mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-22 18:54:02 +01:00
e25a1a8f41
First patch in a series adding MC layer support for the Arm Scalable
Matrix Extension.
This patch adds the following features:
sme, sme-i64, sme-f64
The sme-i64 and sme-f64 flags are for the optional I16I64 and F64F64
features.
If a target supports I16I64 then the following instructions are
implemented:
* 64-bit integer ADDHA and ADDVA variants (D105570).
* SMOPA, SMOPS, SUMOPA, SUMOPS, UMOPA, UMOPS, USMOPA, and USMOPS
instructions that accumulate 16-bit integer outer products into 64-bit
integer tiles.
If a target supports F64F64 then the FMOPA and FMOPS instructions that
accumulate double-precision floating-point outer products into
double-precision tiles are implemented.
Outer products are implemented in D105571.
The reference can be found here:
https://developer.arm.com/documentation/ddi0602/2021-06
Reviewed By: CarolineConcatto
Differential Revision: https://reviews.llvm.org/D105569
297 lines
15 KiB
TableGen
297 lines
15 KiB
TableGen
//==- AArch64SchedA53.td - Cortex-A53 Scheduling Definitions -*- 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 defines the itinerary class data for the ARM Cortex A53 processors.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
// ===---------------------------------------------------------------------===//
|
|
// The following definitions describe the simpler per-operand machine model.
|
|
// This works with MachineScheduler. See MCSchedule.h for details.
|
|
|
|
// Cortex-A53 machine model for scheduling and other instruction cost heuristics.
|
|
def CortexA53Model : SchedMachineModel {
|
|
let MicroOpBufferSize = 0; // Explicitly set to zero since A53 is in-order.
|
|
let IssueWidth = 2; // 2 micro-ops are dispatched per cycle.
|
|
let LoadLatency = 3; // Optimistic load latency assuming bypass.
|
|
// This is overriden by OperandCycles if the
|
|
// Itineraries are queried instead.
|
|
let MispredictPenalty = 9; // Based on "Cortex-A53 Software Optimisation
|
|
// Specification - Instruction Timings"
|
|
// v 1.0 Spreadsheet
|
|
let CompleteModel = 1;
|
|
|
|
list<Predicate> UnsupportedFeatures = !listconcat(SVEUnsupported.F,
|
|
PAUnsupported.F,
|
|
SMEUnsupported.F);
|
|
}
|
|
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
// Define each kind of processor resource and number available.
|
|
|
|
// Modeling each pipeline as a ProcResource using the BufferSize = 0 since
|
|
// Cortex-A53 is in-order.
|
|
|
|
def A53UnitALU : ProcResource<2> { let BufferSize = 0; } // Int ALU
|
|
def A53UnitMAC : ProcResource<1> { let BufferSize = 0; } // Int MAC
|
|
def A53UnitDiv : ProcResource<1> { let BufferSize = 0; } // Int Division
|
|
def A53UnitLdSt : ProcResource<1> { let BufferSize = 0; } // Load/Store
|
|
def A53UnitB : ProcResource<1> { let BufferSize = 0; } // Branch
|
|
def A53UnitFPALU : ProcResource<1> { let BufferSize = 0; } // FP ALU
|
|
def A53UnitFPMDS : ProcResource<1> { let BufferSize = 0; } // FP Mult/Div/Sqrt
|
|
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
// Subtarget-specific SchedWrite types which both map the ProcResources and
|
|
// set the latency.
|
|
|
|
let SchedModel = CortexA53Model in {
|
|
|
|
// ALU - Despite having a full latency of 4, most of the ALU instructions can
|
|
// forward a cycle earlier and then two cycles earlier in the case of a
|
|
// shift-only instruction. These latencies will be incorrect when the
|
|
// result cannot be forwarded, but modeling isn't rocket surgery.
