[RISCV][CodeGen] Account for LMUL for Vector Integer Arithmetic Instructions

[michaelmaitland]

Summary

It is likley that subtargets act differently for a vector integer arithmetic instruction based on the LMUL.
This patch creates seperate SchedRead, SchedWrite, WriteRes, ReadAdvance for each relevant LMUL. It also
introduces the concept of an "UpperBound LMUL" which allows us to describe how an instruction should behave
when the LMUL is unknown. All base instructions use the UpperBound resources because they are not tied to
a specific LMUL. This gives subtargetes the flexibility to describe their own upper bounds on each vector
instruction.

I have a series of patches for the rest of the vector instruction set ready to go, but I would like to first
get feedback on the first one of the series (this one).

[michaelmaitland]

Commit 1fab0ac559a1eb4e74d193c772264b590ad3c317 renamed ReadVIALUCV->ReadVICALUV to match WriteVICALUV.
Needed to update this patch to reflect that.

[michaelmaitland]

Sorry my last update replaced in the wrong direction. This update should do what I meant :)

[michaelmaitland]

I needed to bring back some of the classes that I killed. The subroutines that this change
brings back can be killed at the end of this series of patches.

[craig.topper]

llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td

2394–2408

I think this needs to be indented 2 spaces

2538–2549

Indent

llvm/lib/Target/RISCV/RISCVScheduleV.td

117–119

Is the empty string needed?

[michaelmaitland]

llvm/lib/Target/RISCV/RISCVScheduleV.td

117–119

We do need the double quotes. According to the docs,

The name is parsed in the same special mode used by def. If the name is not included, an unspecified but globally unique name is provided. That is, the following examples end up with different names:
defm    : SomeMultiClass<...>;   // A globally unique name.
defm "" : SomeMultiClass<...>;   // An empty name.

Its no good if we have an empty name because it would generate something like <globally_unique_name>WriteVIALUV_UpperBound when we really want WriteVIALUV_UpperBound

[michaelmaitland]

Fix indentation

[michaelmaitland]

llvm/lib/Target/RISCV/RISCVScheduleV.td

117–119

Meant to say no good if we have no quotes (i.e. a globally unique name)

[craig.topper]

LGTM

llvm/lib/Target/RISCV/RISCVInstrInfoV.td

foreach n = EEWList in {
     def VSOXEI # n # _V :
       VIndexedStore<MOPSTIndexedOrder, w, "vsoxei" # n # ".v">,
       VSXSched<n, "O">;
   }
 }
 
 multiclass VALU_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIALUV, ReadVIALUV, ReadVIALUV, ReadVMask]>;
+           Sched<[WriteVIALUV_UpperBound, ReadVIALUV_UpperBound,
+             ReadVIALUV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIALUX, ReadVIALUV, ReadVIALUX, ReadVMask]>;
+           Sched<[WriteVIALUX_UpperBound, ReadVIALUV_UpperBound,
+             ReadVIALUX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVIALUI, ReadVIALUV, ReadVMask]>;
+           Sched<[WriteVIALUI_UpperBound, ReadVIALUV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VALU_IV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIALUV, ReadVIALUV, ReadVIALUV, ReadVMask]>;
+           Sched<[WriteVIALUV_UpperBound, ReadVIALUV_UpperBound,
+             ReadVIALUV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIALUX, ReadVIALUV, ReadVIALUX, ReadVMask]>;
+           Sched<[WriteVIALUX_UpperBound, ReadVIALUV_UpperBound,
+             ReadVIALUX_UpperBound, ReadVMask]>;
 }
 
 multiclass VALU_IV_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIALUV, ReadVIALUV, ReadVIALUX, ReadVMask]>;
+           Sched<[WriteVIALUV_UpperBound, ReadVIALUV_UpperBound,
+             ReadVIALUX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVIALUI, ReadVIALUV, ReadVMask]>;
+           Sched<[WriteVIALUI_UpperBound, ReadVIALUV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VALU_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIWALUV, ReadVIWALUV, ReadVIWALUV, ReadVMask]>;
+           Sched<[WriteVIWALUV_UpperBound, ReadVIWALUV_UpperBound,
+             ReadVIWALUV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIWALUX, ReadVIWALUV, ReadVIWALUX, ReadVMask]>;
+           Sched<[WriteVIWALUX_UpperBound, ReadVIWALUV_UpperBound,
+             ReadVIWALUX_UpperBound, ReadVMask]>;
 }
 
 multiclass VMAC_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V : VALUrVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-          Sched<[WriteVIMulAddV, ReadVIMulAddV, ReadVIMulAddV, ReadVMask]>;
+          Sched<[WriteVIMulAddV_UpperBound, ReadVIMulAddV_UpperBound,
+            ReadVIMulAddV_UpperBound, ReadVMask]>;
   def X : VALUrVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-          Sched<[WriteVIMulAddX, ReadVIMulAddV, ReadVIMulAddX, ReadVMask]>;
+          Sched<[WriteVIMulAddX_UpperBound, ReadVIMulAddV_UpperBound,
+            ReadVIMulAddX_UpperBound, ReadVMask]>;
 }
 
 multiclass VWMAC_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V : VALUrVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-          Sched<[WriteVIWMulAddV, ReadVIWMulAddV, ReadVIWMulAddV, ReadVMask]>;
+          Sched<[WriteVIWMulAddV_UpperBound, ReadVIWMulAddV_UpperBound,
+            ReadVIWMulAddV_UpperBound, ReadVMask]>;
   def X : VALUrVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-          Sched<[WriteVIWMulAddX, ReadVIWMulAddV, ReadVIWMulAddX, ReadVMask]>;
+          Sched<[WriteVIWMulAddX_UpperBound, ReadVIWMulAddV_UpperBound,
+            ReadVIWMulAddX_UpperBound, ReadVMask]>;
 }
 
 multiclass VWMAC_MV_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def X : VALUrVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-          Sched<[WriteVIWMulAddX, ReadVIWMulAddV, ReadVIWMulAddX, ReadVMask]>;
+          Sched<[WriteVIWMulAddX_UpperBound, ReadVIWMulAddV_UpperBound,
+            ReadVIWMulAddX_UpperBound, ReadVMask]>;
 }
 
 multiclass VALU_MV_VS2<string opcodestr, bits<6> funct6, bits<5> vs1> {
   def "" : VALUVs2<funct6, vs1, OPMVV, opcodestr>,
-           Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
+           Sched<[WriteVExtV_UpperBound, ReadVExtV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VALUm_IV_V_X_I<string opcodestr, bits<6> funct6> {
   def VM : VALUmVV<funct6, OPIVV, opcodestr # ".vvm">,
-           Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
+           Sched<[WriteVICALUV_UpperBound, ReadVICALUV_UpperBound,
+             ReadVICALUV_UpperBound, ReadVMask]>;
   def XM : VALUmVX<funct6, OPIVX, opcodestr # ".vxm">,
-           Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
+           Sched<[WriteVICALUX_UpperBound, ReadVICALUV_UpperBound,
+             ReadVICALUX_UpperBound, ReadVMask]>;
   def IM : VALUmVI<funct6, opcodestr # ".vim">,
-           Sched<[WriteVICALUI, ReadVICALUV, ReadVMask]>;
+           Sched<[WriteVICALUI_UpperBound, ReadVICALUV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VMRG_IV_V_X_I<string opcodestr, bits<6> funct6> {
   def VM : VALUmVV<funct6, OPIVV, opcodestr # ".vvm">,
-           Sched<[WriteVIMergeV, ReadVIMergeV, ReadVIMergeV, ReadVMask]>;
+           Sched<[WriteVIMergeV_UpperBound, ReadVIMergeV_UpperBound,
+             ReadVIMergeV_UpperBound, ReadVMask]>;
   def XM : VALUmVX<funct6, OPIVX, opcodestr # ".vxm">,
-           Sched<[WriteVIMergeX, ReadVIMergeV, ReadVIMergeX, ReadVMask]>;
+           Sched<[WriteVIMergeX_UpperBound, ReadVIMergeV_UpperBound,
+             ReadVIMergeX_UpperBound, ReadVMask]>;
   def IM : VALUmVI<funct6, opcodestr # ".vim">,
-           Sched<[WriteVIMergeI, ReadVIMergeV, ReadVMask]>;
+           Sched<[WriteVIMergeI_UpperBound, ReadVIMergeV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VALUm_IV_V_X<string opcodestr, bits<6> funct6> {
   def VM : VALUmVV<funct6, OPIVV, opcodestr # ".vvm">,
-           Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
+           Sched<[WriteVICALUV_UpperBound, ReadVICALUV_UpperBound,
+             ReadVICALUV_UpperBound, ReadVMask]>;
   def XM : VALUmVX<funct6, OPIVX, opcodestr # ".vxm">,
-           Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
+           Sched<[WriteVICALUX_UpperBound, ReadVICALUV_UpperBound,
+             ReadVICALUX_UpperBound, ReadVMask]>;
 }
 
 multiclass VALUNoVm_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5> {
   def V : VALUVVNoVm<funct6, OPIVV, opcodestr # ".vv">,
-          Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV]>;
+          Sched<[WriteVICALUV_UpperBound, ReadVICALUV_UpperBound,
+            ReadVICALUV_UpperBound]>;
   def X : VALUVXNoVm<funct6, OPIVX, opcodestr # ".vx">,
-          Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX]>;
+          Sched<[WriteVICALUX_UpperBound, ReadVICALUV_UpperBound
+            , ReadVICALUX_UpperBound]>;
   def I : VALUVINoVm<funct6, opcodestr # ".vi", optype>,
-          Sched<[WriteVICALUI, ReadVICALUV]>;
+          Sched<[WriteVICALUI_UpperBound, ReadVICALUV_UpperBound]>;
 }
 
 multiclass VALUNoVm_IV_V_X<string opcodestr, bits<6> funct6> {
   def V : VALUVVNoVm<funct6, OPIVV, opcodestr # ".vv">,
-          Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV]>;
+          Sched<[WriteVICALUV_UpperBound, ReadVICALUV_UpperBound,
+            ReadVICALUV_UpperBound]>;
   def X : VALUVXNoVm<funct6, OPIVX, opcodestr # ".vx">,
-          Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX]>;
+          Sched<[WriteVICALUX_UpperBound, ReadVICALUV_UpperBound,
+            ReadVICALUX_UpperBound]>;
 }
 
 multiclass VALU_FV_V_F<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V : VALUVV<funct6, OPFVV, opcodestr # "." # vw # "v">,
           Sched<[WriteVFALUV, ReadVFALUV, ReadVFALUV, ReadVMask]>;
   def F : VALUVF<funct6, OPFVF, opcodestr # "." # vw # "f">,
           Sched<[WriteVFALUF, ReadVFALUV, ReadVFALUF, ReadVMask]>;
 }

 
 multiclass VMIOT_MV_V<string opcodestr, bits<6> funct6, bits<5> vs1> {
   def "" : VALUVs2<funct6, vs1, OPMVV, opcodestr>,
            Sched<[WriteVMIotV, ReadVMIotV, ReadVMask]>;
 }
 
