[NVPTX] Add TTI support for folding isspacep in InferAS #114486

AlexMaclean · 2024-10-31T23:18:45Z

This change enables constant folding of '@llvm.nvvm.isspacep.*' intrinsics if the address space can be propagated in InferAdressSpace.

llvmbot · 2024-10-31T23:19:20Z

@llvm/pr-subscribers-backend-nvptx

Author: Alex MacLean (AlexMaclean)

Changes

This change enables constant folding of '@llvm.nvvm.isspacep.*' intrinsics if the address space can be propagated in InferAdressSpace.

Full diff: https://github.com/llvm/llvm-project/pull/114486.diff

3 Files Affected:

(modified) llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp (+59-24)
(modified) llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h (+6)
(added) llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll (+144)

diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
index 3507573df1869f..5df211726882ec 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
@@ -416,33 +416,34 @@ static Instruction *convertNvvmIntrinsicToLlvm(InstCombiner &IC,
   llvm_unreachable("All SpecialCase enumerators should be handled in switch.");
 }
 
+  // Returns true/false when we know the answer, nullopt otherwise.
+static std::optional<bool> evaluateIsSpace(Intrinsic::ID IID, unsigned AS) {
+  if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC ||
+      AS == NVPTXAS::ADDRESS_SPACE_PARAM)
+    return std::nullopt; // Got to check at run-time.
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_global:
+    return AS == NVPTXAS::ADDRESS_SPACE_GLOBAL;
+  case Intrinsic::nvvm_isspacep_local:
+    return AS == NVPTXAS::ADDRESS_SPACE_LOCAL;
+  case Intrinsic::nvvm_isspacep_shared:
+    return AS == NVPTXAS::ADDRESS_SPACE_SHARED;
+  case Intrinsic::nvvm_isspacep_shared_cluster:
+    // We can't tell shared from shared_cluster at compile time from AS alone,
+    // but it can't be either is AS is not shared.
+    return AS == NVPTXAS::ADDRESS_SPACE_SHARED ? std::nullopt
+                                               : std::optional{false};
+  case Intrinsic::nvvm_isspacep_const:
+    return AS == NVPTXAS::ADDRESS_SPACE_CONST;
+  default:
+    llvm_unreachable("Unexpected intrinsic");
+  }
+}
+
 // Returns an instruction pointer (may be nullptr if we do not know the answer).
 // Returns nullopt if `II` is not one of the `isspacep` intrinsics.
 static std::optional<Instruction *>
 handleSpaceCheckIntrinsics(InstCombiner &IC, IntrinsicInst &II) {
-  // Returns true/false when we know the answer, nullopt otherwise.
-  auto CheckASMatch = [](unsigned IID, unsigned AS) -> std::optional<bool> {
-    if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC ||
-        AS == NVPTXAS::ADDRESS_SPACE_PARAM)
-      return std::nullopt; // Got to check at run-time.
-    switch (IID) {
-    case Intrinsic::nvvm_isspacep_global:
-      return AS == NVPTXAS::ADDRESS_SPACE_GLOBAL;
-    case Intrinsic::nvvm_isspacep_local:
-      return AS == NVPTXAS::ADDRESS_SPACE_LOCAL;
-    case Intrinsic::nvvm_isspacep_shared:
-      return AS == NVPTXAS::ADDRESS_SPACE_SHARED;
-    case Intrinsic::nvvm_isspacep_shared_cluster:
-      // We can't tell shared from shared_cluster at compile time from AS alone,
-      // but it can't be either is AS is not shared.
-      return AS == NVPTXAS::ADDRESS_SPACE_SHARED ? std::nullopt
-                                                 : std::optional{false};
-    case Intrinsic::nvvm_isspacep_const:
-      return AS == NVPTXAS::ADDRESS_SPACE_CONST;
-    default:
-      llvm_unreachable("Unexpected intrinsic");
-    }
-  };
 
   switch (auto IID = II.getIntrinsicID()) {
   case Intrinsic::nvvm_isspacep_global:
@@ -458,7 +459,7 @@ handleSpaceCheckIntrinsics(InstCombiner &IC, IntrinsicInst &II) {
       if (auto *ASCO = dyn_cast<AddrSpaceCastOperator>(Op0))
         AS = ASCO->getOperand(0)->getType()->getPointerAddressSpace();
 
