[Mlir-commits] [mlir] 0c65351 - [mlir][rocdl] Add GlobalLoadAsyncToLDS operation (#165374)

[llvmlistbot at llvm.org]

Author: Pablo Antonio Martinez
Date: 2025-11-04T12:47:18+01:00
New Revision: 0c653514cc2f7e1005fadc330a3f3f904a5a7a52

URL: https://github.com/llvm/llvm-project/commit/0c653514cc2f7e1005fadc330a3f3f904a5a7a52
DIFF: https://github.com/llvm/llvm-project/commit/0c653514cc2f7e1005fadc330a3f3f904a5a7a52.diff

LOG: [mlir][rocdl] Add GlobalLoadAsyncToLDS operation (#165374)

Adds `global.load.async.to.lds` op to rocdl, supporting `b8`, `b32`, `b64` and `b128`. The op is lowered to the appropriate `llvm.amdgcn.global.load.async.to.lds.bXX` intrinsic.

This is available on gfx1250+

Added: 
    

Modified: 
    mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td
    mlir/test/Dialect/LLVMIR/rocdl.mlir
    mlir/test/Target/LLVMIR/rocdl.mlir

Removed: 
    


################################################################################
diff  --git a/mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td b/mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td
index 5241f9a6f2b43..921fdf36a59b0 100644
--- a/mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td
+++ b/mlir/include/mlir/Dialect/LLVMIR/ROCDLOps.td
@@ -692,6 +692,38 @@ def ROCDL_GlobalLoadLDSOp :
   }];
 }
 
+//===---------------------------------------------------------------------===//
+// Async load to LDS intrinsic (available in GFX1250)
+//===---------------------------------------------------------------------===//
+
+foreach bitsVal = [8, 32, 64, 128] in {
+  defvar bitsStr = "b" # !cast<string>(bitsVal);
+  def ROCDL_GlobalLoadAsyncToLDS # !toupper(bitsStr) # Op :
+    ROCDL_IntrOp<"global.load.async.to.lds." # bitsStr, [], [], [], 0, 0, 1, 0, [2, 3], ["offset", "aux"]> {
+    dag args = (ins Arg<ROCDLGlobalBuffer, "", [MemRead]>:$globalPtr,
+                   Arg<ROCDLBufferLDS, "", [MemWrite]>:$ldsPtr,
+                   I32Attr:$offset,
+                   I32Attr:$aux);
+    let arguments = !con(args, baseArgs);
+    let assemblyFormat = [{
+      $globalPtr `,`  $ldsPtr `,` $offset `,` $aux
+      attr-dict `:` type($globalPtr) `,` type($ldsPtr)
+    }];
+    let description = [{
+      Asynchronously loads }] # !cast<string>(bitsVal) # [{ bits of data from a global memory pointer
+      to a Local Data Share (LDS) pointer.
+
+      Available on gfx1250+.
+    }];
+
+    let extraClassDefinition = [{
+      ::llvm::SmallVector<::mlir::Value> $cppClass::getAccessedOperands() {
+        return {getGlobalPtr(), getLdsPtr()};
+      }
+    }];
+  }
+}
+
 //===---------------------------------------------------------------------===//
 // Tensor load/store intrinsics (available in GFX1250)
 //===---------------------------------------------------------------------===//

diff  --git a/mlir/test/Dialect/LLVMIR/rocdl.mlir b/mlir/test/Dialect/LLVMIR/rocdl.mlir
index e703600c71c8e..5e857599b65ea 100644
--- a/mlir/test/Dialect/LLVMIR/rocdl.mlir
+++ b/mlir/test/Dialect/LLVMIR/rocdl.mlir
@@ -664,6 +664,19 @@ llvm.func @rocdl.global.load.lds(%src : !llvm.ptr<1>, %dst: !llvm.ptr<3>) {
   llvm.return
 }
 
+llvm.func @rocdl.global.load.async.to.lds(%src : !llvm.ptr<1>, %dst: !llvm.ptr<3>) {
+  // CHECK-LABEL @rocdl.global.load.async.to.lds
+  // CHECK: rocdl.global.load.async.to.lds.b8 %{{.*}}, %{{.*}}, 0, 0
+  // CHECK: rocdl.global.load.async.to.lds.b32 %{{.*}}, %{{.*}}, 0, 0
+  // CHECK: rocdl.global.load.async.to.lds.b64 %{{.*}}, %{{.*}}, 0, 0
+  // CHECK: rocdl.global.load.async.to.lds.b128 %{{.*}}, %{{.*}}, 0, 0
+  rocdl.global.load.async.to.lds.b8 %src, %dst, 0, 0 : <1>, <3>
+  rocdl.global.load.async.to.lds.b32 %src, %dst, 0, 0 : <1>, <3>
+  rocdl.global.load.async.to.lds.b64 %src, %dst, 0, 0 : <1>, <3>
+  rocdl.global.load.async.to.lds.b128 %src, %dst, 0, 0 : <1>, <3>
+  llvm.return
+}
+
 // CHECK-LABEL @rocdl.tensor.load.to.lds
 llvm.func @rocdl.tensor.load.to.lds(%dgroup0 : vector<4xi32>, %dgroup1 : vector<8xi32>,
                                     %dgroup2 : vector<4xi32>, %dgroup3 : vector<4xi32>) {

diff  --git a/mlir/test/Target/LLVMIR/rocdl.mlir b/mlir/test/Target/LLVMIR/rocdl.mlir
index 8a848221a50dd..3fbd9e0567948 100644
--- a/mlir/test/Target/LLVMIR/rocdl.mlir
+++ b/mlir/test/Target/LLVMIR/rocdl.mlir
@@ -1040,6 +1040,19 @@ llvm.func @rocdl.global.load.lds(%src : !llvm.ptr<1>, %dst: !llvm.ptr<3>) {
   llvm.return
 }
 
+// CHECK-LABEL: rocdl.global.load.async.to.lds
+llvm.func @rocdl.global.load.async.to.lds(%src : !llvm.ptr<1>, %dst: !llvm.ptr<3>) {
+  // CHECK: call void @llvm.amdgcn.global.load.async.to.lds.b8
+  rocdl.global.load.async.to.lds.b8 %src, %dst, 0, 0 : !llvm.ptr<1>, !llvm.ptr<3>
+  // CHECK: call void @llvm.amdgcn.global.load.async.to.lds.b32
+  rocdl.global.load.async.to.lds.b32 %src, %dst, 0, 0 : !llvm.ptr<1>, !llvm.ptr<3>
+  // CHECK: call void @llvm.amdgcn.global.load.async.to.lds.b64
+  rocdl.global.load.async.to.lds.b64 %src, %dst, 0, 0 : !llvm.ptr<1>, !llvm.ptr<3>
+  // CHECK: call void @llvm.amdgcn.global.load.async.to.lds.b128
+  rocdl.global.load.async.to.lds.b128 %src, %dst, 0, 0 : !llvm.ptr<1>, !llvm.ptr<3>
+  llvm.return
+}
+
 // CHECK-LABEL: rocdl.tensor.load.to.lds
 llvm.func @rocdl.tensor.load.to.lds(%dgroup0 : vector<4xi32>, %dgroup1 : vector<8xi32>,
                                     %dgroup2 : vector<4xi32>, %dgroup3 : vector<4xi32>) {