 multiclass VSHT_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVShiftV, ReadVShiftV, ReadVShiftV, ReadVMask]>;
+           Sched<[WriteVShiftV_UpperBound, ReadVShiftV_UpperBound,
+             ReadVShiftV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVShiftX, ReadVShiftV, ReadVShiftX, ReadVMask]>;
+           Sched<[WriteVShiftX_UpperBound, ReadVShiftV_UpperBound,
+             ReadVShiftX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVShiftI, ReadVShiftV, ReadVMask]>;
+           Sched<[WriteVShiftI_UpperBound, ReadVShiftV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VNSHT_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVNShiftV, ReadVNShiftV, ReadVNShiftV, ReadVMask]>;
+           Sched<[WriteVNShiftV_UpperBound, ReadVNShiftV_UpperBound,
+             ReadVNShiftV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVNShiftX, ReadVNShiftV, ReadVNShiftX, ReadVMask]>;
+           Sched<[WriteVNShiftX_UpperBound, ReadVNShiftV_UpperBound,
+             ReadVNShiftX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVNShiftI, ReadVNShiftV, ReadVMask]>;
+           Sched<[WriteVNShiftI_UpperBound, ReadVNShiftV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VCMP_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpV, ReadVMask]>;
+           Sched<[WriteVICmpV_UpperBound, ReadVICmpV_UpperBound,
+             ReadVICmpV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
+           Sched<[WriteVICmpX_UpperBound, ReadVICmpV_UpperBound,
+             ReadVICmpX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVICmpI, ReadVICmpV, ReadVMask]>;
+           Sched<[WriteVICmpI_UpperBound, ReadVICmpV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VCMP_IV_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpX, ReadVMask]>;
+           Sched<[WriteVICmpV_UpperBound, ReadVICmpV_UpperBound,
+             ReadVICmpX_UpperBound, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,
-           Sched<[WriteVICmpI, ReadVICmpV, ReadVMask]>;
+           Sched<[WriteVICmpI_UpperBound, ReadVICmpV_UpperBound,
+             ReadVMask]>;
 }
 
 multiclass VCMP_IV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpV, ReadVMask]>;
+           Sched<[WriteVICmpV_UpperBound, ReadVICmpV_UpperBound,
+             ReadVICmpV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
+           Sched<[WriteVICmpX_UpperBound, ReadVICmpV_UpperBound,
+             ReadVICmpX_UpperBound, ReadVMask]>;
 }
 
 multiclass VMUL_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIMulV, ReadVIMulV, ReadVIMulV, ReadVMask]>;
+           Sched<[WriteVIMulV_UpperBound, ReadVIMulV_UpperBound,
+             ReadVIMulV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIMulX, ReadVIMulV, ReadVIMulX, ReadVMask]>;
+           Sched<[WriteVIMulX_UpperBound, ReadVIMulV_UpperBound,
+             ReadVIMulX_UpperBound, ReadVMask]>;
 }
 
 multiclass VWMUL_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIWMulV, ReadVIWMulV, ReadVIWMulV, ReadVMask]>;
+           Sched<[WriteVIWMulV_UpperBound, ReadVIWMulV_UpperBound,
+             ReadVIWMulV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIWMulX, ReadVIWMulV, ReadVIWMulX, ReadVMask]>;
+           Sched<[WriteVIWMulX_UpperBound, ReadVIWMulV_UpperBound,
+             ReadVIWMulX_UpperBound, ReadVMask]>;
 }
 
 multiclass VDIV_MV_V_X<string opcodestr, bits<6> funct6, string vw = "v"> {
   def V  : VALUVV<funct6, OPMVV, opcodestr # "." # vw # "v">,
-           Sched<[WriteVIDivV, ReadVIDivV, ReadVIDivV, ReadVMask]>;
+           Sched<[WriteVIDivV_UpperBound, ReadVIDivV_UpperBound,
+             ReadVIDivV_UpperBound, ReadVMask]>;
   def X  : VALUVX<funct6, OPMVX, opcodestr # "." # vw # "x">,
-           Sched<[WriteVIDivX, ReadVIDivV, ReadVIDivX, ReadVMask]>;
+           Sched<[WriteVIDivX_UpperBound, ReadVIDivV_UpperBound,
+             ReadVIDivX_UpperBound, ReadVMask]>;
 }
 
 multiclass VSALU_IV_V_X_I<string opcodestr, bits<6> funct6, Operand optype = simm5, string vw = "v"> {
   def V  : VALUVV<funct6, OPIVV, opcodestr # "." # vw # "v">,
            Sched<[WriteVSALUV, ReadVSALUV, ReadVSALUV, ReadVMask]>;
   def X  : VALUVX<funct6, OPIVX, opcodestr # "." # vw # "x">,
            Sched<[WriteVSALUX, ReadVSALUV, ReadVSALUX, ReadVMask]>;
   def I  : VALUVI<funct6, opcodestr # "." # vw # "i", optype>,

 defm VMERGE_V : VMRG_IV_V_X_I<"vmerge", 0b010111>;
 
 // Vector Integer Move Instructions
 let hasSideEffects = 0, mayLoad = 0, mayStore = 0, vs2 = 0, vm = 1,
     RVVConstraint = NoConstraint  in {
 // op vd, vs1
 def VMV_V_V : RVInstVV<0b010111, OPIVV, (outs VR:$vd),
                        (ins VR:$vs1), "vmv.v.v", "$vd, $vs1">,
-              Sched<[WriteVIMovV, ReadVIMovV]>;
+              Sched<[WriteVIMovV_UpperBound, ReadVIMovV_UpperBound]>;
 // op vd, rs1
 def VMV_V_X : RVInstVX<0b010111, OPIVX, (outs VR:$vd),
                        (ins GPR:$rs1), "vmv.v.x", "$vd, $rs1">,
-              Sched<[WriteVIMovX, ReadVIMovX]>;
+              Sched<[WriteVIMovX_UpperBound, ReadVIMovX_UpperBound]>;
 // op vd, imm
 def VMV_V_I : RVInstIVI<0b010111, (outs VR:$vd),
                        (ins simm5:$imm), "vmv.v.i", "$vd, $imm">,
-              Sched<[WriteVIMovI]>;
+              Sched<[WriteVIMovI_UpperBound]>;
 } // hasSideEffects = 0, mayLoad = 0, mayStore = 0
 
 // Vector Fixed-Point Arithmetic Instructions
 defm VSADDU_V : VSALU_IV_V_X_I<"vsaddu", 0b100000>;
 defm VSADD_V : VSALU_IV_V_X_I<"vsadd", 0b100001>;
 defm VSSUBU_V : VSALU_IV_V_X<"vssubu", 0b100010>;
 defm VSSUB_V : VSALU_IV_V_X<"vssub", 0b100011>;
 

llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td

multiclass VPseudoTiedBinary<VReg RetClass,
   }
 }
 
 multiclass VPseudoBinaryV_VV<string Constraint = ""> {
   foreach m = MxList in
     defm _VV : VPseudoBinary<m.vrclass, m.vrclass, m.vrclass, m, Constraint>;
 }
 
+multiclass VPseudoBinaryV_VV_LMUL<LMULInfo m, string Constraint = ""> {
+  defm _VV : VPseudoBinary<m.vrclass, m.vrclass, m.vrclass, m, Constraint>;
+}
+
 // Similar to VPseudoBinaryV_VV, but uses MxListF.
 multiclass VPseudoBinaryFV_VV<string Constraint = ""> {
   foreach m = MxListF in
     defm _VV : VPseudoBinary<m.vrclass, m.vrclass, m.vrclass, m, Constraint>;
 }
 
 multiclass VPseudoVGTR_VV_EEW<int eew, string Constraint = ""> {
   foreach m = MxList in {
multiclass VPseudoVGTR_VV_EEW<int eew, string Constraint = ""> {
   }
 }
 
 multiclass VPseudoBinaryV_VX<string Constraint = ""> {
   foreach m = MxList in
     defm "_VX" : VPseudoBinary<m.vrclass, m.vrclass, GPR, m, Constraint>;
 }
 
+multiclass VPseudoBinaryV_VX_LMUL<LMULInfo m, string Constraint = ""> {
+  defm "_VX" : VPseudoBinary<m.vrclass, m.vrclass, GPR, m, Constraint>;
+}
+
 multiclass VPseudoVSLD1_VX<string Constraint = ""> {
   foreach m = MxList in
     defm "_VX" : VPseudoBinary<m.vrclass, m.vrclass, GPR, m, Constraint>,
                  Sched<[WriteVISlide1X, ReadVISlideV, ReadVISlideX, ReadVMask]>;
 }
 
 multiclass VPseudoBinaryV_VF<string Constraint = ""> {
   foreach f = FPList in
foreach m = f.MxList in
         Sched<[WriteVFSlide1F, ReadVFSlideV, ReadVFSlideF, ReadVMask]>;
 }
 
 multiclass VPseudoBinaryV_VI<Operand ImmType = simm5, string Constraint = ""> {
   foreach m = MxList in
     defm _VI : VPseudoBinary<m.vrclass, m.vrclass, ImmType, m, Constraint>;
 }
 
+multiclass VPseudoBinaryV_VI_LMUL<Operand ImmType = simm5, LMULInfo m, string Constraint = ""> {
+  defm _VI : VPseudoBinary<m.vrclass, m.vrclass, ImmType, m, Constraint>;
+}
+
 multiclass VPseudoVALU_MM {
   foreach m = MxList in
     let VLMul = m.value in {
       def "_MM_" # m.MX : VPseudoBinaryNoMask<VR, VR, VR, "", /*DummyMask*/0>,
                           Sched<[WriteVMALUV, ReadVMALUV, ReadVMALUV]>;
     }
 }
 
 // We use earlyclobber here due to
 // * The destination EEW is smaller than the source EEW and the overlap is
 //   in the lowest-numbered part of the source register group is legal.
 //   Otherwise, it is illegal.
 // * The destination EEW is greater than the source EEW, the source EMUL is
 //   at least 1, and the overlap is in the highest-numbered part of the
 //   destination register group is legal. Otherwise, it is illegal.
 multiclass VPseudoBinaryW_VV<list<LMULInfo> mxlist = MxListW> {
   foreach m = mxlist in
     defm _VV : VPseudoBinary<m.wvrclass, m.vrclass, m.vrclass, m,
                              "@earlyclobber $rd">;
 }
 
-multiclass VPseudoBinaryW_VX {
-  foreach m = MxListW in
-    defm "_VX" : VPseudoBinary<m.wvrclass, m.vrclass, GPR, m,
-                               "@earlyclobber $rd">;
+multiclass VPseudoBinaryW_VV_LMUL<LMULInfo m> {
+  defm _VV : VPseudoBinary<m.wvrclass, m.vrclass, m.vrclass, m,
+                           "@earlyclobber $rd">;
+}
+
+multiclass VPseudoBinaryW_VX_LMUL<LMULInfo m> {
 }
 
 multiclass VPseudoBinaryW_VF {
   foreach f = FPListW in
     foreach m = f.MxList in
       defm "_V" # f.FX : VPseudoBinary<m.wvrclass, m.vrclass,
                                        f.fprclass, m,
                                        "@earlyclobber $rd">;
 }
 
 multiclass VPseudoBinaryW_WV<list<LMULInfo> mxlist = MxListW> {
   foreach m = mxlist in {
     defm _WV : VPseudoBinary<m.wvrclass, m.wvrclass, m.vrclass, m,
                              "@earlyclobber $rd">;
     defm _WV : VPseudoTiedBinary<m.wvrclass, m.vrclass, m,
                                  "@earlyclobber $rd">;
   }
 }
 
-multiclass VPseudoBinaryW_WX {
-  foreach m = MxListW in
-    defm "_WX" : VPseudoBinary<m.wvrclass, m.wvrclass, GPR, m>;
+multiclass VPseudoBinaryW_WV_LMUL<LMULInfo m> {
+  defm _WV : VPseudoBinary<m.wvrclass, m.wvrclass, m.vrclass, m,
+                           "@earlyclobber $rd">;
+  defm _WV : VPseudoTiedBinary<m.wvrclass, m.vrclass, m,
+                               "@earlyclobber $rd">;
+}
+
+multiclass VPseudoBinaryW_WX_LMUL<LMULInfo m> {
 }
 
 multiclass VPseudoBinaryW_WF {
   foreach f = FPListW in
     foreach m = f.MxList in
       defm "_W" # f.FX : VPseudoBinary<m.wvrclass, m.wvrclass,
                                        f.fprclass, m>;
 }
 
 // Narrowing instructions like vnsrl/vnsra/vnclip(u) don't need @earlyclobber
 // if the source and destination have an LMUL<=1. This matches this overlap
 // exception from the spec.
 // "The destination EEW is smaller than the source EEW and the overlap is in the
 //  lowest-numbered part of the source register group."
 multiclass VPseudoBinaryV_WV {
   foreach m = MxListW in
     defm _WV : VPseudoBinary<m.vrclass, m.wvrclass, m.vrclass, m,
                              !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
 }
 
