1 files changed, 227 insertions, 0 deletions
diff --git a/llvm/lib/Target/RISCV/RISCVSchedRocket32.td b/llvm/lib/Target/RISCV/RISCVSchedRocket32.td
new file mode 100644
index 0000000000000..305e2b9b5927a
--- /dev/null
+++ b/llvm/lib/Target/RISCV/RISCVSchedRocket32.td
@@ -0,0 +1,227 @@
+//==- RISCVSchedRocket32.td - Rocket 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
+//
+//===----------------------------------------------------------------------===//
+
+// ===---------------------------------------------------------------------===//
+// The following definitions describe the simpler per-operand machine model.
+// This works with MachineScheduler. See MCSchedule.h for details.
+
+// Rocket machine model for scheduling and other instruction cost heuristics.
+def Rocket32Model : SchedMachineModel {
+  let MicroOpBufferSize = 0; // Explicitly set to zero since Rocket is in-order.
+  let IssueWidth = 1;        // 1 micro-ops are dispatched per cycle.
+  let LoadLatency = 3;
+  let MispredictPenalty = 3;
+  let CompleteModel = 1;
+  let UnsupportedFeatures = [HasStdExtV];
+}
+
+//===----------------------------------------------------------------------===//
+// Define each kind of processor resource and number available.
+
+// Modeling each pipeline as a ProcResource using the BufferSize = 0 since
+// Rocket is in-order.
+
+let BufferSize = 0 in {
+def Rocket32UnitALU        : ProcResource<1>; // Int ALU
+def Rocket32UnitIMul       : ProcResource<1>; // Int Multiply
+def Rocket32UnitMem        : ProcResource<1>; // Load/Store
+def Rocket32UnitB          : ProcResource<1>; // Branch
+
+def Rocket32UnitFPALU      : ProcResource<1>; // FP ALU
+}
+
+let BufferSize = 1 in {
+def Rocket32UnitIDiv       : ProcResource<1>; // Int Division
+def Rocket32UnitFPDivSqrt  : ProcResource<1>; // FP Divide/Sqrt'
+}
+
+//===----------------------------------------------------------------------===//
+// Subtarget-specific SchedWrite types which both map the ProcResources and
+// set the latency.
+
+let SchedModel = Rocket32Model in {
+
+def : WriteRes<WriteJmp, [Rocket32UnitB]>;
+def : WriteRes<WriteJal, [Rocket32UnitB]>;
+def : WriteRes<WriteJalr, [Rocket32UnitB]>;
+def : WriteRes<WriteJmpReg, [Rocket32UnitB]>;
+
+def : WriteRes<WriteIALU, [Rocket32UnitALU]>;
+def : WriteRes<WriteShift, [Rocket32UnitALU]>;
+
+// Multiplies on Rocket differ by implementation; placeholder until
+// we can determine how to read from command line
+def : WriteRes<WriteIMul, [Rocket32UnitIMul]> { let Latency = 4; }
+
+// 32-bit divides have worse case latency of 34 cycle
+def : WriteRes<WriteIDiv, [Rocket32UnitIDiv]> {
+  let Latency = 34;
+  let ResourceCycles = [34];
+}
+
+// Memory
+def : WriteRes<WriteSTB, [Rocket32UnitMem]>;
+def : WriteRes<WriteSTH, [Rocket32UnitMem]>;
+def : WriteRes<WriteSTW, [Rocket32UnitMem]>;
+def : WriteRes<WriteFST32, [Rocket32UnitMem]>;
+def : WriteRes<WriteFST64, [Rocket32UnitMem]>;
+
+let Latency = 3 in {
+def : WriteRes<WriteLDB, [Rocket32UnitMem]>;
+def : WriteRes<WriteLDH, [Rocket32UnitMem]>;
+def : WriteRes<WriteCSR, [Rocket32UnitALU]>;
+}
+
+let Latency = 2 in {
+def : WriteRes<WriteLDW, [Rocket32UnitMem]>;
+def : WriteRes<WriteFLD32, [Rocket32UnitMem]>;
+def : WriteRes<WriteFLD64, [Rocket32UnitMem]>;
+
+def : WriteRes<WriteAtomicW, [Rocket32UnitMem]>;
+def : WriteRes<WriteAtomicLDW, [Rocket32UnitMem]>;
+}
+
+def : WriteRes<WriteAtomicSTW, [Rocket32UnitMem]>;
+
+// Most FP single precision operations are 4 cycles
+let Latency = 4 in {
+def : WriteRes<WriteFALU32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFSGNJ32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMinMax32, [Rocket32UnitFPALU]>;
+}
+
+// Most FP double precision operations are 6 cycles
+let Latency = 6 in {
+def : WriteRes<WriteFALU64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFSGNJ64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMinMax64, [Rocket32UnitFPALU]>;
+}
+
+let Latency = 2 in {
+def : WriteRes<WriteFCvtI32ToF32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCvtI32ToF64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCvtF32ToI32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCvtF64ToI32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCvtF32ToF64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCvtF64ToF32, [Rocket32UnitFPALU]>;
+
