort23]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup124],
|
|
(instregex "VF(N)?M(ADD|SUB|ADDSUB|SUBADD)(132|213|231)P(D|S)Ym")>;
|
|
|
|
def BWWriteResGroup126 : SchedWriteRes<[BWPort0,BWPort015]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup126], (instregex "VRCPPSYr",
|
|
"VRSQRTPSYr")>;
|
|
|
|
def BWWriteResGroup127 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup127], (instregex "(V?)ROUNDPDm",
|
|
"(V?)ROUNDPSm",
|
|
"(V?)ROUNDSDm",
|
|
"(V?)ROUNDSSm")>;
|
|
|
|
def BWWriteResGroup128 : SchedWriteRes<[BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup128], (instregex "VCVTDQ2PDYrm")>;
|
|
|
|
def BWWriteResGroup129 : SchedWriteRes<[BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup129], (instregex "VHADDPDYrm",
|
|
"VHADDPSYrm",
|
|
"VHSUBPDYrm",
|
|
"VHSUBPSYrm")>;
|
|
|
|
def BWWriteResGroup130 : SchedWriteRes<[BWPort1,BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 6;
|
|
let ResourceCycles = [1,1,1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup130], (instregex "SHLD(16|32|64)mrCL",
|
|
"SHRD(16|32|64)mrCL")>;
|
|
|
|
def BWWriteResGroup131 : SchedWriteRes<[BWPort1,BWPort06,BWPort0156]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 7;
|
|
let ResourceCycles = [2,2,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup131], (instregex "RCL(16|32|64)rCL",
|
|
"RCR(16|32|64)rCL")>;
|
|
|
|
def BWWriteResGroup132 : SchedWriteRes<[BWPort1,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [1,4,1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup132], (instregex "RCL8rCL")>;
|
|
|
|
def BWWriteResGroup133 : SchedWriteRes<[BWPort06,BWPort0156]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 11;
|
|
let ResourceCycles = [2,9];
|
|
}
|
|
def: InstRW<[BWWriteResGroup133], (instrs LOOPE)>;
|
|
def: InstRW<[BWWriteResGroup133], (instrs LOOPNE)>;
|
|
|
|
def BWWriteResGroup135 : SchedWriteRes<[BWPort1,BWPort23]> {
|
|
let Latency = 12;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup135], (instregex "ADD_FI16m",
|
|
"ADD_FI32m",
|
|
"SUBR_FI16m",
|
|
"SUBR_FI32m",
|
|
"SUB_FI16m",
|
|
"SUB_FI32m",
|
|
"VROUNDPDYm",
|
|
"VROUNDPSYm")>;
|
|
|
|
def BWWriteResGroup136 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 12;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup136], (instregex "(V?)MPSADBWrmi")>;
|
|
|
|
def BWWriteResGroup137 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,7];
|
|
}
|
|
def: InstRW<[BWWriteResGroup137], (instregex "(V?)SQRTPSr")>;
|
|
|
|
def BWWriteResGroup137_1 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 11;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup137_1], (instregex "(V?)SQRTSSr")>;
|
|
|
|
def BWWriteResGroup138 : SchedWriteRes<[BWPort0,BWPort5,BWPort23]> {
|
|
let Latency = 13;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,2,1,7];
|
|
}
|
|
def: InstRW<[BWWriteResGroup138], (instregex "VMPSADBWYrmi")>;
|
|
|
|
def BWWriteResGroup139 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,8];
|
|
}
|
|
def: InstRW<[BWWriteResGroup139], (instregex "(V?)DIVPDrr")>;
|
|
|
|
def BWWriteResGroup139_1 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup139_1], (instregex "(V?)DIVSDrr")>;
|
|
|
|
def BWWriteResGroup141 : SchedWriteRes<[BWPort0,BWPort1,BWPort23]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup141], (instregex "MUL_FI16m",
|
|
"MUL_FI32m")>;
|
|
|
|
def BWWriteResGroup142 : SchedWriteRes<[BWPort0,BWPort1,BWPort5]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup142], (instregex "(V?)DPPS(Y?)rri")>;