-    if (std::optional<bool> Answer = CheckASMatch(IID, AS))
+    if (std::optional<bool> Answer = evaluateIsSpace(IID, AS))
       return IC.replaceInstUsesWith(II,
                                     ConstantInt::get(II.getType(), *Answer));
     return nullptr; // Don't know the answer, got to check at run time.
@@ -525,3 +526,37 @@ void NVPTXTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
                                          TTI::PeelingPreferences &PP) {
   BaseT::getPeelingPreferences(L, SE, PP);
 }
+
+bool NVPTXTTIImpl::collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
+                                              Intrinsic::ID IID) const {
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_const:
+  case Intrinsic::nvvm_isspacep_global:
+  case Intrinsic::nvvm_isspacep_local:
+  case Intrinsic::nvvm_isspacep_shared:
+  case Intrinsic::nvvm_isspacep_shared_cluster: {
+    OpIndexes.push_back(0);
+    return true;
+  }
+  }
+  return false;
+}
+
+Value *NVPTXTTIImpl::rewriteIntrinsicWithAddressSpace(IntrinsicInst *II,
+                                                      Value *OldV,
+                                                      Value *NewV) const {
+  const Intrinsic::ID IID = II->getIntrinsicID();
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_const:
+  case Intrinsic::nvvm_isspacep_global:
+  case Intrinsic::nvvm_isspacep_local:
+  case Intrinsic::nvvm_isspacep_shared:
+  case Intrinsic::nvvm_isspacep_shared_cluster: {
+    const unsigned NewAS = NewV->getType()->getPointerAddressSpace();
+    if (const auto R = evaluateIsSpace(IID, NewAS))
+      return ConstantInt::get(II->getType(), *R);
+    return nullptr;
+  }
+  }
+  return nullptr;
+}
\ No newline at end of file
diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
index 86140daa7be489..0f4fb280b2d996 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
@@ -123,6 +123,12 @@ class NVPTXTTIImpl : public BasicTTIImplBase<NVPTXTTIImpl> {
       return true;
     }
   }
+
+  bool collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
+                                  Intrinsic::ID IID) const;
+
+  Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
+                                          Value *NewV) const;
 };
 
 } // end namespace llvm
diff --git a/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll b/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll
new file mode 100644
index 00000000000000..348fa688770df6
--- /dev/null
+++ b/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll
@@ -0,0 +1,144 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -mtriple=nvptx64-nvidia-cuda -passes=infer-address-spaces,bdce %s | FileCheck %s
+
+target datalayout = "e-p:64:64:64-p3:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-i128:128:128-f32:32:32-f64:64:64-v16:16:16-v32:32:32-v64:64:64-v128:128:128-n16:32:64"
+target triple = "nvptx64-nvidia-cuda"
+
+declare i1 @llvm.nvvm.isspacep.const(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.global(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.local(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.shared(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.cluster.shared(ptr) readnone noinline
+
+define i1 @test_isspacep_const_true(ptr addrspace(4) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_const_true(
+; CHECK-SAME: ptr addrspace(4) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(4) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.const(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_const_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_const_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.const(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_global_true(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_global_true(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.global(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_global_false(ptr addrspace(4) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_global_false(
+; CHECK-SAME: ptr addrspace(4) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(4) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.global(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_local_true(ptr addrspace(5) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_local_true(
+; CHECK-SAME: ptr addrspace(5) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(5) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.local(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_local_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_local_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.local(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_shared_true(ptr addrspace(3) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_shared_true(
+; CHECK-SAME: ptr addrspace(3) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(3) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_shared_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_shared_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_cluster_shared_unsure(ptr addrspace(3) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_cluster_shared_unsure(
+; CHECK-SAME: ptr addrspace(3) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[ADDR1:%.*]] = getelementptr i8, ptr addrspace(3) [[ADDR]], i32 10
+; CHECK-NEXT:    [[TMP0:%.*]] = addrspacecast ptr addrspace(3) [[ADDR1]] to ptr
+; CHECK-NEXT:    [[VAL:%.*]] = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr [[TMP0]])
+; CHECK-NEXT:    ret i1 [[VAL]]
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(3) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_cluster_shared_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_cluster_shared_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr %addr1)
+  ret i1 %val
+}

llvmbot · 2024-10-31T23:19:21Z

@llvm/pr-subscribers-llvm-transforms

Author: Alex MacLean (AlexMaclean)

Changes

This change enables constant folding of '@llvm.nvvm.isspacep.*' intrinsics if the address space can be propagated in InferAdressSpace.