+def : WriteRes<WriteFClass32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFClass64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCmp32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFCmp64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMovF32ToI32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMovI32ToF32, [Rocket32UnitFPALU]>;
+}
+
+let Latency = 5 in {
+def : WriteRes<WriteFMul32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMulAdd32, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMulSub32, [Rocket32UnitFPALU]>;
+}
+
+let Latency = 7 in {
+def : WriteRes<WriteFMul64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMulAdd64, [Rocket32UnitFPALU]>;
+def : WriteRes<WriteFMulSub64, [Rocket32UnitFPALU]>;
+}
+
+// FP Divide unit on Rocket is not pipelined, so set resource cycles to latency
+let Latency = 20, ResourceCycles = [20] in {
+def : WriteRes<WriteFDiv32, [Rocket32UnitFPDivSqrt]>;
+def : WriteRes<WriteFDiv64, [Rocket32UnitFPDivSqrt]>;
+}
+
+// FP Sqrt unit on Rocket is not pipelined, so set resource cycles to latency
+def : WriteRes<WriteFSqrt32, [Rocket32UnitFPDivSqrt]> { let Latency = 20;
+                                                        let ResourceCycles = [20];}
+def : WriteRes<WriteFSqrt64, [Rocket32UnitFPDivSqrt]> { let Latency = 25;
+                                                        let ResourceCycles = [25];}
+
+def : WriteRes<WriteNop, []>;
+
+def : InstRW<[WriteIALU], (instrs COPY)>;
+
+let Unsupported = 1 in {
+def : WriteRes<WriteIALU32, []>;
+def : WriteRes<WriteShift32, []>;
+def : WriteRes<WriteIMul32, []>;
+def : WriteRes<WriteIDiv32, []>;
+def : WriteRes<WriteSTD, []>;
+def : WriteRes<WriteLDWU, []>;
+def : WriteRes<WriteLDD, []>;
+def : WriteRes<WriteAtomicD, []>;
+def : WriteRes<WriteAtomicLDD, []>;
+def : WriteRes<WriteAtomicSTD, []>;
+def : WriteRes<WriteFCvtI64ToF32, []>;
+def : WriteRes<WriteFCvtI64ToF64, []>;
+def : WriteRes<WriteFCvtF64ToI64, []>;
+def : WriteRes<WriteFCvtF32ToI64, []>;
+def : WriteRes<WriteFMovI64ToF64, []>;
+def : WriteRes<WriteFMovF64ToI64, []>;
+}
+
+//===----------------------------------------------------------------------===//
+// Subtarget-specific SchedRead types with cycles.
+// Dummy definitions for RocketCore.
+def : ReadAdvance<ReadJmp, 0>;
+def : ReadAdvance<ReadJalr, 0>;
+def : ReadAdvance<ReadCSR, 0>;
+def : ReadAdvance<ReadStoreData, 0>;
+def : ReadAdvance<ReadMemBase, 0>;
+def : ReadAdvance<ReadIALU, 0>;
+def : ReadAdvance<ReadIALU32, 0>;
+def : ReadAdvance<ReadShift, 0>;
+def : ReadAdvance<ReadShift32, 0>;
+def : ReadAdvance<ReadIDiv, 0>;
+def : ReadAdvance<ReadIDiv32, 0>;
+def : ReadAdvance<ReadIMul, 0>;
+def : ReadAdvance<ReadIMul32, 0>;
+def : ReadAdvance<ReadAtomicWA, 0>;
+def : ReadAdvance<ReadAtomicWD, 0>;
+def : ReadAdvance<ReadAtomicDA, 0>;
+def : ReadAdvance<ReadAtomicDD, 0>;
+def : ReadAdvance<ReadAtomicLDW, 0>;
+def : ReadAdvance<ReadAtomicLDD, 0>;
+def : ReadAdvance<ReadAtomicSTW, 0>;
+def : ReadAdvance<ReadAtomicSTD, 0>;
+def : ReadAdvance<ReadFMemBase, 0>;
+def : ReadAdvance<ReadFALU32, 0>;
+def : ReadAdvance<ReadFALU64, 0>;
+def : ReadAdvance<ReadFMul32, 0>;
+def : ReadAdvance<ReadFMulAdd32, 0>;
+def : ReadAdvance<ReadFMulSub32, 0>;
+def : ReadAdvance<ReadFMul64, 0>;
+def : ReadAdvance<ReadFMulAdd64, 0>;
+def : ReadAdvance<ReadFMulSub64, 0>;
+def : ReadAdvance<ReadFDiv32, 0>;
+def : ReadAdvance<ReadFDiv64, 0>;
+def : ReadAdvance<ReadFSqrt32, 0>;
+def : ReadAdvance<ReadFSqrt64, 0>;
+def : ReadAdvance<ReadFCmp32, 0>;
+def : ReadAdvance<ReadFCmp64, 0>;
+def : ReadAdvance<ReadFSGNJ32, 0>;
+def : ReadAdvance<ReadFSGNJ64, 0>;
+def : ReadAdvance<ReadFMinMax32, 0>;
+def : ReadAdvance<ReadFMinMax64, 0>;
+def : ReadAdvance<ReadFCvtF32ToI32, 0>;
+def : ReadAdvance<ReadFCvtF32ToI64, 0>;
+def : ReadAdvance<ReadFCvtF64ToI32, 0>;
+def : ReadAdvance<ReadFCvtF64ToI64, 0>;
+def : ReadAdvance<ReadFCvtI32ToF32, 0>;
+def : ReadAdvance<ReadFCvtI32ToF64, 0>;
+def : ReadAdvance<ReadFCvtI64ToF32, 0>;
+def : ReadAdvance<ReadFCvtI64ToF64, 0>;
+def : ReadAdvance<ReadFCvtF32ToF64, 0>;
+def : ReadAdvance<ReadFCvtF64ToF32, 0>;
+def : ReadAdvance<ReadFMovF32ToI32, 0>;
+def : ReadAdvance<ReadFMovI32ToF32, 0>;
+def : ReadAdvance<ReadFMovF64ToI64, 0>;
+def : ReadAdvance<ReadFMovI64ToF64, 0>;
+def : ReadAdvance<ReadFClass32, 0>;
+def : ReadAdvance<ReadFClass64, 0>;
+}