|
|
|
|
def BWWriteResGroup143 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [1,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup143], (instregex "(V?)DPPDrmi")>;
|
|
|
|
def BWWriteResGroup144 : SchedWriteRes<[BWPort1,BWPort6,BWPort23,BWPort0156]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [2,2,1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup144], (instregex "LAR(16|32|64)rr")>;
|
|
|
|
def BWWriteResGroup145 : SchedWriteRes<[BWPort1,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 10;
|
|
let ResourceCycles = [2,3,1,4];
|
|
}
|
|
def: InstRW<[BWWriteResGroup145], (instregex "RCR8rCL")>;
|
|
|
|
def BWWriteResGroup146 : SchedWriteRes<[BWPort0,BWPort1,BWPort6,BWPort0156]> {
|
|
let Latency = 14;
|
|
let NumMicroOps = 12;
|
|
let ResourceCycles = [2,1,4,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup146], (instregex "XCH_F")>;
|
|
|
|
def BWWriteResGroup147 : SchedWriteRes<[BWPort0]> {
|
|
let Latency = 15;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup147], (instregex "DIVR_FPrST0",
|
|
"DIVR_FST0r",
|
|
"DIVR_FrST0")>;
|
|
|
|
def BWWriteResGroup149 : SchedWriteRes<[BWPort1,BWPort23,BWPort237,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 15;
|
|
let NumMicroOps = 10;
|
|
let ResourceCycles = [1,1,1,4,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup149], (instregex "RCL(8|16|32|64)mCL")>;
|
|
|
|
def BWWriteResGroup150 : SchedWriteRes<[BWPort0,BWPort23,BWFPDivider]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup150], (instregex "(V?)DIVPSrm",
|
|
"(V?)DIVSSrm")>;
|
|
|
|
def BWWriteResGroup151 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup151], (instregex "VPMULLDYrm")>;
|
|
|
|
def BWWriteResGroup153 : SchedWriteRes<[BWPort4,BWPort23,BWPort237,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 14;
|
|
let ResourceCycles = [1,1,1,4,2,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup153], (instregex "CMPXCHG8B")>;
|
|
|
|
def BWWriteResGroup154 : SchedWriteRes<[BWPort5]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 16;
|
|
let ResourceCycles = [16];
|
|
}
|
|
def: InstRW<[BWWriteResGroup154], (instrs VZEROALL)>;
|
|
|
|
def BWWriteResGroup155 : SchedWriteRes<[BWPort0,BWPort015,BWFPDivider]> {
|
|
let Latency = 17;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1,10];
|
|
}
|
|
def: InstRW<[BWWriteResGroup155], (instregex "VDIVPSYrr")>;
|
|
|
|
def BWWriteResGroup156 : SchedWriteRes<[BWPort0,BWPort23,BWPort015]> {
|
|
let Latency = 17;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup156], (instregex "VRCPPSYm",
|
|
"VRSQRTPSYm")>;
|
|
|
|
def BWWriteResGroup157 : SchedWriteRes<[BWPort0,BWPort23,BWFPDivider]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1,7];
|
|
}
|
|
def: InstRW<[BWWriteResGroup157], (instregex "(V?)SQRTPSm",
|
|
"(V?)SQRTSSm")>;
|
|
|
|
def BWWriteResGroup159 : SchedWriteRes<[BWPort5,BWPort6,BWPort06,BWPort0156]> {
|
|
let Latency = 18;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [1,1,1,5];
|
|
}
|
|
def: InstRW<[BWWriteResGroup159], (instrs CPUID)>;
|
|
def: InstRW<[BWWriteResGroup159], (instrs RDTSC)>;
|
|
|
|
def BWWriteResGroup160 : SchedWriteRes<[BWPort1,BWPort23,BWPort237,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 18;
|
|
let NumMicroOps = 11;
|
|
let ResourceCycles = [2,1,1,3,1,3];
|
|
}
|
|
def: InstRW<[BWWriteResGroup160], (instregex "RCR(8|16|32|64)mCL")>;
|
|
|
|
def BWWriteResGroup161 : SchedWriteRes<[BWPort0,BWPort23,BWFPDivider]> {
|
|