Full diff: https://github.com/llvm/llvm-project/pull/114486.diff

3 Files Affected:

(modified) llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp (+59-24)
(modified) llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h (+6)
(added) llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll (+144)

diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
index 3507573df1869f..5df211726882ec 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp
@@ -416,33 +416,34 @@ static Instruction *convertNvvmIntrinsicToLlvm(InstCombiner &IC,
   llvm_unreachable("All SpecialCase enumerators should be handled in switch.");
 }
 
+  // Returns true/false when we know the answer, nullopt otherwise.
+static std::optional<bool> evaluateIsSpace(Intrinsic::ID IID, unsigned AS) {
+  if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC ||
+      AS == NVPTXAS::ADDRESS_SPACE_PARAM)
+    return std::nullopt; // Got to check at run-time.
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_global:
+    return AS == NVPTXAS::ADDRESS_SPACE_GLOBAL;
+  case Intrinsic::nvvm_isspacep_local:
+    return AS == NVPTXAS::ADDRESS_SPACE_LOCAL;
+  case Intrinsic::nvvm_isspacep_shared:
+    return AS == NVPTXAS::ADDRESS_SPACE_SHARED;
+  case Intrinsic::nvvm_isspacep_shared_cluster:
+    // We can't tell shared from shared_cluster at compile time from AS alone,
+    // but it can't be either is AS is not shared.
+    return AS == NVPTXAS::ADDRESS_SPACE_SHARED ? std::nullopt
+                                               : std::optional{false};
+  case Intrinsic::nvvm_isspacep_const:
+    return AS == NVPTXAS::ADDRESS_SPACE_CONST;
+  default:
+    llvm_unreachable("Unexpected intrinsic");
+  }
+}
+
 // Returns an instruction pointer (may be nullptr if we do not know the answer).
 // Returns nullopt if `II` is not one of the `isspacep` intrinsics.
 static std::optional<Instruction *>
 handleSpaceCheckIntrinsics(InstCombiner &IC, IntrinsicInst &II) {
-  // Returns true/false when we know the answer, nullopt otherwise.
-  auto CheckASMatch = [](unsigned IID, unsigned AS) -> std::optional<bool> {
-    if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC ||
-        AS == NVPTXAS::ADDRESS_SPACE_PARAM)
-      return std::nullopt; // Got to check at run-time.
-    switch (IID) {
-    case Intrinsic::nvvm_isspacep_global:
-      return AS == NVPTXAS::ADDRESS_SPACE_GLOBAL;
-    case Intrinsic::nvvm_isspacep_local:
-      return AS == NVPTXAS::ADDRESS_SPACE_LOCAL;
-    case Intrinsic::nvvm_isspacep_shared:
-      return AS == NVPTXAS::ADDRESS_SPACE_SHARED;
-    case Intrinsic::nvvm_isspacep_shared_cluster:
-      // We can't tell shared from shared_cluster at compile time from AS alone,
-      // but it can't be either is AS is not shared.
-      return AS == NVPTXAS::ADDRESS_SPACE_SHARED ? std::nullopt
-                                                 : std::optional{false};
-    case Intrinsic::nvvm_isspacep_const:
-      return AS == NVPTXAS::ADDRESS_SPACE_CONST;
-    default:
-      llvm_unreachable("Unexpected intrinsic");
-    }
-  };
 
   switch (auto IID = II.getIntrinsicID()) {
   case Intrinsic::nvvm_isspacep_global:
@@ -458,7 +459,7 @@ handleSpaceCheckIntrinsics(InstCombiner &IC, IntrinsicInst &II) {
       if (auto *ASCO = dyn_cast<AddrSpaceCastOperator>(Op0))
         AS = ASCO->getOperand(0)->getType()->getPointerAddressSpace();
 
-    if (std::optional<bool> Answer = CheckASMatch(IID, AS))
+    if (std::optional<bool> Answer = evaluateIsSpace(IID, AS))
       return IC.replaceInstUsesWith(II,
                                     ConstantInt::get(II.getType(), *Answer));
     return nullptr; // Don't know the answer, got to check at run time.
@@ -525,3 +526,37 @@ void NVPTXTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
                                          TTI::PeelingPreferences &PP) {
   BaseT::getPeelingPreferences(L, SE, PP);
 }
+
+bool NVPTXTTIImpl::collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
+                                              Intrinsic::ID IID) const {
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_const:
+  case Intrinsic::nvvm_isspacep_global:
+  case Intrinsic::nvvm_isspacep_local:
+  case Intrinsic::nvvm_isspacep_shared:
+  case Intrinsic::nvvm_isspacep_shared_cluster: {
+    OpIndexes.push_back(0);
+    return true;
+  }
+  }
+  return false;
+}
+
+Value *NVPTXTTIImpl::rewriteIntrinsicWithAddressSpace(IntrinsicInst *II,
+                                                      Value *OldV,
+                                                      Value *NewV) const {
+  const Intrinsic::ID IID = II->getIntrinsicID();
+  switch (IID) {
+  case Intrinsic::nvvm_isspacep_const:
+  case Intrinsic::nvvm_isspacep_global:
+  case Intrinsic::nvvm_isspacep_local:
+  case Intrinsic::nvvm_isspacep_shared:
+  case Intrinsic::nvvm_isspacep_shared_cluster: {
+    const unsigned NewAS = NewV->getType()->getPointerAddressSpace();
+    if (const auto R = evaluateIsSpace(IID, NewAS))
+      return ConstantInt::get(II->getType(), *R);
+    return nullptr;
+  }
+  }
+  return nullptr;
+}
\ No newline at end of file
diff --git a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
index 86140daa7be489..0f4fb280b2d996 100644
--- a/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
+++ b/llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.h
@@ -123,6 +123,12 @@ class NVPTXTTIImpl : public BasicTTIImplBase<NVPTXTTIImpl> {
       return true;
     }
   }
+
+  bool collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
+                                  Intrinsic::ID IID) const;
+
+  Value *rewriteIntrinsicWithAddressSpace(IntrinsicInst *II, Value *OldV,
+                                          Value *NewV) const;
 };
 
