[mlir][ArmSVE] Restructure sources to match ArmSME dialect (NFC) by MacDue · Pull Request #68399 · llvm/llvm-project

MacDue · 2023-10-06T09:43:55Z

This rearranges the Arm SVE dialect to have the same structure of the Arm SME dialect. So this just moves around some source files and adds a ArmSVE_IntrOp base class for SVE intrinsics. This makes later changes a little easier and more consistent other dialects.

llvmbot · 2023-10-06T09:44:58Z

@llvm/pr-subscribers-mlir-sve
@llvm/pr-subscribers-mlir-llvm

@llvm/pr-subscribers-mlir

Changes

This rearranges the Arm SVE dialect to have the same structure of the Arm SME dialect. So this just moves around some source files and adds a ArmSVE_IntrOp base class for SVE intrinsics. This makes later changes a little easier and more consistent other dialects.

Full diff: https://github.com/llvm/llvm-project/pull/68399.diff

10 Files Affected:

(modified) mlir/include/mlir/Dialect/ArmSVE/CMakeLists.txt (+1-6)
(renamed) mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td (+11-5)
(renamed) mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h (+2-2)
(added) mlir/include/mlir/Dialect/ArmSVE/IR/CMakeLists.txt (+6)
(renamed) mlir/include/mlir/Dialect/ArmSVE/Transforms/Transforms.h ()
(modified) mlir/include/mlir/InitAllDialects.h (+1-1)
(modified) mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp (+2-2)
(modified) mlir/lib/Dialect/ArmSVE/IR/ArmSVEDialect.cpp (+5-5)
(modified) mlir/lib/Dialect/ArmSVE/Transforms/LegalizeForLLVMExport.cpp (+2-2)
(modified) mlir/lib/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.cpp (+2-2)

diff --git a/mlir/include/mlir/Dialect/ArmSVE/CMakeLists.txt b/mlir/include/mlir/Dialect/ArmSVE/CMakeLists.txt
index 06595b7088a1e1b..f33061b2d87cffc 100644
--- a/mlir/include/mlir/Dialect/ArmSVE/CMakeLists.txt
+++ b/mlir/include/mlir/Dialect/ArmSVE/CMakeLists.txt
@@ -1,6 +1 @@
-add_mlir_dialect(ArmSVE arm_sve ArmSVE)
-add_mlir_doc(ArmSVE ArmSVE Dialects/ -gen-dialect-doc -dialect=arm_sve)
-
-set(LLVM_TARGET_DEFINITIONS ArmSVE.td)
-mlir_tablegen(ArmSVEConversions.inc -gen-llvmir-conversions)
-add_public_tablegen_target(MLIRArmSVEConversionsIncGen)
+add_subdirectory(IR)
diff --git a/mlir/include/mlir/Dialect/ArmSVE/ArmSVE.td b/mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td
similarity index 95%
rename from mlir/include/mlir/Dialect/ArmSVE/ArmSVE.td
rename to mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td
index 5c86df1ef21f4b9..22f57a21fb9ae24 100644
--- a/mlir/include/mlir/Dialect/ArmSVE/ArmSVE.td
+++ b/mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td
@@ -28,7 +28,6 @@ def ArmSVE_Dialect : Dialect {
     This dialect contains the definitions necessary to target specific Arm SVE
     scalable vector operations.
   }];
-
 }
 
 //===----------------------------------------------------------------------===//
@@ -38,16 +37,23 @@ def ArmSVE_Dialect : Dialect {
 class ArmSVE_Op<string mnemonic, list<Trait> traits = []> :
   Op<ArmSVE_Dialect, mnemonic, traits> {}
 
-class ArmSVE_IntrBinaryOverloadedOp<string mnemonic,
-                                    list<Trait> traits = []> :
+class ArmSVE_IntrOp<string mnemonic,
+                                    list<Trait> traits = [],
+                                    list<int> overloadedOperands = [],
+                                    list<int> overloadedResults = []> :
   LLVM_IntrOpBase</*Dialect dialect=*/ArmSVE_Dialect,
                   /*string opName=*/"intr." # mnemonic,
                   /*string enumName=*/"aarch64_sve_" # !subst(".", "_", mnemonic),
-                  /*list<int> overloadedResults=*/[0],
-                  /*list<int> overloadedOperands=*/[], // defined by result overload
+                  /*list<int> overloadedResults=*/overloadedResults,
+                  /*list<int> overloadedOperands=*/overloadedOperands,
                   /*list<Trait> traits=*/traits,
                   /*int numResults=*/1>;
 