+multiclass VPseudoBinaryV_WV_LMUL<LMULInfo m> {
+  defm _WV : VPseudoBinary<m.vrclass, m.wvrclass, m.vrclass, m,
+                           !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
+}
+
 multiclass VPseudoBinaryV_WX {
   foreach m = MxListW in
     defm _WX : VPseudoBinary<m.vrclass, m.wvrclass, GPR, m,
                              !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
 }
 
+multiclass VPseudoBinaryV_WX_LMUL<LMULInfo m> {
+  defm _WX : VPseudoBinary<m.vrclass, m.wvrclass, GPR, m,
+                           !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
+}
+
 multiclass VPseudoBinaryV_WI {
   foreach m = MxListW in
     defm _WI : VPseudoBinary<m.vrclass, m.wvrclass, uimm5, m,
                              !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
 }
 
+multiclass VPseudoBinaryV_WI_LMUL<LMULInfo m> {
+  defm _WI : VPseudoBinary<m.vrclass, m.wvrclass, uimm5, m,
+                           !if(!ge(m.octuple, 8), "@earlyclobber $rd", "")>;
+}
+
 // For vadc and vsbc, the instruction encoding is reserved if the destination
 // vector register is v0.
 // For vadc and vsbc, CarryIn == 1 and CarryOut == 0
-multiclass VPseudoBinaryV_VM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoBinaryV_VM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                              string Constraint = ""> {
-  foreach m = MxList in
-    def "_VV" # !if(CarryIn, "M", "") # "_" # m.MX :
-      VPseudoBinaryCarryIn<!if(CarryOut, VR,
-                           !if(!and(CarryIn, !not(CarryOut)),
-                               GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                           m.vrclass, m.vrclass, m, CarryIn, Constraint>;
 }
 
-multiclass VPseudoTiedBinaryV_VM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoTiedBinaryV_VM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                                  string Constraint = ""> {
-  foreach m = MxList in
-    def "_VV" # !if(CarryIn, "M", "") # "_" # m.MX # "_TU" :
-      VPseudoTiedBinaryCarryIn<!if(CarryOut, VR,
-                               !if(!and(CarryIn, !not(CarryOut)),
-                                   GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                               m.vrclass, m.vrclass, m, CarryIn, Constraint>;
 }
 
-multiclass VPseudoBinaryV_XM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoBinaryV_XM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                              string Constraint = ""> {
-  foreach m = MxList in
-    def "_VX" # !if(CarryIn, "M", "") # "_" # m.MX :
-      VPseudoBinaryCarryIn<!if(CarryOut, VR,
-                           !if(!and(CarryIn, !not(CarryOut)),
-                               GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                           m.vrclass, GPR, m, CarryIn, Constraint>;
 }
 
-multiclass VPseudoTiedBinaryV_XM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoTiedBinaryV_XM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                                  string Constraint = ""> {
-  foreach m = MxList in
-    def "_VX" # !if(CarryIn, "M", "") # "_" # m.MX # "_TU":
-      VPseudoTiedBinaryCarryIn<!if(CarryOut, VR,
-                               !if(!and(CarryIn, !not(CarryOut)),
-                                   GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                               m.vrclass, GPR, m, CarryIn, Constraint>;
 }
 
 multiclass VPseudoVMRG_FM {
   foreach f = FPList in
     foreach m = f.MxList in {
       def "_V" # f.FX # "M_" # m.MX :
         VPseudoBinaryCarryIn<GetVRegNoV0<m.vrclass>.R,
                              m.vrclass, f.fprclass, m, /*CarryIn=*/1, "">,
         Sched<[WriteVFMergeV, ReadVFMergeV, ReadVFMergeF, ReadVMask]>;
       // Tied version to allow codegen control over the tail elements
       def "_V" # f.FX # "M_" # m.MX # "_TU":
         VPseudoTiedBinaryCarryIn<GetVRegNoV0<m.vrclass>.R,
                                  m.vrclass, f.fprclass, m, /*CarryIn=*/1, "">,
         Sched<[WriteVFMergeV, ReadVFMergeV, ReadVFMergeF, ReadVMask]>;
     }
 }
 
-multiclass VPseudoBinaryV_IM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoBinaryV_IM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                              string Constraint = ""> {
-  foreach m = MxList in
-    def "_VI" # !if(CarryIn, "M", "") # "_" # m.MX :
-      VPseudoBinaryCarryIn<!if(CarryOut, VR,
-                           !if(!and(CarryIn, !not(CarryOut)),
-                               GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                           m.vrclass, simm5, m, CarryIn, Constraint>;
 }
 
-multiclass VPseudoTiedBinaryV_IM<bit CarryOut = 0, bit CarryIn = 1,
+multiclass VPseudoTiedBinaryV_IM<LMULInfo m, bit CarryOut = 0, bit CarryIn = 1,
                                  string Constraint = ""> {
-  foreach m = MxList in
-    def "_VI" # !if(CarryIn, "M", "") # "_" # m.MX # "_TU":
-      VPseudoTiedBinaryCarryIn<!if(CarryOut, VR,
-                               !if(!and(CarryIn, !not(CarryOut)),
-                                   GetVRegNoV0<m.vrclass>.R, m.vrclass)),
-                               m.vrclass, simm5, m, CarryIn, Constraint>;
 }
 
 multiclass VPseudoUnaryVMV_V_X_I {
   foreach m = MxList in {
     let VLMul = m.value in {
-      def "_V_" # m.MX : VPseudoUnaryNoDummyMask<m.vrclass, m.vrclass>,
-                         Sched<[WriteVIMovV, ReadVIMovV]>;
-      def "_X_" # m.MX : VPseudoUnaryNoDummyMask<m.vrclass, GPR>,
-                         Sched<[WriteVIMovX, ReadVIMovX]>;
-      def "_I_" # m.MX : VPseudoUnaryNoDummyMask<m.vrclass, simm5>,
-                         Sched<[WriteVIMovI]>;
-      def "_V_" # m.MX # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, m.vrclass>,
-                         Sched<[WriteVIMovV, ReadVIMovV]>;
-      def "_X_" # m.MX # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, GPR>,
-                         Sched<[WriteVIMovX, ReadVIMovX]>;
-      def "_I_" # m.MX # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, simm5>,
-                         Sched<[WriteVIMovI]>;
+      defvar mx = m.MX;
+      defvar WriteVIMovV_MX = !cast<SchedWrite>("WriteVIMovV_" # mx);
+      defvar WriteVIMovX_MX = !cast<SchedWrite>("WriteVIMovX_" # mx);
+      defvar WriteVIMovI_MX = !cast<SchedWrite>("WriteVIMovI_" # mx);
+      defvar ReadVIMovV_MX = !cast<SchedRead>("ReadVIMovV_" # mx);
+      defvar ReadVIMovX_MX = !cast<SchedRead>("ReadVIMovX_" # mx);
+
+      let VLMul = m.value in {
+        def "_V_" # mx : VPseudoUnaryNoDummyMask<m.vrclass, m.vrclass>,
+                           Sched<[WriteVIMovV_MX, ReadVIMovV_MX]>;
+        def "_X_" # mx : VPseudoUnaryNoDummyMask<m.vrclass, GPR>,
+                           Sched<[WriteVIMovX_MX, ReadVIMovX_MX]>;
+        def "_I_" # mx : VPseudoUnaryNoDummyMask<m.vrclass, simm5>,
+                           Sched<[WriteVIMovI_MX]>;
+        def "_V_" # mx # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, m.vrclass>,
+                           Sched<[WriteVIMovV_MX, ReadVIMovV_MX]>;
+        def "_X_" # mx # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, GPR>,
+                           Sched<[WriteVIMovX_MX, ReadVIMovX_MX]>;
+        def "_I_" # mx # "_TU": VPseudoUnaryNoDummyMaskTU<m.vrclass, simm5>,
+                           Sched<[WriteVIMovI_MX]>;
+      }
     }
   }
 }
 
 multiclass VPseudoVMV_F {
   foreach f = FPList in {
     foreach m = f.MxList in {
       let VLMul = m.value in {
foreach m = MxListF in {
     }
   }
 }
 
 multiclass PseudoVEXT_VF2 {
   defvar constraints = "@earlyclobber $rd";
   foreach m = MxListVF2 in
   {
+    defvar mx = m.MX;
+    defvar WriteVExtV_MX = !cast<SchedWrite>("WriteVExtV_" # mx);
+    defvar ReadVExtV_MX = !cast<SchedRead>("ReadVExtV_" # mx);
+
     let VLMul = m.value in {
-      def "_" # m.MX : VPseudoUnaryNoMask<m.vrclass, m.f2vrclass, constraints>,
-                       Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f2vrclass, constraints>,
-                              Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_MASK" :
+      def "_" # mx : VPseudoUnaryNoMask<m.vrclass, m.f2vrclass, constraints>,
+                       Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f2vrclass, constraints>,
+                              Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_MASK" :
         VPseudoUnaryMaskTA<m.vrclass, m.f2vrclass, constraints>,
         RISCVMaskedPseudo</*MaskOpIdx*/ 2>,
-        Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
+        Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
     }
   }
 }
 
 multiclass PseudoVEXT_VF4 {
   defvar constraints = "@earlyclobber $rd";
   foreach m = MxListVF4 in
   {
+    defvar mx = m.MX;
+    defvar WriteVExtV_MX = !cast<SchedWrite>("WriteVExtV_" # mx);
+    defvar ReadVExtV_MX = !cast<SchedRead>("ReadVExtV_" # mx);
+
     let VLMul = m.value in {
-      def "_" # m.MX : VPseudoUnaryNoMask<m.vrclass, m.f4vrclass, constraints>,
-                       Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f4vrclass, constraints>,
-                              Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_MASK" :
+      def "_" # mx : VPseudoUnaryNoMask<m.vrclass, m.f4vrclass, constraints>,
+                       Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f4vrclass, constraints>,
+                              Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_MASK" :
         VPseudoUnaryMaskTA<m.vrclass, m.f4vrclass, constraints>,
         RISCVMaskedPseudo</*MaskOpIdx*/ 2>,
-        Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
+        Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
     }
   }
 }
 
 multiclass PseudoVEXT_VF8 {
   defvar constraints = "@earlyclobber $rd";
   foreach m = MxListVF8 in
   {
+    defvar mx = m.MX;
+    defvar WriteVExtV_MX = !cast<SchedWrite>("WriteVExtV_" # mx);
+    defvar ReadVExtV_MX = !cast<SchedRead>("ReadVExtV_" # mx);
+
     let VLMul = m.value in {
-      def "_" # m.MX : VPseudoUnaryNoMask<m.vrclass, m.f8vrclass, constraints>,
-                       Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f8vrclass, constraints>,
-                              Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
-      def "_" # m.MX # "_MASK" :
+      def "_" # mx : VPseudoUnaryNoMask<m.vrclass, m.f8vrclass, constraints>,
+                       Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_TU": VPseudoUnaryNoMaskTU<m.vrclass, m.f8vrclass, constraints>,
+                              Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
+      def "_" # mx # "_MASK" :
         VPseudoUnaryMaskTA<m.vrclass, m.f8vrclass, constraints>,
         RISCVMaskedPseudo</*MaskOpIdx*/ 2>,
-        Sched<[WriteVExtV, ReadVExtV, ReadVMask]>;
+        Sched<[WriteVExtV_MX, ReadVExtV_MX, ReadVMask]>;
     }
   }
 }
 