 } // end namespace llvm
diff --git a/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll b/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll
new file mode 100644
index 00000000000000..348fa688770df6
--- /dev/null
+++ b/llvm/test/Transforms/InferAddressSpaces/NVPTX/isspacep.ll
@@ -0,0 +1,144 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -mtriple=nvptx64-nvidia-cuda -passes=infer-address-spaces,bdce %s | FileCheck %s
+
+target datalayout = "e-p:64:64:64-p3:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-i128:128:128-f32:32:32-f64:64:64-v16:16:16-v32:32:32-v64:64:64-v128:128:128-n16:32:64"
+target triple = "nvptx64-nvidia-cuda"
+
+declare i1 @llvm.nvvm.isspacep.const(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.global(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.local(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.shared(ptr) readnone noinline
+declare i1 @llvm.nvvm.isspacep.cluster.shared(ptr) readnone noinline
+
+define i1 @test_isspacep_const_true(ptr addrspace(4) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_const_true(
+; CHECK-SAME: ptr addrspace(4) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(4) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.const(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_const_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_const_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.const(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_global_true(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_global_true(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.global(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_global_false(ptr addrspace(4) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_global_false(
+; CHECK-SAME: ptr addrspace(4) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(4) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.global(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_local_true(ptr addrspace(5) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_local_true(
+; CHECK-SAME: ptr addrspace(5) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(5) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.local(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_local_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_local_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.local(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_shared_true(ptr addrspace(3) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_shared_true(
+; CHECK-SAME: ptr addrspace(3) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 true
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(3) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_shared_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_shared_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_cluster_shared_unsure(ptr addrspace(3) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_cluster_shared_unsure(
+; CHECK-SAME: ptr addrspace(3) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[ADDR1:%.*]] = getelementptr i8, ptr addrspace(3) [[ADDR]], i32 10
+; CHECK-NEXT:    [[TMP0:%.*]] = addrspacecast ptr addrspace(3) [[ADDR1]] to ptr
+; CHECK-NEXT:    [[VAL:%.*]] = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr [[TMP0]])
+; CHECK-NEXT:    ret i1 [[VAL]]
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(3) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr %addr1)
+  ret i1 %val
+}
+
+define i1 @test_isspacep_cluster_shared_false(ptr addrspace(1) %addr) {
+; CHECK-LABEL: define i1 @test_isspacep_cluster_shared_false(
+; CHECK-SAME: ptr addrspace(1) [[ADDR:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    ret i1 false
+;
+entry:
+  %addr0 = addrspacecast ptr addrspace(1) %addr to ptr
+  %addr1 = getelementptr i8, ptr %addr0, i32 10
+  %val = call i1 @llvm.nvvm.isspacep.shared.cluster(ptr %addr1)
+  ret i1 %val
+}

github-actions · 2024-10-31T23:22:03Z

✅ With the latest revision this PR passed the C/C++ code formatter.

justinfargnoli

LGTM

gchak · 2024-11-01T03:10:41Z

Could be taken up in a separate PR, but could we get these nvvm.isspacep.* intrinsics documented in the NVPTX usage guide?

AlexMaclean · 2024-11-04T17:14:40Z

@Artem-B When you have a minute could I please get your review on this as well? Thanks!

AlexMaclean · 2024-11-04T17:48:19Z

Could be taken up in a separate PR, but could we get these nvvm.isspacep.* intrinsics documented in the NVPTX usage guide?

I've created #114839 to address this.