+class ArmSVE_IntrBinaryOverloadedOp<string mnemonic,
+                                    list<Trait> traits = []>:
+  ArmSVE_IntrOp<mnemonic, traits,
+    /*overloadedOperands=*/[], /*overloadedResults=*/[0]>;
+
 class ScalableMaskedFOp<string mnemonic, string op_description,
                         list<Trait> traits = []> :
   ArmSVE_Op<mnemonic, !listconcat(traits,
diff --git a/mlir/include/mlir/Dialect/ArmSVE/ArmSVEDialect.h b/mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h
similarity index 89%
rename from mlir/include/mlir/Dialect/ArmSVE/ArmSVEDialect.h
rename to mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h
index 6ecee1e4552b95c..a424e109b06d3e8 100644
--- a/mlir/include/mlir/Dialect/ArmSVE/ArmSVEDialect.h
+++ b/mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h
@@ -19,9 +19,9 @@
 #include "mlir/IR/OpDefinition.h"
 #include "mlir/Interfaces/SideEffectInterfaces.h"
 
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h.inc"
 
 #define GET_OP_CLASSES
-#include "mlir/Dialect/ArmSVE/ArmSVE.h.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVE.h.inc"
 
 #endif // MLIR_DIALECT_ARMSVE_ARMSVEDIALECT_H
diff --git a/mlir/include/mlir/Dialect/ArmSVE/IR/CMakeLists.txt b/mlir/include/mlir/Dialect/ArmSVE/IR/CMakeLists.txt
new file mode 100644
index 000000000000000..06595b7088a1e1b
--- /dev/null
+++ b/mlir/include/mlir/Dialect/ArmSVE/IR/CMakeLists.txt
@@ -0,0 +1,6 @@
+add_mlir_dialect(ArmSVE arm_sve ArmSVE)
+add_mlir_doc(ArmSVE ArmSVE Dialects/ -gen-dialect-doc -dialect=arm_sve)
+
+set(LLVM_TARGET_DEFINITIONS ArmSVE.td)
+mlir_tablegen(ArmSVEConversions.inc -gen-llvmir-conversions)
+add_public_tablegen_target(MLIRArmSVEConversionsIncGen)
diff --git a/mlir/include/mlir/Dialect/ArmSVE/Transforms.h b/mlir/include/mlir/Dialect/ArmSVE/Transforms/Transforms.h
similarity index 100%
rename from mlir/include/mlir/Dialect/ArmSVE/Transforms.h
rename to mlir/include/mlir/Dialect/ArmSVE/Transforms/Transforms.h
diff --git a/mlir/include/mlir/InitAllDialects.h b/mlir/include/mlir/InitAllDialects.h
index 5ec36a7f289e586..d04ed373ecf045a 100644
--- a/mlir/include/mlir/InitAllDialects.h
+++ b/mlir/include/mlir/InitAllDialects.h
@@ -24,7 +24,7 @@
 #include "mlir/Dialect/Arith/Transforms/BufferizableOpInterfaceImpl.h"
 #include "mlir/Dialect/ArmNeon/ArmNeonDialect.h"
 #include "mlir/Dialect/ArmSME/IR/ArmSME.h"
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h"
 #include "mlir/Dialect/Async/IR/Async.h"
 #include "mlir/Dialect/Bufferization/IR/Bufferization.h"
 #include "mlir/Dialect/Bufferization/Transforms/FuncBufferizableOpInterfaceImpl.h"
diff --git a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
index 2929823bad32adf..b865a2671fff762 100644
--- a/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
+++ b/mlir/lib/Conversion/VectorToLLVM/ConvertVectorToLLVMPass.cpp
@@ -17,8 +17,8 @@
 #include "mlir/Dialect/ArmSME/IR/ArmSME.h"
 #include "mlir/Dialect/ArmSME/Transforms/Passes.h"
 #include "mlir/Dialect/ArmSME/Transforms/Transforms.h"
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
-#include "mlir/Dialect/ArmSVE/Transforms.h"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h"
+#include "mlir/Dialect/ArmSVE/Transforms/Transforms.h"
 #include "mlir/Dialect/Func/IR/FuncOps.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/Dialect/MemRef/IR/MemRef.h"
diff --git a/mlir/lib/Dialect/ArmSVE/IR/ArmSVEDialect.cpp b/mlir/lib/Dialect/ArmSVE/IR/ArmSVEDialect.cpp
index 4af836a93c2a169..b7f1020deba1e40 100644
--- a/mlir/lib/Dialect/ArmSVE/IR/ArmSVEDialect.cpp
+++ b/mlir/lib/Dialect/ArmSVE/IR/ArmSVEDialect.cpp
@@ -10,7 +10,7 @@
 //
 //===----------------------------------------------------------------------===//
 
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h"
 #include "mlir/Dialect/LLVMIR/LLVMTypes.h"
 #include "mlir/IR/Builders.h"
 #include "mlir/IR/DialectImplementation.h"
@@ -38,17 +38,17 @@ static Type getI1SameShape(Type type) {
 // Tablegen Definitions
 //===----------------------------------------------------------------------===//
 