let Latency = 19;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1,8];
|
|
}
|
|
def: InstRW<[BWWriteResGroup161], (instregex "(V?)DIVPDrm",
|
|
"(V?)DIVSDrm")>;
|
|
|
|
def BWWriteResGroup163 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 19;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [2,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup163], (instregex "(V?)DPPSrmi")>;
|
|
|
|
def BWWriteResGroup165 : SchedWriteRes<[BWPort0]> {
|
|
let Latency = 20;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup165], (instregex "DIV_FPrST0",
|
|
"DIV_FST0r",
|
|
"DIV_FrST0")>;
|
|
|
|
def BWWriteResGroup166 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23]> {
|
|
let Latency = 20;
|
|
let NumMicroOps = 5;
|
|
let ResourceCycles = [2,1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup166], (instregex "VDPPSYrmi")>;
|
|
|
|
def BWWriteResGroup167 : SchedWriteRes<[BWPort4,BWPort5,BWPort6,BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 20;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [1,1,1,1,1,1,2];
|
|
}
|
|
def: InstRW<[BWWriteResGroup167], (instregex "INSB",
|
|
"INSL",
|
|
"INSW")>;
|
|
|
|
def BWWriteResGroup168 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,14];
|
|
}
|
|
def: InstRW<[BWWriteResGroup168], (instregex "(V?)SQRTPDr")>;
|
|
|
|
def BWWriteResGroup168_1 : SchedWriteRes<[BWPort0,BWFPDivider]> {
|
|
let Latency = 16;
|
|
let NumMicroOps = 1;
|
|
let ResourceCycles = [1,8];
|
|
}
|
|
def: InstRW<[BWWriteResGroup168_1], (instregex "(V?)SQRTSDr")>;
|
|
|
|
def BWWriteResGroup169 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 21;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup169], (instregex "DIV_F32m",
|
|
"DIV_F64m")>;
|
|
|
|
def BWWriteResGroup170 : SchedWriteRes<[BWPort0,BWPort015,BWFPDivider]> {
|
|
let Latency = 21;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1,14];
|
|
}
|
|
def: InstRW<[BWWriteResGroup170], (instregex "VSQRTPSYr")>;
|
|
|
|
def BWWriteResGroup171 : SchedWriteRes<[BWPort0,BWPort4,BWPort5,BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 21;
|
|
let NumMicroOps = 19;
|
|
let ResourceCycles = [2,1,4,1,1,4,6];
|
|
}
|
|
def: InstRW<[BWWriteResGroup171], (instregex "CMPXCHG16B")>;
|
|
|
|
def BWWriteResGroup172 : SchedWriteRes<[BWPort6,BWPort23,BWPort0156]> {
|
|
let Latency = 22;
|
|
let NumMicroOps = 18;
|
|
let ResourceCycles = [1,1,16];
|
|
}
|
|
def: InstRW<[BWWriteResGroup172], (instregex "POPF64")>;
|
|
|
|
def BWWriteResGroup173 : SchedWriteRes<[BWPort0,BWPort015,BWFPDivider]> {
|
|
let Latency = 23;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1,16];
|
|
}
|
|
def: InstRW<[BWWriteResGroup173], (instregex "VDIVPDYrr")>;
|
|
|
|
def BWWriteResGroup174 : SchedWriteRes<[BWPort0,BWPort23,BWPort015,BWFPDivider]> {
|
|
let Latency = 23;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1,10];
|
|
}
|
|
def: InstRW<[BWWriteResGroup174], (instregex "VDIVPSYrm")>;
|
|
|
|
def BWWriteResGroup176 : SchedWriteRes<[BWPort6,BWPort23,BWPort0156]> {
|
|
let Latency = 23;
|
|
let NumMicroOps = 19;
|
|
let ResourceCycles = [3,1,15];
|
|
}
|
|
def: InstRW<[BWWriteResGroup176], (instregex "XRSTOR(64)?")>;
|
|
|
|
def BWWriteResGroup177 : SchedWriteRes<[BWPort0,BWPort1,BWPort23]> {
|
|
let Latency = 24;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup177], (instregex "DIV_FI16m",
|
|
"DIV_FI32m")>;
|
|
|
|
def BWWriteResGroup179 : SchedWriteRes<[BWPort0,BWPort23,BWFPDivider]> {
|
|
let Latency = 21;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1,14];