|
|
def : WriteRes<WriteImm, [A53UnitALU]> { let Latency = 3; }
|
|
def : WriteRes<WriteI, [A53UnitALU]> { let Latency = 3; }
|
|
def : WriteRes<WriteISReg, [A53UnitALU]> { let Latency = 3; }
|
|
def : WriteRes<WriteIEReg, [A53UnitALU]> { let Latency = 3; }
|
|
def : WriteRes<WriteIS, [A53UnitALU]> { let Latency = 2; }
|
|
def : WriteRes<WriteExtr, [A53UnitALU]> { let Latency = 3; }
|
|
|
|
// MAC
|
|
def : WriteRes<WriteIM32, [A53UnitMAC]> { let Latency = 4; }
|
|
def : WriteRes<WriteIM64, [A53UnitMAC]> { let Latency = 4; }
|
|
|
|
// Div
|
|
def : WriteRes<WriteID32, [A53UnitDiv]> { let Latency = 4; }
|
|
def : WriteRes<WriteID64, [A53UnitDiv]> { let Latency = 4; }
|
|
|
|
// Load
|
|
def : WriteRes<WriteLD, [A53UnitLdSt]> { let Latency = 4; }
|
|
def : WriteRes<WriteLDIdx, [A53UnitLdSt]> { let Latency = 4; }
|
|
def : WriteRes<WriteLDHi, [A53UnitLdSt]> { let Latency = 4; }
|
|
|
|
// Vector Load - Vector loads take 1-5 cycles to issue. For the WriteVecLd
|
|
// below, choosing the median of 3 which makes the latency 6.
|
|
// May model this more carefully in the future. The remaining
|
|
// A53WriteVLD# types represent the 1-5 cycle issues explicitly.
|
|
def : WriteRes<WriteVLD, [A53UnitLdSt]> { let Latency = 6;
|
|
let ResourceCycles = [3]; }
|
|
def A53WriteVLD1 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 4; }
|
|
def A53WriteVLD2 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 5;
|
|
let ResourceCycles = [2]; }
|
|
def A53WriteVLD3 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 6;
|
|
let ResourceCycles = [3]; }
|
|
def A53WriteVLD4 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 7;
|
|
let ResourceCycles = [4]; }
|
|
def A53WriteVLD5 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 8;
|
|
let ResourceCycles = [5]; }
|
|
|
|
// Pre/Post Indexing - Performed as part of address generation which is already
|
|
// accounted for in the WriteST* latencies below
|
|
def : WriteRes<WriteAdr, []> { let Latency = 0; }
|
|
|
|
// Store
|
|
def : WriteRes<WriteST, [A53UnitLdSt]> { let Latency = 4; }
|
|
def : WriteRes<WriteSTP, [A53UnitLdSt]> { let Latency = 4; }
|
|
def : WriteRes<WriteSTIdx, [A53UnitLdSt]> { let Latency = 4; }
|
|
def : WriteRes<WriteSTX, [A53UnitLdSt]> { let Latency = 4; }
|
|
|
|
// Vector Store - Similar to vector loads, can take 1-3 cycles to issue.
|
|
def : WriteRes<WriteVST, [A53UnitLdSt]> { let Latency = 5;
|
|
let ResourceCycles = [2];}
|
|
def A53WriteVST1 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 4; }
|
|
def A53WriteVST2 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 5;
|
|
let ResourceCycles = [2]; }
|
|
def A53WriteVST3 : SchedWriteRes<[A53UnitLdSt]> { let Latency = 6;
|
|
let ResourceCycles = [3]; }
|
|
|
|
def : WriteRes<WriteAtomic, []> { let Unsupported = 1; }
|
|
|
|
// Branch
|
|
def : WriteRes<WriteBr, [A53UnitB]>;
|
|
def : WriteRes<WriteBrReg, [A53UnitB]>;
|
|
def : WriteRes<WriteSys, [A53UnitB]>;
|
|
def : WriteRes<WriteBarrier, [A53UnitB]>;
|
|
def : WriteRes<WriteHint, [A53UnitB]>;
|
|
|
|
// FP ALU
|
|
def : WriteRes<WriteF, [A53UnitFPALU]> { let Latency = 6; }
|
|
def : WriteRes<WriteFCmp, [A53UnitFPALU]> { let Latency = 6; }
|
|
def : WriteRes<WriteFCvt, [A53UnitFPALU]> { let Latency = 6; }
|
|
def : WriteRes<WriteFCopy, [A53UnitFPALU]> { let Latency = 6; }
|
|
def : WriteRes<WriteFImm, [A53UnitFPALU]> { let Latency = 6; }
|
|
def : WriteRes<WriteV, [A53UnitFPALU]> { let Latency = 6; }
|
|
|
|
// FP Mul, Div, Sqrt
|
|
def : WriteRes<WriteFMul, [A53UnitFPMDS]> { let Latency = 6; }
|
|
def : WriteRes<WriteFDiv, [A53UnitFPMDS]> { let Latency = 33;
|
|
let ResourceCycles = [29]; }
|
|
def A53WriteFMAC : SchedWriteRes<[A53UnitFPMDS]> { let Latency = 10; }
|
|
def A53WriteFDivSP : SchedWriteRes<[A53UnitFPMDS]> { let Latency = 18;
|
|
let ResourceCycles = [14]; }
|
|
def A53WriteFDivDP : SchedWriteRes<[A53UnitFPMDS]> { let Latency = 33;
|
|
let ResourceCycles = [29]; }
|
|
def A53WriteFSqrtSP : SchedWriteRes<[A53UnitFPMDS]> { let Latency = 17;
|
|
let ResourceCycles = [13]; }
|
|
def A53WriteFSqrtDP : SchedWriteRes<[A53UnitFPMDS]> { let Latency = 32;
|
|
let ResourceCycles = [28]; }
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
// Subtarget-specific SchedRead types.
