[LoongArch] Add support for getNumberOfRegisters()

[wangleiat]

The TTI hooks are used during vectorization for calculating register pressure. The default implementation defined wrong value for register number (all register class are 8 registers).

This patch also defines LoongArch's own register classes.

[llvmbot]

@llvm/pr-subscribers-backend-loongarch

@llvm/pr-subscribers-llvm-transforms

Author: wanglei (wangleiat)

Changes

The TTI hooks are used during vectorization for calculating register pressure. The default implementation definitly wrong value for register number (all register class are 8 registers).

This patch also defines LoongArch's own register classes.

---

Full diff: https://github.com/llvm/llvm-project/pull/88372.diff

3 Files Affected:

• (modified) llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.cpp (+41)
• (modified) llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.h (+4)
• (added) llvm/test/Transforms/LoopVectorize/LoongArch/reg-usage.ll (+44)

diff --git a/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.cpp b/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.cpp
index d47dded9ea6ecf..add1c60d89d21c 100644
--- a/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.cpp
+++ b/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.cpp
@@ -40,4 +40,45 @@ TypeSize LoongArchTTIImpl::getRegisterBitWidth(
   llvm_unreachable("Unsupported register kind");
 }
 
+unsigned LoongArchTTIImpl::getNumberOfRegisters(unsigned ClassID) const {
+  switch (ClassID) {
+  case LoongArchRegisterClass::GPRRC:
+    // 30 = 32 GPRs - r0 (zero register) - r21 (non-allocatable)
+    return 30;
+  case LoongArchRegisterClass::FPRRC:
+    return ST->hasBasicF() ? 32 : 0;
+  case LoongArchRegisterClass::VRRC:
+    return ST->hasExtLSX() ? 32 : 0;
+  }
+  llvm_unreachable("unknown register class");
+}
+
+unsigned LoongArchTTIImpl::getRegisterClassForType(bool Vector,
+                                                   Type *Ty) const {
+  if (Vector)
+    return LoongArchRegisterClass::VRRC;
+  if (!Ty)
+    return LoongArchRegisterClass::GPRRC;
+
+  Type *ScalarTy = Ty->getScalarType();
+  if ((ScalarTy->isFloatTy() && ST->hasBasicF()) ||
+      (ScalarTy->isDoubleTy() && ST->hasBasicD())) {
+    return LoongArchRegisterClass::FPRRC;
+  }
+
+  return LoongArchRegisterClass::GPRRC;
+}
+
+const char *LoongArchTTIImpl::getRegisterClassName(unsigned ClassID) const {
+  switch (ClassID) {
+  case LoongArchRegisterClass::GPRRC:
+    return "LoongArch::GPRRC";
+  case LoongArchRegisterClass::FPRRC:
+    return "LoongArch::FPRRC";
+  case LoongArchRegisterClass::VRRC:
+    return "LoongArch::VRRC";
+  }
+  llvm_unreachable("unknown register class");
+}
+
 // TODO: Implement more hooks to provide TTI machinery for LoongArch.
diff --git a/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.h b/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.h
index d296c9ed576fbd..34c18163bbdb6e 100644
--- a/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.h
+++ b/llvm/lib/Target/LoongArch/LoongArchTargetTransformInfo.h
@@ -28,6 +28,7 @@ class LoongArchTTIImpl : public BasicTTIImplBase<LoongArchTTIImpl> {
   typedef TargetTransformInfo TTI;
   friend BaseT;
 
+  enum LoongArchRegisterClass { GPRRC, FPRRC, VRRC };
   const LoongArchSubtarget *ST;
   const LoongArchTargetLowering *TLI;
 
@@ -40,6 +41,9 @@ class LoongArchTTIImpl : public BasicTTIImplBase<LoongArchTTIImpl> {
         TLI(ST->getTargetLowering()) {}
 
   TypeSize getRegisterBitWidth(TargetTransformInfo::RegisterKind K) const;
+  unsigned getNumberOfRegisters(unsigned ClassID) const;
+  unsigned getRegisterClassForType(bool Vector, Type *Ty = nullptr) const;
+  const char *getRegisterClassName(unsigned ClassID) const;
 
   // TODO: Implement more hooks to provide TTI machinery for LoongArch.
 };
diff --git a/llvm/test/Transforms/LoopVectorize/LoongArch/reg-usage.ll b/llvm/test/Transforms/LoopVectorize/LoongArch/reg-usage.ll
new file mode 100644
index 00000000000000..f45a2f0f5b7e8f
--- /dev/null
+++ b/llvm/test/Transforms/LoopVectorize/LoongArch/reg-usage.ll
@@ -0,0 +1,44 @@
+; REQUIRES: asserts
+; RUN: opt --passes=loop-vectorize --mtriple loongarch64-linux-gnu \
+; RUN:   --mattr=+lsx -debug-only=loop-vectorize --force-vector-width=1 \
+; RUN:   -S < %s 2>&1 | FileCheck %s --check-prefix=CHECK-SCALAR
+; RUN: opt --passes=loop-vectorize --mtriple loongarch64-linux-gnu \
+; RUN:   --mattr=+lsx -debug-only=loop-vectorize --force-vector-width=4 \
+; RUN:   -S < %s 2>&1 | FileCheck %s --check-prefix=CHECK-VECTOR
+
+define void @bar(ptr %A, i32 signext %n) {
+; CHECK-LABEL: bar
+; CHECK-SCALAR:      LV(REG): Found max usage: 2 item
+; CHECK-SCALAR-NEXT: LV(REG): RegisterClass: LoongArch::GPRRC, 2 registers
+; CHECK-SCALAR-NEXT: LV(REG): RegisterClass: LoongArch::FPRRC, 1 registers
+; CHECK-SCALAR-NEXT: LV(REG): Found invariant usage: 1 item
+; CHECK-SCALAR-NEXT: LV(REG): RegisterClass: LoongArch::GPRRC, 1 registers
+; CHECK-SCALAR-NEXT: LV: The target has 30 registers of LoongArch::GPRRC register class
+; CHECK-SCALAR-NEXT: LV: The target has 32 registers of LoongArch::FPRRC register class
+; CHECK-VECTOR:      LV(REG): Found max usage: 1 item
+; CHECK-VECTOR-NEXT: LV(REG): RegisterClass: LoongArch::VRRC, 3 registers
+; CHECK-VECTOR-NEXT: LV(REG): Found invariant usage: 1 item
+; CHECK-VECTOR-NEXT: LV(REG): RegisterClass: LoongArch::GPRRC, 1 registers
+; CHECK-VECTOR-NEXT: LV: The target has 32 registers of LoongArch::VRRC register class
+
+entry:
+  %cmp4 = icmp sgt i32 %n, 0
+  br i1 %cmp4, label %for.body.preheader, label %for.cond.cleanup
+
+for.body.preheader:                               ; preds = %entry
+  %wide.trip.count = zext nneg i32 %n to i64
+  br label %for.body
+
+for.cond.cleanup:                                 ; preds = %for.body, %entry
+  ret void
+
+for.body:                                         ; preds = %for.body.preheader, %for.body
+  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %0 = trunc i64 %indvars.iv to i32
+  %conv = sitofp i32 %0 to float
+  %arrayidx = getelementptr inbounds float, ptr %A, i64 %indvars.iv
+  store float %conv, ptr %arrayidx, align 4
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond.not = icmp eq i64 %indvars.iv.next, %wide.trip.count
+  br i1 %exitcond.not, label %for.cond.cleanup, label %for.body
+}