AlexMaclean · 2024-11-07T18:16:31Z

@Artem-B ping for review

Artem-B · 2024-11-07T18:33:53Z

llvm/lib/Target/NVPTX/NVPTXTargetTransformInfo.cpp

      if (auto *ASCO = dyn_cast<AddrSpaceCastOperator>(Op0))
        AS = ASCO->getOperand(0)->getType()->getPointerAddressSpace();


Do we still need this attempt to peek through the ASC then? If we do manage to see original AS, so would the infer-address-spaces pass, and it would have folded the intrinsic into a constant.

Yea, I would assume that infer-address-spaces would handle all these cases we do in instcombine, Though maybe doing it in instcombine will allow some cases to happen earlier.

Where do we run AS inference pass in the pipeline? If it's early, then I think it would make sense to keep isspacep folding in one place only, and AS inference pass looks like a better place for that.

Looks like it is only run in llc. I see AMDGPU runs this pass in opt so we should maybe consider doing the same, but I'm not sure how we'd make that decision.

OK. Add a TODO comment there for now and we'll handle it separately.

This change enables constant folding of '`@llvm.nvvm.isspacep.*`' intrinsics if the address space can be propagated in InferAdressSpace.

AlexMaclean requested review from Artem-B and justinfargnoli October 31, 2024 23:18

AlexMaclean self-assigned this Oct 31, 2024

llvmbot added backend:NVPTX llvm:transforms labels Oct 31, 2024

justinfargnoli approved these changes Nov 1, 2024

View reviewed changes

[NVPTX][docs] Add isspacep.* to usage doc

ab58488

Artem-B reviewed Nov 7, 2024

View reviewed changes

Artem-B approved these changes Nov 7, 2024

View reviewed changes

AlexMaclean added 3 commits November 7, 2024 23:02

[NVPTX] Add TTI support for folding isspacep in InferAS

b832e93

address formatting

431f4f7

address comment

05f4e4e

AlexMaclean force-pushed the dev/amaclean/upstream-isspacep-mso-1 branch from 730f9bd to 05f4e4e Compare November 7, 2024 23:02

AlexMaclean merged commit 1febd71 into llvm:main Nov 8, 2024
9 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[NVPTX] Add TTI support for folding isspacep in InferAS #114486

[NVPTX] Add TTI support for folding isspacep in InferAS #114486

Uh oh!

AlexMaclean commented Oct 31, 2024

Uh oh!

llvmbot commented Oct 31, 2024

Uh oh!

llvmbot commented Oct 31, 2024

Uh oh!

github-actions bot commented Oct 31, 2024 •

edited

Loading

Uh oh!

justinfargnoli left a comment

Uh oh!

gchak commented Nov 1, 2024

Uh oh!

AlexMaclean commented Nov 4, 2024

Uh oh!

AlexMaclean commented Nov 4, 2024

Uh oh!

AlexMaclean commented Nov 7, 2024

Uh oh!

Artem-B Nov 7, 2024

Uh oh!

AlexMaclean Nov 7, 2024

Uh oh!

Artem-B Nov 7, 2024

Uh oh!

AlexMaclean Nov 7, 2024

Uh oh!

Artem-B Nov 7, 2024

Uh oh!

Uh oh!

Uh oh!

		if (auto *ASCO = dyn_cast<AddrSpaceCastOperator>(Op0))
		AS = ASCO->getOperand(0)->getType()->getPointerAddressSpace();

[NVPTX] Add TTI support for folding isspacep in InferAS #114486

[NVPTX] Add TTI support for folding isspacep in InferAS #114486

Uh oh!

Conversation

AlexMaclean commented Oct 31, 2024

Uh oh!

llvmbot commented Oct 31, 2024

Uh oh!

llvmbot commented Oct 31, 2024

Uh oh!

github-actions bot commented Oct 31, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

justinfargnoli left a comment

Choose a reason for hiding this comment

Uh oh!

gchak commented Nov 1, 2024

Uh oh!

AlexMaclean commented Nov 4, 2024

Uh oh!

AlexMaclean commented Nov 4, 2024

Uh oh!

AlexMaclean commented Nov 7, 2024

Uh oh!

Artem-B Nov 7, 2024

Choose a reason for hiding this comment

Uh oh!

AlexMaclean Nov 7, 2024

Choose a reason for hiding this comment

Uh oh!

Artem-B Nov 7, 2024

Choose a reason for hiding this comment

Uh oh!

AlexMaclean Nov 7, 2024

Choose a reason for hiding this comment

Uh oh!

Artem-B Nov 7, 2024

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Oct 31, 2024 •

edited

Loading