|
|
}
|
|
def: InstRW<[BWWriteResGroup179], (instregex "(V?)SQRTPDm",
|
|
"(V?)SQRTSDm")>;
|
|
|
|
def BWWriteResGroup180 : SchedWriteRes<[BWPort0,BWPort23]> {
|
|
let Latency = 26;
|
|
let NumMicroOps = 2;
|
|
let ResourceCycles = [1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup180], (instregex "DIVR_F32m",
|
|
"DIVR_F64m")>;
|
|
|
|
def BWWriteResGroup181 : SchedWriteRes<[BWPort0,BWPort23,BWPort015,BWFPDivider]> {
|
|
let Latency = 27;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1,14];
|
|
}
|
|
def: InstRW<[BWWriteResGroup181], (instregex "VSQRTPSYm")>;
|
|
|
|
def BWWriteResGroup182 : SchedWriteRes<[BWPort0,BWPort1,BWPort23]> {
|
|
let Latency = 29;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [1,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup182], (instregex "DIVR_FI16m",
|
|
"DIVR_FI32m")>;
|
|
|
|
def BWWriteResGroup183 : SchedWriteRes<[BWPort0,BWPort23,BWPort015,BWFPDivider]> {
|
|
let Latency = 29;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1,16];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183], (instregex "VDIVPDYrm")>;
|
|
|
|
def BWWriteResGroup183_1 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 22;
|
|
let NumMicroOps = 7;
|
|
let ResourceCycles = [1,3,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_1], (instrs VGATHERQPDrm)>;
|
|
|
|
def BWWriteResGroup183_2 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 23;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [1,3,4,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_2], (instrs VGATHERQPDYrm)>;
|
|
|
|
def BWWriteResGroup183_3 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 24;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [1,5,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_3], (instrs VGATHERQPSYrm)>;
|
|
|
|
def BWWriteResGroup183_4 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 25;
|
|
let NumMicroOps = 7;
|
|
let ResourceCycles = [1,3,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_4], (instrs VGATHERDPDrm,
|
|
VGATHERDPSrm)>;
|
|
|
|
def BWWriteResGroup183_5 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 26;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [1,5,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_5], (instrs VGATHERDPDYrm)>;
|
|
|
|
def BWWriteResGroup183_6 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 26;
|
|
let NumMicroOps = 14;
|
|
let ResourceCycles = [1,4,8,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_6], (instrs VGATHERDPSYrm)>;
|
|
|
|
def BWWriteResGroup183_7 : SchedWriteRes<[BWPort4, BWPort5, BWPort23, BWPort0156]> {
|
|
let Latency = 27;
|
|
let NumMicroOps = 9;
|
|
let ResourceCycles = [1,5,2,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup183_7], (instrs VGATHERQPSrm)>;
|
|
|
|
def BWWriteResGroup185 : SchedWriteRes<[BWPort4,BWPort6,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 29;
|
|
let NumMicroOps = 27;
|
|
let ResourceCycles = [1,5,1,1,19];
|
|
}
|
|
def: InstRW<[BWWriteResGroup185], (instregex "XSAVE64")>;
|
|
|
|
def BWWriteResGroup186 : SchedWriteRes<[BWPort4,BWPort6,BWPort23,BWPort237,BWPort0156]> {
|
|
let Latency = 30;
|
|
let NumMicroOps = 28;
|
|
let ResourceCycles = [1,6,1,1,19];
|
|
}
|
|
def: InstRW<[BWWriteResGroup186], (instregex "^XSAVE$", "XSAVEC", "XSAVES", "XSAVEOPT")>;
|
|
|
|
def BWWriteResGroup187 : SchedWriteRes<[BWPort01,BWPort15,BWPort015,BWPort0156]> {
|
|
let Latency = 31;
|
|
let NumMicroOps = 31;
|
|
let ResourceCycles = [8,1,21,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup187], (instregex "MMX_EMMS")>;