 // The destination EEW is 1 since "For the purposes of register group overlap
 // constraints, mask elements have EEW=1."
 // The source EEW is 8, 16, 32, or 64.
 // When the destination EEW is different from source EEW, we need to use
 // @earlyclobber to avoid the overlap between destination and source registers.
 // We don't need @earlyclobber for LMUL<=1 since that matches this overlap
 // exception from the spec
 // "The destination EEW is smaller than the source EEW and the overlap is in the
 //  lowest-numbered part of the source register group".
 // With LMUL<=1 the source and dest occupy a single register so any overlap
 // is in the lowest-numbered part.
 multiclass VPseudoBinaryM_VV<list<LMULInfo> mxlist = MxList> {
   foreach m = mxlist in
     defm _VV : VPseudoBinaryM<VR, m.vrclass, m.vrclass, m,
                               !if(!ge(m.octuple, 16), "@earlyclobber $rd", "")>;
 }
 
-multiclass VPseudoBinaryM_VX {
-  foreach m = MxList in
-    defm "_VX" :
-      VPseudoBinaryM<VR, m.vrclass, GPR, m,
-                     !if(!ge(m.octuple, 16), "@earlyclobber $rd", "")>;
+multiclass VPseudoBinaryM_VV_LMUL<LMULInfo m> {
+  defm _VV : VPseudoBinaryM<VR, m.vrclass, m.vrclass, m,
+                            !if(!ge(m.octuple, 16), "@earlyclobber $rd", "")>;
+}
+
+multiclass VPseudoBinaryM_VX<LMULInfo m> {
 }
 
 multiclass VPseudoBinaryM_VF {
   foreach f = FPList in
     foreach m = f.MxList in
       defm "_V" # f.FX :
         VPseudoBinaryM<VR, m.vrclass, f.fprclass, m,
                        !if(!ge(m.octuple, 16), "@earlyclobber $rd", "")>;
 }
 
-multiclass VPseudoBinaryM_VI {
-  foreach m = MxList in
-    defm _VI : VPseudoBinaryM<VR, m.vrclass, simm5, m,
-                              !if(!ge(m.octuple, 16), "@earlyclobber $rd", "")>;
+multiclass VPseudoBinaryM_VI<LMULInfo m> {
 }
 
 multiclass VPseudoVGTR_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
   defm "" : VPseudoBinaryV_VV<Constraint>,
             Sched<[WriteVGatherV, ReadVGatherV, ReadVGatherV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX<Constraint>,
             Sched<[WriteVGatherX, ReadVGatherV, ReadVGatherX, ReadVMask]>;
   defm "" : VPseudoBinaryV_VI<ImmType, Constraint>,
             Sched<[WriteVGatherI, ReadVGatherV, ReadVMask]>;
 }
 
 multiclass VPseudoVSALU_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
   defm "" : VPseudoBinaryV_VV<Constraint>,
             Sched<[WriteVSALUV, ReadVSALUV, ReadVSALUV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX<Constraint>,
             Sched<[WriteVSALUX, ReadVSALUV, ReadVSALUX, ReadVMask]>;
   defm "" : VPseudoBinaryV_VI<ImmType, Constraint>,
             Sched<[WriteVSALUI, ReadVSALUV, ReadVMask]>;
 }
 
 
 multiclass VPseudoVSHT_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
-  defm "" : VPseudoBinaryV_VV<Constraint>,
-            Sched<[WriteVShiftV, ReadVShiftV, ReadVShiftV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX<Constraint>,
-            Sched<[WriteVShiftX, ReadVShiftV, ReadVShiftX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VI<ImmType, Constraint>,
-            Sched<[WriteVShiftI, ReadVShiftV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVShiftV_MX = !cast<SchedWrite>("WriteVShiftV_" # mx);
+    defvar WriteVShiftX_MX = !cast<SchedWrite>("WriteVShiftX_" # mx);
+    defvar WriteVShiftI_MX = !cast<SchedWrite>("WriteVShiftI_" # mx);
+    defvar ReadVShiftV_MX = !cast<SchedRead>("ReadVShiftV_" # mx);
+    defvar ReadVShiftX_MX = !cast<SchedRead>("ReadVShiftX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m, Constraint>,
+              Sched<[WriteVShiftV_MX, ReadVShiftV_MX, ReadVShiftV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m, Constraint>,
+              Sched<[WriteVShiftX_MX, ReadVShiftV_MX, ReadVShiftX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VI_LMUL<ImmType, m, Constraint>,
+              Sched<[WriteVShiftI_MX, ReadVShiftV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVSSHT_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
   defm "" : VPseudoBinaryV_VV<Constraint>,
             Sched<[WriteVSShiftV, ReadVSShiftV, ReadVSShiftV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX<Constraint>,
             Sched<[WriteVSShiftX, ReadVSShiftV, ReadVSShiftX, ReadVMask]>;
   defm "" : VPseudoBinaryV_VI<ImmType, Constraint>,
             Sched<[WriteVSShiftI, ReadVSShiftV, ReadVMask]>;
 }
 
 multiclass VPseudoVALU_VV_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
-  defm "" : VPseudoBinaryV_VV<Constraint>,
-            Sched<[WriteVIALUV, ReadVIALUV, ReadVIALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX<Constraint>,
-            Sched<[WriteVIALUX, ReadVIALUV, ReadVIALUX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VI<ImmType, Constraint>,
-            Sched<[WriteVIALUI, ReadVIALUV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIALUV_MX = !cast<SchedWrite>("WriteVIALUV_" # mx);
+    defvar WriteVIALUX_MX = !cast<SchedWrite>("WriteVIALUX_" # mx);
+    defvar WriteVIALUI_MX = !cast<SchedWrite>("WriteVIALUI_" # mx);
+    defvar ReadVIALUV_MX = !cast<SchedRead>("ReadVIALUV_" # mx);
+    defvar ReadVIALUX_MX = !cast<SchedRead>("ReadVIALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m, Constraint>,
+            Sched<[WriteVIALUV_MX, ReadVIALUV_MX, ReadVIALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m, Constraint>,
+            Sched<[WriteVIALUX_MX, ReadVIALUV_MX, ReadVIALUX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VI_LMUL<ImmType, m, Constraint>,
+            Sched<[WriteVIALUI_MX, ReadVIALUV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVSALU_VV_VX {
   defm "" : VPseudoBinaryV_VV,
             Sched<[WriteVSALUV, ReadVSALUV, ReadVSALUV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX,
             Sched<[WriteVSALUX, ReadVSALUV, ReadVSALUX, ReadVMask]>;
 }
 
 multiclass VPseudoVSMUL_VV_VX {
   defm "" : VPseudoBinaryV_VV,
             Sched<[WriteVSMulV, ReadVSMulV, ReadVSMulV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX,
             Sched<[WriteVSMulX, ReadVSMulV, ReadVSMulX, ReadVMask]>;
 }
 
 multiclass VPseudoVAALU_VV_VX {
   defm "" : VPseudoBinaryV_VV,
             Sched<[WriteVAALUV, ReadVAALUV, ReadVAALUV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VX,
             Sched<[WriteVAALUX, ReadVAALUV, ReadVAALUX, ReadVMask]>;
 }
 
 multiclass VPseudoVMINMAX_VV_VX {
-  defm "" : VPseudoBinaryV_VV,
-            Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX,
-            Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICmpV_MX = !cast<SchedWrite>("WriteVICmpV_" # mx);
+    defvar WriteVICmpX_MX = !cast<SchedWrite>("WriteVICmpX_" # mx);
+    defvar ReadVICmpV_MX = !cast<SchedRead>("ReadVICmpV_" # mx);
+    defvar ReadVICmpX_MX = !cast<SchedRead>("ReadVICmpX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m>,
+              Sched<[WriteVICmpV_MX, ReadVICmpV_MX, ReadVICmpV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m>,
+              Sched<[WriteVICmpX_MX, ReadVICmpV_MX, ReadVICmpX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVMUL_VV_VX {
-  defm "" : VPseudoBinaryV_VV,
-            Sched<[WriteVIMulV, ReadVIMulV, ReadVIMulV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX,
-            Sched<[WriteVIMulX, ReadVIMulV, ReadVIMulX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIMulV_MX = !cast<SchedWrite>("WriteVIMulV_" # mx);
+    defvar WriteVIMulX_MX = !cast<SchedWrite>("WriteVIMulX_" # mx);
+    defvar ReadVIMulV_MX = !cast<SchedRead>("ReadVIMulV_" # mx);
+    defvar ReadVIMulX_MX = !cast<SchedRead>("ReadVIMulX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m>,
+              Sched<[WriteVIMulV_MX, ReadVIMulV_MX, ReadVIMulV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m>,
+              Sched<[WriteVIMulX_MX, ReadVIMulV_MX, ReadVIMulX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVDIV_VV_VX {
-  defm "" : VPseudoBinaryV_VV,
-            Sched<[WriteVIDivV, ReadVIDivV, ReadVIDivV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX,
-            Sched<[WriteVIDivX, ReadVIDivV, ReadVIDivX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIDivV_MX = !cast<SchedWrite>("WriteVIDivV_" # mx);
+    defvar WriteVIDivX_MX = !cast<SchedWrite>("WriteVIDivX_" # mx);
+    defvar ReadVIDivV_MX = !cast<SchedRead>("ReadVIDivV_" # mx);
+    defvar ReadVIDivX_MX = !cast<SchedRead>("ReadVIDivX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m>,
+              Sched<[WriteVIDivV_MX, ReadVIDivV_MX, ReadVIDivV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m>,
+              Sched<[WriteVIDivX_MX, ReadVIDivV_MX, ReadVIDivX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVFMUL_VV_VF {
   defm "" : VPseudoBinaryFV_VV,
             Sched<[WriteVFMulV, ReadVFMulV, ReadVFMulV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VF,
             Sched<[WriteVFMulF, ReadVFMulV, ReadVFMulF, ReadVMask]>;
 }
 
 multiclass VPseudoVFDIV_VV_VF {
   defm "" : VPseudoBinaryFV_VV,
             Sched<[WriteVFDivV, ReadVFDivV, ReadVFDivV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VF,
             Sched<[WriteVFDivF, ReadVFDivV, ReadVFDivF, ReadVMask]>;
 }
 
 multiclass VPseudoVFRDIV_VF {
   defm "" : VPseudoBinaryV_VF,
             Sched<[WriteVFDivF, ReadVFDivV, ReadVFDivF, ReadVMask]>;
 }
 
 multiclass VPseudoVALU_VV_VX {
-  defm "" : VPseudoBinaryV_VV,
-            Sched<[WriteVIALUV, ReadVIALUV, ReadVIALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VX,
-            Sched<[WriteVIALUX, ReadVIALUV, ReadVIALUX, ReadVMask]>;
+ foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIALUV_MX = !cast<SchedWrite>("WriteVIALUV_" # mx);
+    defvar WriteVIALUX_MX = !cast<SchedWrite>("WriteVIALUV_" # mx);
+    defvar ReadVIALUV_MX = !cast<SchedRead>("ReadVIALUV_" # mx);
+    defvar ReadVIALUX_MX = !cast<SchedRead>("ReadVIALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VV_LMUL<m>,
+            Sched<[WriteVIALUV_MX, ReadVIALUV_MX, ReadVIALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VX_LMUL<m>,
+            Sched<[WriteVIALUX_MX, ReadVIALUV_MX, ReadVIALUX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVSGNJ_VV_VF {
   defm "" : VPseudoBinaryFV_VV,
             Sched<[WriteVFSgnjV, ReadVFSgnjV, ReadVFSgnjV, ReadVMask]>;
   defm "" : VPseudoBinaryV_VF,
             Sched<[WriteVFSgnjF, ReadVFSgnjV, ReadVFSgnjF, ReadVMask]>;
 }

