[llvm-exegesis][NFC] Simplify code.

[courbet]

Summary

What we're redoing already exists in the X86 backend, it's called
X86II::getOperandBias.

[gchatelet]

Nice!

llvm/tools/llvm-exegesis/lib/X86/Target.cpp

 void ExegesisX86Target::fillMemoryOperands(InstructionTemplate &IT,
                                            unsigned Reg,
                                            unsigned Offset) const {
   assert(!isInvalidMemoryInstr(IT.getInstr()) &&
          "fillMemoryOperands requires a valid memory instruction");
   int MemOpIdx = X86II::getMemoryOperandNo(IT.getInstr().Description.TSFlags);
   assert(MemOpIdx >= 0 && "invalid memory operand index");
   // getMemoryOperandNo() ignores tied operands, so we have to add them back.
-  for (unsigned I = 0; I <= static_cast<unsigned>(MemOpIdx); ++I) {
-    const auto &Op = IT.getInstr().Operands[I];
-    if (Op.isTied() && Op.getTiedToIndex() < I) {
-      ++MemOpIdx;
-    }
-  }
+  MemOpIdx += X86II::getOperandBias(IT.getInstr().Description);
   setMemOp(IT, MemOpIdx + 0, MCOperand::createReg(Reg));    // BaseReg
   setMemOp(IT, MemOpIdx + 1, MCOperand::createImm(1));      // ScaleAmt
   setMemOp(IT, MemOpIdx + 2, MCOperand::createReg(0));      // IndexReg
   setMemOp(IT, MemOpIdx + 3, MCOperand::createImm(Offset)); // Disp
   setMemOp(IT, MemOpIdx + 4, MCOperand::createReg(0));      // Segment
 }
 
 void ExegesisX86Target::decrementLoopCounterAndJump(