Use subtarget feature

Author: SamTebbs33

llvm/lib/Target/AArch64/AArch64Features.td

@@ -894,6 +894,11 @@ def FeatureUseFixedOverScalableIfEqualCost : SubtargetFeature<"use-fixed-over-sc
   "UseFixedOverScalableIfEqualCost", "true",
   "Prefer fixed width loop vectorization over scalable if the cost-model assigns equal costs">;
 
+def FeatureDisableMaximizeScalableBandwidth : SubtargetFeature< "disable-maximize-scalable-bandwidth",
+    "DisableMaximizeScalableBandwidth", "true",
+    "Determine the maximum scalable vector length for a loop by the "
+    "largest scalar type rather than the smallest">;
+
 // For performance reasons we prefer to use ldapr to ldapur on certain cores.
 def FeatureAvoidLDAPUR : SubtargetFeature<"avoid-ldapur", "AvoidLDAPUR", "true",
   "Prefer add+ldapr to offset ldapur">;

llvm/lib/Target/AArch64/AArch64Processors.td

@@ -593,6 +593,7 @@ def TuneNeoverseN2 : SubtargetFeature<"neoversen2", "ARMProcFamily", "NeoverseN2
                                       FeatureALULSLFast,
                                       FeaturePostRAScheduler,
                                       FeatureEnableSelectOptimize,
+                                      FeatureDisableMaximizeScalableBandwidth,
                                       FeaturePredictableSelectIsExpensive]>;
 
 def TuneNeoverseN3 : SubtargetFeature<"neoversen3", "ARMProcFamily", "NeoverseN3",
@@ -626,6 +627,7 @@ def TuneNeoverseV1 : SubtargetFeature<"neoversev1", "ARMProcFamily", "NeoverseV1
                                       FeaturePostRAScheduler,
                                       FeatureEnableSelectOptimize,
                                       FeaturePredictableSelectIsExpensive,
+                                      FeatureDisableMaximizeScalableBandwidth,
                                       FeatureNoSVEFPLD1R]>;
 
 def TuneNeoverseV2 : SubtargetFeature<"neoversev2", "ARMProcFamily", "NeoverseV2",

llvm/lib/Target/AArch64/AArch64Subtarget.cpp

@@ -267,28 +267,21 @@ void AArch64Subtarget::initializeProperties(bool HasMinSize) {
     PrefLoopAlignment = Align(32);
     MaxBytesForLoopAlignment = 16;
     break;
-  case NeoverseN2:
-    ShouldMaximizeScalableVectorBandwidth = false;
-    PrefFunctionAlignment = Align(16);
-    PrefLoopAlignment = Align(32);
-    MaxBytesForLoopAlignment = 16;
-    VScaleForTuning = 1;
-    break;
   case NeoverseV2:
   case NeoverseV3:
     CacheLineSize = 64;
     EpilogueVectorizationMinVF = 8;
     MaxInterleaveFactor = 4;
     ScatterOverhead = 13;
     [[fallthrough]];
+  case NeoverseN2:
   case NeoverseN3:
     PrefFunctionAlignment = Align(16);
     PrefLoopAlignment = Align(32);
     MaxBytesForLoopAlignment = 16;
     VScaleForTuning = 1;
     break;
   case NeoverseV1:
-    ShouldMaximizeScalableVectorBandwidth = false;
     PrefFunctionAlignment = Align(16);
     PrefLoopAlignment = Align(32);
     MaxBytesForLoopAlignment = 16;

llvm/lib/Target/AArch64/AArch64Subtarget.h

@@ -57,7 +57,6 @@ class AArch64Subtarget final : public AArch64GenSubtargetInfo {
 #include "AArch64GenSubtargetInfo.inc"
 
   unsigned EpilogueVectorizationMinVF = 16;
-  bool ShouldMaximizeScalableVectorBandwidth = true;
   uint8_t MaxInterleaveFactor = 2;
   uint8_t VectorInsertExtractBaseCost = 2;
   uint16_t CacheLineSize = 0;
@@ -252,9 +251,6 @@ class AArch64Subtarget final : public AArch64GenSubtargetInfo {
   unsigned getEpilogueVectorizationMinVF() const {
     return EpilogueVectorizationMinVF;
   }
-  bool shouldMaximizeScalableVectorBandwidth() const {
-    return ShouldMaximizeScalableVectorBandwidth;
-  }
   unsigned getMaxInterleaveFactor() const { return MaxInterleaveFactor; }
   unsigned getVectorInsertExtractBaseCost() const;
   unsigned getCacheLineSize() const override { return CacheLineSize; }

llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp

@@ -381,7 +381,7 @@ bool AArch64TTIImpl::shouldMaximizeVectorBandwidth(
 
   return K == TargetTransformInfo::RGK_ScalableVector &&
          ST->isSVEorStreamingSVEAvailable() &&
-         ST->shouldMaximizeScalableVectorBandwidth();
+         !ST->disableMaximizeScalableBandwidth();
 }
 
 /// Calculate the cost of materializing a 64-bit value. This helper