|
|
|
|
// No forwarding for these reads.
|
|
def : ReadAdvance<ReadExtrHi, 0>;
|
|
def : ReadAdvance<ReadAdrBase, 0>;
|
|
def : ReadAdvance<ReadVLD, 0>;
|
|
|
|
// ALU - Most operands in the ALU pipes are not needed for two cycles. Shiftable
|
|
// operands are needed one cycle later if and only if they are to be
|
|
// shifted. Otherwise, they too are needed two cycles later. This same
|
|
// ReadAdvance applies to Extended registers as well, even though there is
|
|
// a separate SchedPredicate for them.
|
|
def : ReadAdvance<ReadI, 2, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
def A53ReadShifted : SchedReadAdvance<1, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
def A53ReadNotShifted : SchedReadAdvance<2, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
def A53ReadISReg : SchedReadVariant<[
|
|
SchedVar<RegShiftedPred, [A53ReadShifted]>,
|
|
SchedVar<NoSchedPred, [A53ReadNotShifted]>]>;
|
|
def : SchedAlias<ReadISReg, A53ReadISReg>;
|
|
|
|
def A53ReadIEReg : SchedReadVariant<[
|
|
SchedVar<RegExtendedPred, [A53ReadShifted]>,
|
|
SchedVar<NoSchedPred, [A53ReadNotShifted]>]>;
|
|
def : SchedAlias<ReadIEReg, A53ReadIEReg>;
|
|
|
|
// MAC - Operands are generally needed one cycle later in the MAC pipe.
|
|
// Accumulator operands are needed two cycles later.
|
|
def : ReadAdvance<ReadIM, 1, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
def : ReadAdvance<ReadIMA, 2, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
|
|
// Div
|
|
def : ReadAdvance<ReadID, 1, [WriteImm,WriteI,
|
|
WriteISReg, WriteIEReg,WriteIS,
|
|
WriteID32,WriteID64,
|
|
WriteIM32,WriteIM64]>;
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
// Subtarget-specific InstRWs.