 }
 
 multiclass VPseudoVALU_VF {
   defm "" : VPseudoBinaryV_VF,
             Sched<[WriteVFALUF, ReadVFALUV, ReadVFALUF, ReadVMask]>;
 }
 
 multiclass VPseudoVALU_VX_VI<Operand ImmType = simm5> {
-  defm "" : VPseudoBinaryV_VX,
-            Sched<[WriteVIALUX, ReadVIALUV, ReadVIALUX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_VI<ImmType>,
-            Sched<[WriteVIALUI, ReadVIALUV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIALUX_MX = !cast<SchedWrite>("WriteVIALUX_" # mx);
+    defvar WriteVIALUI_MX = !cast<SchedWrite>("WriteVIALUI_" # mx);
+    defvar ReadVIALUV_MX = !cast<SchedRead>("ReadVIALUV_" # mx);
+    defvar ReadVIALUX_MX = !cast<SchedRead>("ReadVIALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VX_LMUL<m>,
+            Sched<[WriteVIALUX_MX, ReadVIALUV_MX, ReadVIALUX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_VI_LMUL<ImmType, m>,
+            Sched<[WriteVIALUI_MX, ReadVIALUV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVWALU_VV_VX {
-  defm "" : VPseudoBinaryW_VV,
-            Sched<[WriteVIWALUV, ReadVIWALUV, ReadVIWALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryW_VX,
-            Sched<[WriteVIWALUX, ReadVIWALUV, ReadVIWALUX, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVIWALUV_MX = !cast<SchedWrite>("WriteVIWALUV_" # mx);
+    defvar WriteVIWALUX_MX = !cast<SchedWrite>("WriteVIWALUX_" # mx);
+    defvar ReadVIWALUV_MX = !cast<SchedRead>("ReadVIWALUV_" # mx);
+    defvar ReadVIWALUX_MX = !cast<SchedRead>("ReadVIWALUX_" # mx);
+
+    defm "" : VPseudoBinaryW_VV_LMUL<m>,
+            Sched<[WriteVIWALUV_MX, ReadVIWALUV_MX, ReadVIWALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryW_VX_LMUL<m>,
+            Sched<[WriteVIWALUX_MX, ReadVIWALUV_MX, ReadVIWALUX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVWMUL_VV_VX {
-  defm "" : VPseudoBinaryW_VV,
-            Sched<[WriteVIWMulV, ReadVIWMulV, ReadVIWMulV, ReadVMask]>;
-  defm "" : VPseudoBinaryW_VX,
-            Sched<[WriteVIWMulX, ReadVIWMulV, ReadVIWMulX, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVIWMulV_MX = !cast<SchedWrite>("WriteVIWMulV_" # mx);
+    defvar WriteVIWMulX_MX = !cast<SchedWrite>("WriteVIWMulX_" # mx);
+    defvar ReadVIWMulV_MX = !cast<SchedRead>("ReadVIWMulV_" # mx);
+    defvar ReadVIWMulX_MX = !cast<SchedRead>("ReadVIWMulX_" # mx);
+
+    defm "" : VPseudoBinaryW_VV_LMUL<m>,
+              Sched<[WriteVIWMulV_MX, ReadVIWMulV_MX, ReadVIWMulV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryW_VX_LMUL<m>,
+              Sched<[WriteVIWMulX_MX, ReadVIWMulV_MX, ReadVIWMulX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVWMUL_VV_VF {
   defm "" : VPseudoBinaryW_VV<MxListFW>,
             Sched<[WriteVFWMulV, ReadVFWMulV, ReadVFWMulV, ReadVMask]>;
   defm "" : VPseudoBinaryW_VF,
             Sched<[WriteVFWMulF, ReadVFWMulV, ReadVFWMulF, ReadVMask]>;
 }
 
 multiclass VPseudoVWALU_WV_WX {
-  defm "" : VPseudoBinaryW_WV,
-            Sched<[WriteVIWALUV, ReadVIWALUV, ReadVIWALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryW_WX,
-            Sched<[WriteVIWALUX, ReadVIWALUV, ReadVIWALUX, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVIWALUV_MX = !cast<SchedWrite>("WriteVIWALUV_" # mx);
+    defvar WriteVIWALUX_MX = !cast<SchedWrite>("WriteVIWALUX_" # mx);
+    defvar ReadVIWALUV_MX = !cast<SchedRead>("ReadVIWALUV_" # mx);
+    defvar ReadVIWALUX_MX = !cast<SchedRead>("ReadVIWALUX_" # mx);
+
+    defm "" : VPseudoBinaryW_WV_LMUL<m>,
+              Sched<[WriteVIWALUV_MX, ReadVIWALUV_MX, ReadVIWALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryW_WX_LMUL<m>,
+              Sched<[WriteVIWALUX_MX, ReadVIWALUV_MX, ReadVIWALUX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVFWALU_VV_VF {
   defm "" : VPseudoBinaryW_VV<MxListFW>,
             Sched<[WriteVFWALUV, ReadVFWALUV, ReadVFWALUV, ReadVMask]>;
   defm "" : VPseudoBinaryW_VF,
             Sched<[WriteVFWALUF, ReadVFWALUV, ReadVFWALUF, ReadVMask]>;
 }
 
 multiclass VPseudoVFWALU_WV_WF {
   defm "" : VPseudoBinaryW_WV<MxListFW>,
             Sched<[WriteVFWALUV, ReadVFWALUV, ReadVFWALUV, ReadVMask]>;
   defm "" : VPseudoBinaryW_WF,
             Sched<[WriteVFWALUF, ReadVFWALUV, ReadVFWALUF, ReadVMask]>;
 }
 
 multiclass VPseudoVMRG_VM_XM_IM {
-  defm "" : VPseudoBinaryV_VM,
-            Sched<[WriteVIMergeV, ReadVIMergeV, ReadVIMergeV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_XM,
-            Sched<[WriteVIMergeX, ReadVIMergeV, ReadVIMergeX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_IM,
-            Sched<[WriteVIMergeI, ReadVIMergeV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIMergeV_MX = !cast<SchedWrite>("WriteVIMergeV_" # mx);
+    defvar WriteVIMergeX_MX = !cast<SchedWrite>("WriteVIMergeX_" # mx);
+    defvar WriteVIMergeI_MX = !cast<SchedWrite>("WriteVIMergeI_" # mx);
+    defvar ReadVIMergeV_MX = !cast<SchedRead>("ReadVIMergeV_" # mx);
+    defvar ReadVIMergeX_MX = !cast<SchedRead>("ReadVIMergeX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m>,
+              Sched<[WriteVIMergeV_MX, ReadVIMergeV_MX, ReadVIMergeV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_XM<m>,
+              Sched<[WriteVIMergeX_MX, ReadVIMergeV_MX, ReadVIMergeX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_IM<m>,
+              Sched<[WriteVIMergeI_MX, ReadVIMergeV_MX, ReadVMask]>;
     // Tied versions to allow codegen control over the tail elements
-  defm "" : VPseudoTiedBinaryV_VM,
-            Sched<[WriteVIMergeV, ReadVIMergeV, ReadVIMergeV, ReadVMask]>;
-  defm "" : VPseudoTiedBinaryV_XM,
-            Sched<[WriteVIMergeX, ReadVIMergeV, ReadVIMergeX, ReadVMask]>;
-  defm "" : VPseudoTiedBinaryV_IM,
-            Sched<[WriteVIMergeI, ReadVIMergeV, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_VM<m>,
+              Sched<[WriteVIMergeV_MX, ReadVIMergeV_MX, ReadVIMergeV_MX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_XM<m>,
+              Sched<[WriteVIMergeX_MX, ReadVIMergeV_MX, ReadVIMergeX_MX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_IM<m>,
+              Sched<[WriteVIMergeI_MX, ReadVIMergeV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCALU_VM_XM_IM {
-  defm "" : VPseudoBinaryV_VM,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_XM,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_IM,
-            Sched<[WriteVICALUI, ReadVICALUV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar WriteVICALUI_MX = !cast<SchedWrite>("WriteVICALUI_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_XM<m>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_IM<m>,
+              Sched<[WriteVICALUI_MX, ReadVICALUV_MX, ReadVMask]>;
     // Tied versions to allow codegen control over the tail elements
-  defm "" : VPseudoTiedBinaryV_VM,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoTiedBinaryV_XM,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
-  defm "" : VPseudoTiedBinaryV_IM,
-            Sched<[WriteVICALUI, ReadVICALUV, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_VM<m>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_XM<m>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_IM<m>,
+              Sched<[WriteVICALUI_MX, ReadVICALUV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCALU_VM_XM {
-  defm "" : VPseudoBinaryV_VM,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_XM,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_XM<m>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
     // Tied versions to allow codegen control over the tail elements
-  defm "" : VPseudoTiedBinaryV_VM,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoTiedBinaryV_XM,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_VM<m>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoTiedBinaryV_XM<m>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCALUM_VM_XM_IM<string Constraint> {
-  defm "" : VPseudoBinaryV_VM</*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_XM</*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_IM</*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
-            Sched<[WriteVICALUI, ReadVICALUV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar WriteVICALUI_MX = !cast<SchedWrite>("WriteVICALUI_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m, /*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_XM<m, /*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_IM<m, /*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
+              Sched<[WriteVICALUI_MX, ReadVICALUV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCALUM_VM_XM<string Constraint> {
-  defm "" : VPseudoBinaryV_VM</*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_XM</*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m, /*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_XM<m, /*CarryOut=*/1, /*CarryIn=*/1, Constraint>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCALUM_V_X_I<string Constraint> {
-  defm "" : VPseudoBinaryV_VM</*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV]>;
-  defm "" : VPseudoBinaryV_XM</*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX]>;
-  defm "" : VPseudoBinaryV_IM</*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
-            Sched<[WriteVICALUI, ReadVICALUV]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar WriteVICALUI_MX = !cast<SchedWrite>("WriteVICALUI_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m, /*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX]>;
+    defm "" : VPseudoBinaryV_XM<m, /*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX]>;
+    defm "" : VPseudoBinaryV_IM<m, /*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
+              Sched<[WriteVICALUI_MX, ReadVICALUV_MX]>;
+  }
 }
 
 multiclass VPseudoVCALUM_V_X<string Constraint> {
-  defm "" : VPseudoBinaryV_VM</*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
-            Sched<[WriteVICALUV, ReadVICALUV, ReadVICALUV]>;
-  defm "" : VPseudoBinaryV_XM</*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
-            Sched<[WriteVICALUX, ReadVICALUV, ReadVICALUX]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICALUV_MX = !cast<SchedWrite>("WriteVICALUV_" # mx);
+    defvar WriteVICALUX_MX = !cast<SchedWrite>("WriteVICALUX_" # mx);
+    defvar ReadVICALUV_MX = !cast<SchedRead>("ReadVICALUV_" # mx);
+    defvar ReadVICALUX_MX = !cast<SchedRead>("ReadVICALUX_" # mx);
+
+    defm "" : VPseudoBinaryV_VM<m, /*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
+              Sched<[WriteVICALUV_MX, ReadVICALUV_MX, ReadVICALUV_MX]>;
+    defm "" : VPseudoBinaryV_XM<m, /*CarryOut=*/1, /*CarryIn=*/0, Constraint>,
+              Sched<[WriteVICALUX_MX, ReadVICALUV_MX, ReadVICALUX_MX]>;
+  }
 }
 
 multiclass VPseudoVNCLP_WV_WX_WI {
   defm "" : VPseudoBinaryV_WV,
             Sched<[WriteVNClipV, ReadVNClipV, ReadVNClipV, ReadVMask]>;
   defm "" : VPseudoBinaryV_WX,
             Sched<[WriteVNClipX, ReadVNClipV, ReadVNClipX, ReadVMask]>;
   defm "" : VPseudoBinaryV_WI,
             Sched<[WriteVNClipI, ReadVNClipV, ReadVMask]>;
 }
 
 multiclass VPseudoVNSHT_WV_WX_WI {
-  defm "" : VPseudoBinaryV_WV,
-            Sched<[WriteVNShiftV, ReadVNShiftV, ReadVNShiftV, ReadVMask]>;
-  defm "" : VPseudoBinaryV_WX,
-            Sched<[WriteVNShiftX, ReadVNShiftV, ReadVNShiftX, ReadVMask]>;
-  defm "" : VPseudoBinaryV_WI,
-            Sched<[WriteVNShiftI, ReadVNShiftV, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVNShiftV_MX = !cast<SchedWrite>("WriteVNShiftV_" # mx);
+    defvar WriteVNShiftX_MX = !cast<SchedWrite>("WriteVNShiftX_" # mx);
+    defvar WriteVNShiftI_MX = !cast<SchedWrite>("WriteVNShiftI_" # mx);
+    defvar ReadVNShiftV_MX = !cast<SchedRead>("ReadVNShiftV_" # mx);
+    defvar ReadVNShiftX_MX = !cast<SchedRead>("ReadVNShiftX_" # mx);
+
+    defm "" : VPseudoBinaryV_WV_LMUL<m>,
+              Sched<[WriteVNShiftV_MX, ReadVNShiftV_MX, ReadVNShiftV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_WX_LMUL<m>,
+              Sched<[WriteVNShiftX_MX, ReadVNShiftV_MX, ReadVNShiftX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryV_WI_LMUL<m>,
+              Sched<[WriteVNShiftI_MX, ReadVNShiftV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoTernary<VReg RetClass,
                           RegisterClass Op1Class,
                           DAGOperand Op2Class,
                           LMULInfo MInfo,
                           string Constraint = ""> {
   let VLMul = MInfo.value in {

 multiclass VPseudoTernaryV_VV_AAXA<string Constraint = "",
                                    list<LMULInfo> mxlist = MxList> {
   foreach m = mxlist in {
     defm _VV : VPseudoTernaryWithPolicy<m.vrclass, m.vrclass, m.vrclass, m,
                                         Constraint, /*Commutable*/1>;
   }
 }
 