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.cpp.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.cpp.inc"
 
 #define GET_OP_CLASSES
-#include "mlir/Dialect/ArmSVE/ArmSVE.cpp.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVE.cpp.inc"
 
 #define GET_TYPEDEF_CLASSES
-#include "mlir/Dialect/ArmSVE/ArmSVETypes.cpp.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVETypes.cpp.inc"
 
 void ArmSVEDialect::initialize() {
   addOperations<
 #define GET_OP_LIST
-#include "mlir/Dialect/ArmSVE/ArmSVE.cpp.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVE.cpp.inc"
       >();
 }
diff --git a/mlir/lib/Dialect/ArmSVE/Transforms/LegalizeForLLVMExport.cpp b/mlir/lib/Dialect/ArmSVE/Transforms/LegalizeForLLVMExport.cpp
index b6723a52e177f5f..abbb978304068e2 100644
--- a/mlir/lib/Dialect/ArmSVE/Transforms/LegalizeForLLVMExport.cpp
+++ b/mlir/lib/Dialect/ArmSVE/Transforms/LegalizeForLLVMExport.cpp
@@ -8,8 +8,8 @@
 
 #include "mlir/Conversion/LLVMCommon/ConversionTarget.h"
 #include "mlir/Conversion/LLVMCommon/Pattern.h"
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
-#include "mlir/Dialect/ArmSVE/Transforms.h"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h"
+#include "mlir/Dialect/ArmSVE/Transforms/Transforms.h"
 #include "mlir/Dialect/Func/IR/FuncOps.h"
 #include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/IR/BuiltinOps.h"
diff --git a/mlir/lib/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.cpp b/mlir/lib/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.cpp
index bc1f0e934fa02f4..cd10811b68f0288 100644
--- a/mlir/lib/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.cpp
+++ b/mlir/lib/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.cpp
@@ -11,7 +11,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "mlir/Target/LLVMIR/Dialect/ArmSVE/ArmSVEToLLVMIRTranslation.h"
-#include "mlir/Dialect/ArmSVE/ArmSVEDialect.h"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEDialect.h"
 #include "mlir/IR/Operation.h"
 #include "mlir/Target/LLVMIR/ModuleTranslation.h"
 
@@ -35,7 +35,7 @@ class ArmSVEDialectLLVMIRTranslationInterface
   convertOperation(Operation *op, llvm::IRBuilderBase &builder,
                    LLVM::ModuleTranslation &moduleTranslation) const final {
     Operation &opInst = *op;
-#include "mlir/Dialect/ArmSVE/ArmSVEConversions.inc"
+#include "mlir/Dialect/ArmSVE/IR/ArmSVEConversions.inc"
 
     return failure();
   }

c-rhodes

one minor comment otherwise LGTM, cheers

Co-authored-by: Cullen Rhodes <cullen.rhodes@arm.com>

These will be used in future pass to ensure that loads/stores of masks are legal (as the LLVM backend does not support this for any type smaller than an svbool, which is vector<[16]xi1>). Depends on #68399

MacDue requested review from banach-space and c-rhodes October 6, 2023 09:43

llvmbot added mlir:llvm mlir mlir:sve labels Oct 6, 2023

MacDue mentioned this pull request Oct 6, 2023

[mlir][ArmSVE] Add convert.from/to.svbool intrinsics #68418

Merged

dcaballe approved these changes Oct 6, 2023

View reviewed changes

c-rhodes approved these changes Oct 8, 2023

View reviewed changes

Comment thread mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td Outdated

Update mlir/include/mlir/Dialect/ArmSVE/IR/ArmSVE.td

b24b31e

Co-authored-by: Cullen Rhodes <cullen.rhodes@arm.com>

MacDue merged commit 7bbfd2a into llvm:main Oct 9, 2023

stepthomas mentioned this pull request Oct 10, 2023

AMDGPU stepthomas atomic csub no rtn forms ver2 stepthomas/llvm-project#1

Closed

Navigation Menu

Uh oh!

[mlir][ArmSVE] Restructure sources to match ArmSME dialect (NFC) - #68399

[mlir][ArmSVE] Restructure sources to match ArmSME dialect (NFC)#68399
MacDue merged 2 commits into
llvm:mainfrom
MacDue:restructure_arm_sve

MacDue commented Oct 6, 2023

Uh oh!

llvmbot commented Oct 6, 2023 •

edited

Loading

Uh oh!

c-rhodes left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

MacDue commented Oct 6, 2023

Uh oh!

llvmbot commented Oct 6, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

c-rhodes left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

llvmbot commented Oct 6, 2023 •

edited

Loading