|
|
|
|
//---
|
|
// Miscellaneous
|
|
//---
|
|
def : InstRW<[WriteI], (instrs COPY)>;
|
|
|
|
//---
|
|
// Vector Loads
|
|
//---
|
|
def : InstRW<[A53WriteVLD1], (instregex "LD1i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVLD1], (instregex "LD1Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD1], (instregex "LD1Onev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD1Twov(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD3], (instregex "LD1Threev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD4], (instregex "LD1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD1, WriteAdr], (instregex "LD1i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVLD1, WriteAdr], (instregex "LD1Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD1, WriteAdr], (instregex "LD1Onev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD1Twov(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD3, WriteAdr], (instregex "LD1Threev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD4, WriteAdr], (instregex "LD1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVLD1], (instregex "LD2i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVLD1], (instregex "LD2Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD2Twov(8b|4h|2s)$")>;
|
|
def : InstRW<[A53WriteVLD4], (instregex "LD2Twov(16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD1, WriteAdr], (instregex "LD2i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVLD1, WriteAdr], (instregex "LD2Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD2Twov(8b|4h|2s)_POST$")>;
|
|
def : InstRW<[A53WriteVLD4, WriteAdr], (instregex "LD2Twov(16b|8h|4s|2d)_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD3i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD3Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD4], (instregex "LD3Threev(8b|4h|2s|1d|16b|8h|4s)$")>;
|
|
def : InstRW<[A53WriteVLD3], (instregex "LD3Threev2d$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD3i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD3Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD4, WriteAdr], (instregex "LD3Threev(8b|4h|2s|1d|16b|8h|4s)_POST$")>;
|
|
def : InstRW<[A53WriteVLD3, WriteAdr], (instregex "LD3Threev2d_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD4i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVLD2], (instregex "LD4Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVLD5], (instregex "LD4Fourv(8b|4h|2s|1d|16b|8h|4s)$")>;
|
|
def : InstRW<[A53WriteVLD4], (instregex "LD4Fourv(2d)$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD4i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVLD2, WriteAdr], (instregex "LD4Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVLD5, WriteAdr], (instregex "LD4Fourv(8b|4h|2s|1d|16b|8h|4s)_POST$")>;
|
|
def : InstRW<[A53WriteVLD4, WriteAdr], (instregex "LD4Fourv(2d)_POST$")>;
|
|
|
|
//---
|
|
// Vector Stores
|
|
//---
|
|
def : InstRW<[A53WriteVST1], (instregex "ST1i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVST1], (instregex "ST1Onev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVST1], (instregex "ST1Twov(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVST2], (instregex "ST1Threev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVST2], (instregex "ST1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVST1, WriteAdr], (instregex "ST1i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVST1, WriteAdr], (instregex "ST1Onev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVST1, WriteAdr], (instregex "ST1Twov(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST1Threev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVST1], (instregex "ST2i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVST1], (instregex "ST2Twov(8b|4h|2s)$")>;
|
|
def : InstRW<[A53WriteVST2], (instregex "ST2Twov(16b|8h|4s|2d)$")>;
|
|
def : InstRW<[A53WriteVST1, WriteAdr], (instregex "ST2i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVST1, WriteAdr], (instregex "ST2Twov(8b|4h|2s)_POST$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST2Twov(16b|8h|4s|2d)_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVST2], (instregex "ST3i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVST3], (instregex "ST3Threev(8b|4h|2s|1d|16b|8h|4s)$")>;
|
|
def : InstRW<[A53WriteVST2], (instregex "ST3Threev(2d)$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST3i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVST3, WriteAdr], (instregex "ST3Threev(8b|4h|2s|1d|16b|8h|4s)_POST$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST3Threev(2d)_POST$")>;
|
|
|
|
def : InstRW<[A53WriteVST2], (instregex "ST4i(8|16|32|64)$")>;
|
|
def : InstRW<[A53WriteVST3], (instregex "ST4Fourv(8b|4h|2s|1d|16b|8h|4s)$")>;
|
|
def : InstRW<[A53WriteVST2], (instregex "ST4Fourv(2d)$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST4i(8|16|32|64)_POST$")>;
|
|
def : InstRW<[A53WriteVST3, WriteAdr], (instregex "ST4Fourv(8b|4h|2s|1d|16b|8h|4s)_POST$")>;
|
|
def : InstRW<[A53WriteVST2, WriteAdr], (instregex "ST4Fourv(2d)_POST$")>;
|
|
|
|
//---
|
|
// Floating Point MAC, DIV, SQRT
|
|
//---
|
|
def : InstRW<[A53WriteFMAC], (instregex "^FN?M(ADD|SUB).*")>;
|
|
def : InstRW<[A53WriteFMAC], (instregex "^FML(A|S).*")>;
|
|
def : InstRW<[A53WriteFDivSP], (instrs FDIVSrr)>;
|
|
def : InstRW<[A53WriteFDivDP], (instrs FDIVDrr)>;
|
|
def : InstRW<[A53WriteFDivSP], (instregex "^FDIVv.*32$")>;
|
|
def : InstRW<[A53WriteFDivDP], (instregex "^FDIVv.*64$")>;
|
|
def : InstRW<[A53WriteFSqrtSP], (instregex "^.*SQRT.*32$")>;
|
|
def : InstRW<[A53WriteFSqrtDP], (instregex "^.*SQRT.*64$")>;
|
|
|
|
}
|