+multiclass VPseudoTernaryV_VV_AAXA_LMUL<LMULInfo m, string Constraint = ""> {
+  defm _VV : VPseudoTernaryWithPolicy<m.vrclass, m.vrclass, m.vrclass, m,
+                                      Constraint, /*Commutable*/1>;
+}
+
 multiclass VPseudoVSLDV_VX<string Constraint = ""> {
   foreach m = MxList in
     defm _VX : VPseudoTernaryWithPolicy<m.vrclass, m.vrclass, GPR, m, Constraint>;
 }
 
-multiclass VPseudoTernaryV_VX_AAXA<string Constraint = ""> {
-  foreach m = MxList in
-    defm "_VX" : VPseudoTernaryWithPolicy<m.vrclass, GPR, m.vrclass, m,
-                                          Constraint, /*Commutable*/1>;
+multiclass VPseudoTernaryV_VX_AAXA<LMULInfo m, string Constraint = ""> {
 }
 
 multiclass VPseudoTernaryV_VF_AAXA<string Constraint = ""> {
   foreach f = FPList in
     foreach m = f.MxList in
       defm "_V" # f.FX : VPseudoTernaryWithPolicy<m.vrclass, f.fprclass,
                                                   m.vrclass, m, Constraint,
                                                   /*Commutable*/1>;
 }
 
 multiclass VPseudoTernaryW_VV<list<LMULInfo> mxlist = MxListW> {
   defvar constraint = "@earlyclobber $rd";
   foreach m = mxlist in
     defm _VV : VPseudoTernaryWithPolicy<m.wvrclass, m.vrclass, m.vrclass, m,
                                         constraint>;
 }
 
-multiclass VPseudoTernaryW_VX {
+multiclass VPseudoTernaryW_VV_LMUL<LMULInfo m> {
+  defvar constraint = "@earlyclobber $rd";
+  defm _VV : VPseudoTernaryWithPolicy<m.wvrclass, m.vrclass, m.vrclass, m,
+                                      constraint>;
+}
+
+multiclass VPseudoTernaryW_VX<LMULInfo m> {
   defvar constraint = "@earlyclobber $rd";
-  foreach m = MxListW in
-    defm "_VX" : VPseudoTernaryWithPolicy<m.wvrclass, GPR, m.vrclass, m,
-                                          constraint>;
 }
 
 multiclass VPseudoTernaryW_VF {
   defvar constraint = "@earlyclobber $rd";
   foreach f = FPListW in
     foreach m = f.MxList in
       defm "_V" # f.FX : VPseudoTernaryWithPolicy<m.wvrclass, f.fprclass,
                                                   m.vrclass, m, constraint>;
 }
 
 multiclass VPseudoVSLDV_VI<Operand ImmType = simm5, string Constraint = ""> {
   foreach m = MxList in
     defm _VI : VPseudoTernaryWithPolicy<m.vrclass, m.vrclass, ImmType, m, Constraint>;
 }
 
 multiclass VPseudoVMAC_VV_VX_AAXA<string Constraint = ""> {
-  defm "" : VPseudoTernaryV_VV_AAXA<Constraint>,
-            Sched<[WriteVIMulAddV, ReadVIMulAddV, ReadVIMulAddV, ReadVIMulAddV, ReadVMask]>;
-  defm "" : VPseudoTernaryV_VX_AAXA<Constraint>,
-            Sched<[WriteVIMulAddX, ReadVIMulAddV, ReadVIMulAddV, ReadVIMulAddX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVIMulAddV_MX = !cast<SchedWrite>("WriteVIMulAddV_" # mx);
+    defvar WriteVIMulAddX_MX = !cast<SchedWrite>("WriteVIMulAddX_" # mx);
+    defvar ReadVIMulAddV_MX = !cast<SchedRead>("ReadVIMulAddV_" # mx);
+    defvar ReadVIMulAddX_MX = !cast<SchedRead>("ReadVIMulAddX_" # mx);
+
+    defm "" : VPseudoTernaryV_VV_AAXA_LMUL<m, Constraint>,
+              Sched<[WriteVIMulAddV_MX, ReadVIMulAddV_MX, ReadVIMulAddV_MX,
+                ReadVIMulAddV_MX, ReadVMask]>;
+    defm "" : VPseudoTernaryV_VX_AAXA<m, Constraint>,
+              Sched<[WriteVIMulAddX_MX, ReadVIMulAddV_MX, ReadVIMulAddV_MX,
+                ReadVIMulAddX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVMAC_VV_VF_AAXA<string Constraint = ""> {
   defm "" : VPseudoTernaryV_VV_AAXA<Constraint, MxListF>,
             Sched<[WriteVFMulAddV, ReadVFMulAddV, ReadVFMulAddV, ReadVFMulAddV, ReadVMask]>;
   defm "" : VPseudoTernaryV_VF_AAXA<Constraint>,
             Sched<[WriteVFMulAddF, ReadVFMulAddV, ReadVFMulAddV, ReadVFMulAddF, ReadVMask]>;
 }
 
 multiclass VPseudoVSLD_VX_VI<Operand ImmType = simm5, string Constraint = ""> {
   defm "" : VPseudoVSLDV_VX<Constraint>,
             Sched<[WriteVISlideX, ReadVISlideV, ReadVISlideV, ReadVISlideX, ReadVMask]>;
   defm "" : VPseudoVSLDV_VI<ImmType, Constraint>,
             Sched<[WriteVISlideI, ReadVISlideV, ReadVISlideV, ReadVMask]>;
 }
 
 multiclass VPseudoVWMAC_VV_VX {
-  defm "" : VPseudoTernaryW_VV,
-            Sched<[WriteVIWMulAddV, ReadVIWMulAddV, ReadVIWMulAddV, ReadVIWMulAddV, ReadVMask]>;
-  defm "" : VPseudoTernaryW_VX,
-            Sched<[WriteVIWMulAddX, ReadVIWMulAddV, ReadVIWMulAddV, ReadVIWMulAddX, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVIWMulAddV_MX = !cast<SchedWrite>("WriteVIWMulAddV_" # mx);
+    defvar WriteVIWMulAddX_MX = !cast<SchedWrite>("WriteVIWMulAddX_" # mx);
+    defvar ReadVIWMulAddV_MX = !cast<SchedRead>("ReadVIWMulAddV_" # mx);
+    defvar ReadVIWMulAddX_MX = !cast<SchedRead>("ReadVIWMulAddX_" # mx);
+
+    defm "" : VPseudoTernaryW_VV_LMUL<m>,
+              Sched<[WriteVIWMulAddV_MX, ReadVIWMulAddV_MX, ReadVIWMulAddV_MX,
+                ReadVIWMulAddV_MX, ReadVMask]>;
+    defm "" : VPseudoTernaryW_VX<m>,
+              Sched<[WriteVIWMulAddX_MX, ReadVIWMulAddV_MX, ReadVIWMulAddV_MX,
+                ReadVIWMulAddX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVWMAC_VX {
-  defm "" : VPseudoTernaryW_VX,
-            Sched<[WriteVIWMulAddX, ReadVIWMulAddV, ReadVIWMulAddV, ReadVIWMulAddX, ReadVMask]>;
+  foreach m = MxListW in {
+    defvar mx = m.MX;
+    defvar WriteVIWMulAddX_MX = !cast<SchedWrite>("WriteVIWMulAddX_" # mx);
+    defvar ReadVIWMulAddV_MX= !cast<SchedRead>("ReadVIWMulAddV_" # mx);
+    defvar ReadVIWMulAddX_MX = !cast<SchedRead>("ReadVIWMulAddX_" # mx);
+
+    defm "" : VPseudoTernaryW_VX<m>,
+              Sched<[WriteVIWMulAddX_MX, ReadVIWMulAddV_MX, ReadVIWMulAddV_MX,
+                ReadVIWMulAddX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVWMAC_VV_VF {
   defm "" : VPseudoTernaryW_VV<MxListFW>,
             Sched<[WriteVFWMulAddV, ReadVFWMulAddV, ReadVFWMulAddV, ReadVFWMulAddV, ReadVMask]>;
   defm "" : VPseudoTernaryW_VF,
             Sched<[WriteVFWMulAddF, ReadVFWMulAddV, ReadVFWMulAddV, ReadVFWMulAddF, ReadVMask]>;
 }
 
 multiclass VPseudoVCMPM_VV_VX_VI {
-  defm "" : VPseudoBinaryM_VV,
-            Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpV, ReadVMask]>;
-  defm "" : VPseudoBinaryM_VX,
-            Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
-  defm "" : VPseudoBinaryM_VI,
-            Sched<[WriteVICmpI, ReadVICmpV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICmpV_MX = !cast<SchedWrite>("WriteVICmpV_" # mx);
+    defvar WriteVICmpX_MX = !cast<SchedWrite>("WriteVICmpX_" # mx);
+    defvar WriteVICmpI_MX = !cast<SchedWrite>("WriteVICmpI_" # mx);
+    defvar ReadVICmpV_MX = !cast<SchedRead>("ReadVICmpV_" # mx);
+    defvar ReadVICmpX_MX = !cast<SchedRead>("ReadVICmpX_" # mx);
+
+    defm "" : VPseudoBinaryM_VV_LMUL<m>,
+              Sched<[WriteVICmpV_MX, ReadVICmpV_MX, ReadVICmpV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryM_VX<m>,
+              Sched<[WriteVICmpX_MX, ReadVICmpV_MX, ReadVICmpX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryM_VI<m>,
+              Sched<[WriteVICmpI_MX, ReadVICmpV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCMPM_VV_VX {
-  defm "" : VPseudoBinaryM_VV,
-            Sched<[WriteVICmpV, ReadVICmpV, ReadVICmpV, ReadVMask]>;
-  defm "" : VPseudoBinaryM_VX,
-            Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICmpV_MX = !cast<SchedWrite>("WriteVICmpV_" # mx);
+    defvar WriteVICmpX_MX = !cast<SchedWrite>("WriteVICmpX_" # mx);
+    defvar ReadVICmpV_MX = !cast<SchedRead>("ReadVICmpV_" # mx);
+    defvar ReadVICmpX_MX = !cast<SchedRead>("ReadVICmpX_" # mx);
+
+    defm "" : VPseudoBinaryM_VV_LMUL<m>,
+              Sched<[WriteVICmpV_MX, ReadVICmpV_MX, ReadVICmpV_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryM_VX<m>,
+              Sched<[WriteVICmpX_MX, ReadVICmpV_MX, ReadVICmpX_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVCMPM_VV_VF {
   defm "" : VPseudoBinaryM_VV<MxListF>,
             Sched<[WriteVFCmpV, ReadVFCmpV, ReadVFCmpV, ReadVMask]>;
   defm "" : VPseudoBinaryM_VF,
             Sched<[WriteVFCmpF, ReadVFCmpV, ReadVFCmpF, ReadVMask]>;
 }
 