|
|
|
|
def BWWriteResGroup189 : SchedWriteRes<[BWPort0,BWPort015,BWFPDivider]> {
|
|
let Latency = 29;
|
|
let NumMicroOps = 3;
|
|
let ResourceCycles = [2,1,28];
|
|
}
|
|
def: InstRW<[BWWriteResGroup189], (instregex "VSQRTPDYr")>;
|
|
|
|
def BWWriteResGroup190 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23,BWPort0156]> {
|
|
let Latency = 34;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [2,2,2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup190], (instregex "DIV(8|16|32|64)m")>;
|
|
|
|
def BWWriteResGroup191 : SchedWriteRes<[BWPort5,BWPort6,BWPort23,BWPort06,BWPort0156]> {
|
|
let Latency = 34;
|
|
let NumMicroOps = 23;
|
|
let ResourceCycles = [1,5,3,4,10];
|
|
}
|
|
def: InstRW<[BWWriteResGroup191], (instregex "IN(8|16|32)ri",
|
|
"IN(8|16|32)rr")>;
|
|
|
|
def BWWriteResGroup193 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort23,BWPort0156]> {
|
|
let Latency = 35;
|
|
let NumMicroOps = 8;
|
|
let ResourceCycles = [2,2,2,1,1];
|
|
}
|
|
def: InstRW<[BWWriteResGroup193], (instregex "IDIV(8|16|32|64)m")>;
|
|
|
|
def BWWriteResGroup194 : SchedWriteRes<[BWPort5,BWPort6,BWPort23,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 35;
|
|
let NumMicroOps = 23;
|
|
let ResourceCycles = [1,5,2,1,4,10];
|
|
}
|
|
def: InstRW<[BWWriteResGroup194], (instregex "OUT(8|16|32)ir",
|
|
"OUT(8|16|32)rr")>;
|
|
|
|
def BWWriteResGroup195 : SchedWriteRes<[BWPort0,BWPort23,BWPort015,BWFPDivider]> {
|
|
let Latency = 35;
|
|
let NumMicroOps = 4;
|
|
let ResourceCycles = [2,1,1,28];
|
|
}
|
|
def: InstRW<[BWWriteResGroup195], (instregex "VSQRTPDYm")>;
|
|
|
|
def BWWriteResGroup196 : SchedWriteRes<[BWPort5,BWPort0156]> {
|
|
let Latency = 42;
|
|
let NumMicroOps = 22;
|
|
let ResourceCycles = [2,20];
|
|
}
|
|
def: InstRW<[BWWriteResGroup196], (instrs RDTSCP)>;
|
|
|
|
def BWWriteResGroup197 : SchedWriteRes<[BWPort0,BWPort01,BWPort23,BWPort05,BWPort06,BWPort015,BWPort0156]> {
|
|
let Latency = 60;
|
|
let NumMicroOps = 64;
|
|
let ResourceCycles = [2,2,8,1,10,2,39];
|
|
}
|
|
def: InstRW<[BWWriteResGroup197], (instregex "FLDENVm")>;
|
|
|
|
def BWWriteResGroup198 : SchedWriteRes<[BWPort0,BWPort6,BWPort23,BWPort05,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 63;
|
|
let NumMicroOps = 88;
|
|
let ResourceCycles = [4,4,31,1,2,1,45];
|
|
}
|
|
def: InstRW<[BWWriteResGroup198], (instrs FXRSTOR64)>;
|
|
|
|
def BWWriteResGroup199 : SchedWriteRes<[BWPort0,BWPort6,BWPort23,BWPort05,BWPort06,BWPort15,BWPort0156]> {
|
|
let Latency = 63;
|
|
let NumMicroOps = 90;
|
|
let ResourceCycles = [4,2,33,1,2,1,47];
|
|
}
|
|
def: InstRW<[BWWriteResGroup199], (instrs FXRSTOR)>;
|
|
|
|
def BWWriteResGroup200 : SchedWriteRes<[BWPort5,BWPort01,BWPort0156]> {
|
|
let Latency = 75;
|
|
let NumMicroOps = 15;
|
|
let ResourceCycles = [6,3,6];
|
|
}
|
|
def: InstRW<[BWWriteResGroup200], (instregex "FNINIT")>;
|
|
|
|
def BWWriteResGroup201 : SchedWriteRes<[BWPort0,BWPort1,BWPort5,BWPort6,BWPort01,BWPort0156]> {
|
|
let Latency = 80;
|
|
let NumMicroOps = 32;
|
|
let ResourceCycles = [7,7,3,3,1,11];
|
|
}
|
|
def: InstRW<[BWWriteResGroup201], (instregex "DIV(16|32|64)r")>;
|
|
|
|
def BWWriteResGroup202 : SchedWriteRes<[BWPort0,BWPort1,BWPort4,BWPort5,BWPort6,BWPort237,BWPort06,BWPort0156]> {
|
|
let Latency = 115;
|
|
let NumMicroOps = 100;
|
|
let ResourceCycles = [9,9,11,8,1,11,21,30];
|
|
}
|
|
def: InstRW<[BWWriteResGroup202], (instregex "FSTENVm")>;
|
|
|
|
} // SchedModel
|
|
|