 multiclass VPseudoVCMPM_VF {
   defm "" : VPseudoBinaryM_VF,
             Sched<[WriteVFCmpF, ReadVFCmpV, ReadVFCmpF, ReadVMask]>;
 }
 
 multiclass VPseudoVCMPM_VX_VI {
-  defm "" : VPseudoBinaryM_VX,
-            Sched<[WriteVICmpX, ReadVICmpV, ReadVICmpX, ReadVMask]>;
-  defm "" : VPseudoBinaryM_VI,
-            Sched<[WriteVICmpI, ReadVICmpV, ReadVMask]>;
+  foreach m = MxList in {
+    defvar mx = m.MX;
+    defvar WriteVICmpX_MX = !cast<SchedWrite>("WriteVICmpX_" # mx);
+    defvar WriteVICmpI_MX = !cast<SchedWrite>("WriteVICmpI_" # mx);
+    defvar ReadVICmpV_MX = !cast<SchedRead>("ReadVICmpV_" # mx);
+    defvar ReadVICmpX_MX = !cast<SchedRead>("ReadVICmpX_" # mx);
+
+    defm "" : VPseudoBinaryM_VX<m>,
+              Sched<[WriteVICmpX_MX, ReadVICmpV_MX, ReadVICmpX_MX, ReadVMask]>;
+    defm "" : VPseudoBinaryM_VI<m>,
+              Sched<[WriteVICmpI_MX, ReadVICmpV_MX, ReadVMask]>;
+  }
 }
 
 multiclass VPseudoVRED_VS {
   foreach m = MxList in {
     defm _VS : VPseudoTernary<V_M1.vrclass, m.vrclass, V_M1.vrclass, m>,
                Sched<[WriteVIRedV, ReadVIRedV, ReadVIRedV, ReadVIRedV, ReadVMask]>;
   }
 }

llvm/lib/Target/RISCV/RISCVScheduleV.td

 //===-- RISCVScheduleV.td - RISCV Scheduling Definitions V -*- 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
 //
 //===----------------------------------------------------------------------===//
 
 //===----------------------------------------------------------------------===//
 /// Define scheduler resources associated with def operands.
 
+defvar SchedMxList = ["UpperBound", "M1", "M2", "M4", "M8", "MF2", "MF4", "MF8"];
+// Used for widening and narrowing instructions as it doesn't contain M8.
+defvar SchedMxListW = ["UpperBound", "MF8", "MF4", "MF2", "M1", "M2", "M4"];
+
+// Creates SchedWrite for each (name, LMUL) pair for LMUL in lmuls argument
+multiclass LMULSchedWrites<string name, list<string> lmuls = SchedMxList> {
+  foreach mx = lmuls in {
+    def name # "_" # mx : SchedWrite;
+  }
+}
+
+// Creates SchedRead for each (name, LMUL) pair for LMUL in lmuls argument
+multiclass LMULSchedReads<string name, list<string> lmuls = SchedMxList> {
+  foreach mx = lmuls in {
+    def name # "_" # mx : SchedRead;
+  }
+}
+
+// Creates WriteRes for each (name, LMUL, resources) tuple for LMUL
+// in lmuls argument
+multiclass LMULWriteRes<string name, list<ProcResourceKind> resources,
+                        list<string> lmuls = SchedMxList> {
+  foreach mx = lmuls in {
+    def : WriteRes<!cast<SchedWrite>(name # "_" # mx), resources>;
+  }
+}
+
+// Creates ReadAdvance for each (name, LMUL, val) tuple for LMUL
+// in lmuls argument
+multiclass LMULReadAdvance<string name, int val, list<string> lmuls = SchedMxList> {
+  foreach mx = lmuls in {
+    def : ReadAdvance<!cast<SchedRead>(name # "_" # mx), val>;
+  }
+}
+
 // 3.6 Vector Byte Length vlenb
 def WriteRdVLENB      : SchedWrite;
 
 // 6. Configuration-Setting Instructions
 def WriteVSETVLI      : SchedWrite;
 def WriteVSETIVLI     : SchedWrite;
 def WriteVSETVL       : SchedWrite;
 

 def WriteVST1R        : SchedWrite;
 def WriteVST2R        : SchedWrite;
 def WriteVST4R        : SchedWrite;
 def WriteVST8R        : SchedWrite;
 
 // 11. Vector Integer Arithmetic Instructions
 // 11.1. Vector Single-Width Integer Add and Subtract
 // 11.5. Vector Bitwise Logical Instructions
-def WriteVIALUV       : SchedWrite;
-def WriteVIALUX       : SchedWrite;
-def WriteVIALUI       : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIALUV">;
+defm "" : LMULSchedWrites<"WriteVIALUX">;
+defm "" : LMULSchedWrites<"WriteVIALUI">;
 // 11.2. Vector Widening Integer Add/Subtract
-def WriteVIWALUV      : SchedWrite;
-def WriteVIWALUX      : SchedWrite;
-def WriteVIWALUI      : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIWALUV", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVIWALUX", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVIWALUI", SchedMxListW>;
 // 11.3. Vector Integer Extension
-def WriteVExtV        : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVExtV">;
 // 11.4. Vector Integer Arithmetic with Carry or Borrow Instructions
-def WriteVICALUV      : SchedWrite;
-def WriteVICALUX      : SchedWrite;
-def WriteVICALUI      : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVICALUV">;
+defm "" : LMULSchedWrites<"WriteVICALUX">;
+defm "" : LMULSchedWrites<"WriteVICALUI">;
 // 11.6. Vector Single-Width Bit Shift Instructions
-def WriteVShiftV      : SchedWrite;
-def WriteVShiftX      : SchedWrite;
-def WriteVShiftI      : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVShiftV">;
+defm "" : LMULSchedWrites<"WriteVShiftX">;
+defm "" : LMULSchedWrites<"WriteVShiftI">;
 // 11.7. Vector Narrowing Integer Right Shift Instructions
-def WriteVNShiftV     : SchedWrite;
-def WriteVNShiftX     : SchedWrite;
-def WriteVNShiftI     : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVNShiftV", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVNShiftX", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVNShiftI", SchedMxListW>;
 // 11.8. Vector Integer Comparison Instructions
 // 11.9. Vector Integer Min/Max Instructions
-def WriteVICmpV       : SchedWrite;
-def WriteVICmpX       : SchedWrite;
-def WriteVICmpI       : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVICmpV">;
+defm "" : LMULSchedWrites<"WriteVICmpX">;
+defm "" : LMULSchedWrites<"WriteVICmpI">;
 // 11.10. Vector Single-Width Integer Multiply Instructions
-def WriteVIMulV       : SchedWrite;
-def WriteVIMulX       : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIMulV">;
+defm "" : LMULSchedWrites<"WriteVIMulX">;
 // 11.11. Vector Integer Divide Instructions
-def WriteVIDivV       : SchedWrite;
-def WriteVIDivX       : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIDivV">;
+defm "" : LMULSchedWrites<"WriteVIDivX">;
 // 11.12. Vector Widening Integer Multiply Instructions
-def WriteVIWMulV      : SchedWrite;
-def WriteVIWMulX      : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIWMulV", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVIWMulX", SchedMxListW>;
 // 11.13. Vector Single-Width Integer Multiply-Add Instructions
-def WriteVIMulAddV    : SchedWrite;
-def WriteVIMulAddX    : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIMulAddV">;
+defm "" : LMULSchedWrites<"WriteVIMulAddX">;
 // 11.14. Vector Widening Integer Multiply-Add Instructions
-def WriteVIWMulAddV   : SchedWrite;
-def WriteVIWMulAddX   : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIWMulAddV", SchedMxListW>;
+defm "" : LMULSchedWrites<"WriteVIWMulAddX", SchedMxListW>;
 // 11.15. Vector Integer Merge Instructions
-def WriteVIMergeV     : SchedWrite;
-def WriteVIMergeX     : SchedWrite;
-def WriteVIMergeI     : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIMergeV">;
+defm "" : LMULSchedWrites<"WriteVIMergeX">;
+defm "" : LMULSchedWrites<"WriteVIMergeI">;
 // 11.16. Vector Integer Move Instructions
-def WriteVIMovV       : SchedWrite;
-def WriteVIMovX       : SchedWrite;
-def WriteVIMovI       : SchedWrite;
+defm "" : LMULSchedWrites<"WriteVIMovV">;
+defm "" : LMULSchedWrites<"WriteVIMovX">;
+defm "" : LMULSchedWrites<"WriteVIMovI">;
 
 // 12. Vector Fixed-Point Arithmetic Instructions
 // 12.1. Vector Single-Width Saturating Add and Subtract
 def WriteVSALUV       : SchedWrite;
 def WriteVSALUX       : SchedWrite;
 def WriteVSALUI       : SchedWrite;
 // 12.2. Vector Single-Width Averaging Add and Subtract
 def WriteVAALUV       : SchedWrite;

 def ReadVST1R         : SchedRead;
 def ReadVST2R         : SchedRead;
 def ReadVST4R         : SchedRead;
 def ReadVST8R         : SchedRead;
 
 // 11. Vector Integer Arithmetic Instructions
 // 11.1. Vector Single-Width Integer Add and Subtract
 // 11.5. Vector Bitwise Logical Instructions
-def ReadVIALUV        : SchedRead;
-def ReadVIALUX        : SchedRead;
+defm "" : LMULSchedReads<"ReadVIALUV">;
+defm "" : LMULSchedReads<"ReadVIALUX">;
 // 11.2. Vector Widening Integer Add/Subtract
-def ReadVIWALUV       : SchedRead;
-def ReadVIWALUX       : SchedRead;
+defm "" : LMULSchedReads<"ReadVIWALUV", SchedMxListW>;
+defm "" : LMULSchedReads<"ReadVIWALUX", SchedMxListW>;
 // 11.3. Vector Integer Extension
-def ReadVExtV         : SchedRead;
+defm "" : LMULSchedReads<"ReadVExtV">;
 // 11.4. Vector Integer Arithmetic with Carry or Borrow Instructions
-def ReadVICALUV       : SchedRead;
-def ReadVICALUX       : SchedRead;
+defm "" : LMULSchedReads<"ReadVICALUV">;
+defm "" : LMULSchedReads<"ReadVICALUX">;
 // 11.6. Vector Single-Width Bit Shift Instructions
-def ReadVShiftV       : SchedRead;
-def ReadVShiftX       : SchedRead;
+defm "" : LMULSchedReads<"ReadVShiftV">;
+defm "" : LMULSchedReads<"ReadVShiftX">;
 // 11.7. Vector Narrowing Integer Right Shift Instructions
-def ReadVNShiftV      : SchedRead;
-def ReadVNShiftX      : SchedRead;
+defm "" : LMULSchedReads<"ReadVNShiftV", SchedMxListW>;
+defm "" : LMULSchedReads<"ReadVNShiftX", SchedMxListW>;
 // 11.8. Vector Integer Comparison Instructions
 // 11.9. Vector Integer Min/Max Instructions
-def ReadVICmpV        : SchedRead;
-def ReadVICmpX        : SchedRead;
+defm "" : LMULSchedReads<"ReadVICmpV">;
+defm "" : LMULSchedReads<"ReadVICmpX">;
 // 11.10. Vector Single-Width Integer Multiply Instructions
-def ReadVIMulV        : SchedRead;
-def ReadVIMulX        : SchedRead;
+defm "" : LMULSchedReads<"ReadVIMulV">;
+defm "" : LMULSchedReads<"ReadVIMulX">;
 // 11.11. Vector Integer Divide Instructions
-def ReadVIDivV        : SchedRead;
-def ReadVIDivX        : SchedRead;
+defm "" : LMULSchedReads<"ReadVIDivV">;
+defm "" : LMULSchedReads<"ReadVIDivX">;
 // 11.12. Vector Widening Integer Multiply Instructions
-def ReadVIWMulV       : SchedRead;
-def ReadVIWMulX       : SchedRead;
+defm "" : LMULSchedReads<"ReadVIWMulV", SchedMxListW>;
+defm "" : LMULSchedReads<"ReadVIWMulX", SchedMxListW>;
 // 11.13. Vector Single-Width Integer Multiply-Add Instructions
-def ReadVIMulAddV     : SchedRead;
-def ReadVIMulAddX     : SchedRead;
+defm "" : LMULSchedReads<"ReadVIMulAddV">;
+defm "" : LMULSchedReads<"ReadVIMulAddX">;
 // 11.14. Vector Widening Integer Multiply-Add Instructions
-def ReadVIWMulAddV    : SchedRead;
-def ReadVIWMulAddX    : SchedRead;
+defm "" : LMULSchedReads<"ReadVIWMulAddV", SchedMxListW>;
+defm "" : LMULSchedReads<"ReadVIWMulAddX", SchedMxListW>;
 // 11.15. Vector Integer Merge Instructions
-def ReadVIMergeV      : SchedRead;
-def ReadVIMergeX      : SchedRead;
+defm "" : LMULSchedReads<"ReadVIMergeV">;
+defm "" : LMULSchedReads<"ReadVIMergeX">;
 // 11.16. Vector Integer Move Instructions
-def ReadVIMovV        : SchedRead;
-def ReadVIMovX        : SchedRead;
+defm "" : LMULSchedReads<"ReadVIMovV">;
+defm "" : LMULSchedReads<"ReadVIMovX">;
 
 // 12. Vector Fixed-Point Arithmetic Instructions
 // 12.1. Vector Single-Width Saturating Add and Subtract
 def ReadVSALUV        : SchedRead;
 def ReadVSALUX        : SchedRead;
 // 12.2. Vector Single-Width Averaging Add and Subtract
 def ReadVAALUV        : SchedRead;
 def ReadVAALUX        : SchedRead;
foreach eew = [8, 16, 32, 64] in {
     def : WriteRes <!cast<SchedWrite>("WriteVSSSEG" # nf # "e" # eew), []>;
     def : WriteRes <!cast<SchedWrite>("WriteVLUXSEG" # nf # "e" # eew), []>;
     def : WriteRes <!cast<SchedWrite>("WriteVLOXSEG" # nf # "e" # eew), []>;
     def : WriteRes <!cast<SchedWrite>("WriteVSUXSEG" # nf # "e" # eew), []>;
     def : WriteRes <!cast<SchedWrite>("WriteVSOXSEG" # nf # "e" # eew), []>;
   }
 }
 
-// 12. Vector Integer Arithmetic Instructions
-def : WriteRes<WriteVIALUV, []>;
-def : WriteRes<WriteVIALUX, []>;
-def : WriteRes<WriteVIALUI, []>;
-def : WriteRes<WriteVIWALUV, []>;
-def : WriteRes<WriteVIWALUX, []>;
-def : WriteRes<WriteVIWALUI, []>;
-def : WriteRes<WriteVExtV, []>;
-def : WriteRes<WriteVICALUV, []>;
-def : WriteRes<WriteVICALUX, []>;
-def : WriteRes<WriteVICALUI, []>;
-def : WriteRes<WriteVShiftV, []>;
-def : WriteRes<WriteVShiftX, []>;
-def : WriteRes<WriteVShiftI, []>;
-def : WriteRes<WriteVNShiftV, []>;
-def : WriteRes<WriteVNShiftX, []>;
-def : WriteRes<WriteVNShiftI, []>;
-def : WriteRes<WriteVICmpV, []>;
-def : WriteRes<WriteVICmpX, []>;
-def : WriteRes<WriteVICmpI, []>;
-def : WriteRes<WriteVIMulV, []>;
-def : WriteRes<WriteVIMulX, []>;
-def : WriteRes<WriteVIDivV, []>;
-def : WriteRes<WriteVIDivX, []>;
-def : WriteRes<WriteVIWMulV, []>;
-def : WriteRes<WriteVIWMulX, []>;
-def : WriteRes<WriteVIMulAddV, []>;
-def : WriteRes<WriteVIMulAddX, []>;
-def : WriteRes<WriteVIWMulAddV, []>;
-def : WriteRes<WriteVIWMulAddX, []>;
-def : WriteRes<WriteVIMergeV, []>;
-def : WriteRes<WriteVIMergeX, []>;
-def : WriteRes<WriteVIMergeI, []>;
-def : WriteRes<WriteVIMovV, []>;
-def : WriteRes<WriteVIMovX, []>;
-def : WriteRes<WriteVIMovI, []>;
+// 11. Vector Integer Arithmetic Instructions
+defm "" : LMULWriteRes<"WriteVIALUV", []>;
+defm "" : LMULWriteRes<"WriteVIALUX", []>;
+defm "" : LMULWriteRes<"WriteVIALUI", []>;
+defm "" : LMULWriteRes<"WriteVIWALUV", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIWALUX", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIWALUI", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVExtV", []>;
+defm "" : LMULWriteRes<"WriteVICALUV", []>;
+defm "" : LMULWriteRes<"WriteVICALUX", []>;
+defm "" : LMULWriteRes<"WriteVICALUI", []>;
+defm "" : LMULWriteRes<"WriteVShiftV", []>;
+defm "" : LMULWriteRes<"WriteVShiftX", []>;
+defm "" : LMULWriteRes<"WriteVShiftI", []>;
+defm "" : LMULWriteRes<"WriteVNShiftV", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVNShiftX", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVNShiftI", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVICmpV", []>;
+defm "" : LMULWriteRes<"WriteVICmpX", []>;
+defm "" : LMULWriteRes<"WriteVICmpI", []>;
+defm "" : LMULWriteRes<"WriteVIMulV", []>;
+defm "" : LMULWriteRes<"WriteVIMulX", []>;
+defm "" : LMULWriteRes<"WriteVIDivV", []>;
+defm "" : LMULWriteRes<"WriteVIDivX", []>;
+defm "" : LMULWriteRes<"WriteVIWMulV", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIWMulX", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIMulAddV", []>;
+defm "" : LMULWriteRes<"WriteVIMulAddX", []>;
+defm "" : LMULWriteRes<"WriteVIWMulAddV", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIWMulAddX", [], SchedMxListW>;
+defm "" : LMULWriteRes<"WriteVIMergeV", []>;
+defm "" : LMULWriteRes<"WriteVIMergeX", []>;
+defm "" : LMULWriteRes<"WriteVIMergeI", []>;
+defm "" : LMULWriteRes<"WriteVIMovV", []>;
+defm "" : LMULWriteRes<"WriteVIMovX", []>;
+defm "" : LMULWriteRes<"WriteVIMovI", []>;
 
 // 13. Vector Fixed-Point Arithmetic Instructions
 def : WriteRes<WriteVSALUV, []>;
 def : WriteRes<WriteVSALUX, []>;
 def : WriteRes<WriteVSALUI, []>;
 def : WriteRes<WriteVAALUV, []>;
 def : WriteRes<WriteVAALUX, []>;
 def : WriteRes<WriteVSMulV, []>;

 def : ReadAdvance<ReadVSTOX16V, 0>;
 def : ReadAdvance<ReadVSTOX32V, 0>;
 def : ReadAdvance<ReadVSTOX64V, 0>;
 def : ReadAdvance<ReadVST1R, 0>;
 def : ReadAdvance<ReadVST2R, 0>;
 def : ReadAdvance<ReadVST4R, 0>;
 def : ReadAdvance<ReadVST8R, 0>;
 
-// 12. Vector Integer Arithmetic Instructions
-def : ReadAdvance<ReadVIALUV, 0>;
-def : ReadAdvance<ReadVIALUX, 0>;
-def : ReadAdvance<ReadVIWALUV, 0>;
-def : ReadAdvance<ReadVIWALUX, 0>;
-def : ReadAdvance<ReadVExtV, 0>;
-def : ReadAdvance<ReadVICALUV, 0>;
-def : ReadAdvance<ReadVICALUX, 0>;
-def : ReadAdvance<ReadVShiftV, 0>;
-def : ReadAdvance<ReadVShiftX, 0>;
-def : ReadAdvance<ReadVNShiftV, 0>;
-def : ReadAdvance<ReadVNShiftX, 0>;
-def : ReadAdvance<ReadVICmpV, 0>;
-def : ReadAdvance<ReadVICmpX, 0>;
-def : ReadAdvance<ReadVIMulV, 0>;
-def : ReadAdvance<ReadVIMulX, 0>;
-def : ReadAdvance<ReadVIDivV, 0>;
-def : ReadAdvance<ReadVIDivX, 0>;
-def : ReadAdvance<ReadVIWMulV, 0>;
-def : ReadAdvance<ReadVIWMulX, 0>;
-def : ReadAdvance<ReadVIMulAddV, 0>;
-def : ReadAdvance<ReadVIMulAddX, 0>;
-def : ReadAdvance<ReadVIWMulAddV, 0>;
-def : ReadAdvance<ReadVIWMulAddX, 0>;
-def : ReadAdvance<ReadVIMergeV, 0>;
-def : ReadAdvance<ReadVIMergeX, 0>;
-def : ReadAdvance<ReadVIMovV, 0>;
-def : ReadAdvance<ReadVIMovX, 0>;
+// 11. Vector Integer Arithmetic Instructions
+defm "" : LMULReadAdvance<"ReadVIALUV", 0>;
+defm "" : LMULReadAdvance<"ReadVIALUX", 0>;
+defm "" : LMULReadAdvance<"ReadVIWALUV", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVIWALUX", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVExtV", 0>;
+defm "" : LMULReadAdvance<"ReadVICALUV", 0>;
+defm "" : LMULReadAdvance<"ReadVICALUX", 0>;
+defm "" : LMULReadAdvance<"ReadVShiftV", 0>;
+defm "" : LMULReadAdvance<"ReadVShiftX", 0>;
+defm "" : LMULReadAdvance<"ReadVNShiftV", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVNShiftX", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVICmpV", 0>;
+defm "" : LMULReadAdvance<"ReadVICmpX", 0>;
+defm "" : LMULReadAdvance<"ReadVIMulV", 0>;
+defm "" : LMULReadAdvance<"ReadVIMulX", 0>;
+defm "" : LMULReadAdvance<"ReadVIDivV", 0>;
+defm "" : LMULReadAdvance<"ReadVIDivX", 0>;
+defm "" : LMULReadAdvance<"ReadVIWMulV", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVIWMulX", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVIMulAddV", 0>;
+defm "" : LMULReadAdvance<"ReadVIMulAddX", 0>;
+defm "" : LMULReadAdvance<"ReadVIWMulAddV", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVIWMulAddX", 0, SchedMxListW>;
+defm "" : LMULReadAdvance<"ReadVIMergeV", 0>;
+defm "" : LMULReadAdvance<"ReadVIMergeX", 0>;
+defm "" : LMULReadAdvance<"ReadVIMovV", 0>;
+defm "" : LMULReadAdvance<"ReadVIMovX", 0>;
 
 // 13. Vector Fixed-Point Arithmetic Instructions
 def : ReadAdvance<ReadVSALUV, 0>;
 def : ReadAdvance<ReadVSALUX, 0>;
 def : ReadAdvance<ReadVAALUV, 0>;
 def : ReadAdvance<ReadVAALUX, 0>;
 def : ReadAdvance<ReadVSMulV, 0>;
 def : ReadAdvance<ReadVSMulX, 0>;