[mlir] AMDGPUToROCDL: handle 1-element vectors #128266
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Buffer intrinsics doesn't support 1-element vectors, cast them to scalars.
|
@llvm/pr-subscribers-backend-amdgpu @llvm/pr-subscribers-mlir Author: Ivan Butygin (Hardcode84) ChangesBuffer intrinsics doesn't support 1-element vectors, cast them to scalars. Full diff: https://github.com/llvm/llvm-project/pull/128266.diff 2 Files Affected:
diff --git a/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp b/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
index b29228ef87ea7..b8574bbbee345 100644
--- a/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
+++ b/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
@@ -167,6 +167,12 @@ struct RawBufferOpLowering : public ConvertOpToLLVMPattern<GpuOp> {
}
}
}
+ if (auto vecType = dyn_cast<VectorType>(llvmBufferValType)) {
+ // Buffer intrinsics doesn't support 1-element vectors, cast them to
+ // scalars.
+ if (vecType.getNumElements() == 1)
+ llvmBufferValType = vecType.getElementType();
+ }
SmallVector<Value, 6> args;
if (storeData) {
diff --git a/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir b/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
index 062b63c076c3c..8b2f5788721a1 100644
--- a/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
+++ b/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
@@ -76,6 +76,15 @@ func.func @gpu_gcn_raw_buffer_load_i32_oob_off(%buf: memref<64xi32>, %idx: i32)
func.return %0 : i32
}
+// CHECK-LABEL: func @gpu_gcn_raw_buffer_load_1xi32
+func.func @gpu_gcn_raw_buffer_load_1xi32(%buf: memref<64xi32>, %idx: i32) -> vector<1xi32> {
+ // CHECK: %[[ret:.*]] = rocdl.raw.ptr.buffer.load %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : i32
+ // CHECK: %[[cast:.*]] = llvm.bitcast %[[ret]] : i32 to vector<1xi32>
+ // CHECK: return %[[cast]]
+ %0 = amdgpu.raw_buffer_load {boundsCheck = true} %buf[%idx] : memref<64xi32>, i32 -> vector<1xi32>
+ func.return %0 : vector<1xi32>
+}
+
// CHECK-LABEL: func @gpu_gcn_raw_buffer_load_2xi32
func.func @gpu_gcn_raw_buffer_load_2xi32(%buf: memref<64xi32>, %idx: i32) -> vector<2xi32> {
// CHECK: %[[ret:.*]] = rocdl.raw.ptr.buffer.load %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : vector<2xi32>
@@ -159,6 +168,14 @@ func.func @gpu_gcn_raw_buffer_store_i32(%value: i32, %buf: memref<64xi32>, %idx:
func.return
}
+// CHECK-LABEL: func @gpu_gcn_raw_buffer_store_1xf32
+func.func @gpu_gcn_raw_buffer_store_1xf32(%value: vector<1xf32>, %buf: memref<64xf32>, %idx: i32) {
+ // CHECK: %[[cast:.*]] = llvm.bitcast %{{.*}} : vector<1xf32> to f32
+ // CHECK: rocdl.raw.ptr.buffer.store %[[cast]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : f32
+ amdgpu.raw_buffer_store {boundsCheck = true} %value -> %buf[%idx] : vector<1xf32> -> memref<64xf32>, i32
+ func.return
+}
+
// CHECK-LABEL: func @gpu_gcn_raw_buffer_store_2xi8
func.func @gpu_gcn_raw_buffer_store_2xi8(%value: vector<2xi8>, %buf: memref<64xi8>, %idx: i32) {
// CHECK: %[[cast:.*]] = llvm.bitcast %{{.*}} : vector<2xi8> to i16
|
|
@llvm/pr-subscribers-mlir-gpu Author: Ivan Butygin (Hardcode84) ChangesBuffer intrinsics doesn't support 1-element vectors, cast them to scalars. Full diff: https://github.com/llvm/llvm-project/pull/128266.diff 2 Files Affected:
diff --git a/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp b/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
index b29228ef87ea7..b8574bbbee345 100644
--- a/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
+++ b/mlir/lib/Conversion/AMDGPUToROCDL/AMDGPUToROCDL.cpp
@@ -167,6 +167,12 @@ struct RawBufferOpLowering : public ConvertOpToLLVMPattern<GpuOp> {
}
}
}
+ if (auto vecType = dyn_cast<VectorType>(llvmBufferValType)) {
+ // Buffer intrinsics doesn't support 1-element vectors, cast them to
+ // scalars.
+ if (vecType.getNumElements() == 1)
+ llvmBufferValType = vecType.getElementType();
+ }
SmallVector<Value, 6> args;
if (storeData) {
diff --git a/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir b/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
index 062b63c076c3c..8b2f5788721a1 100644
--- a/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
+++ b/mlir/test/Conversion/AMDGPUToROCDL/amdgpu-to-rocdl.mlir
@@ -76,6 +76,15 @@ func.func @gpu_gcn_raw_buffer_load_i32_oob_off(%buf: memref<64xi32>, %idx: i32)
func.return %0 : i32
}
+// CHECK-LABEL: func @gpu_gcn_raw_buffer_load_1xi32
+func.func @gpu_gcn_raw_buffer_load_1xi32(%buf: memref<64xi32>, %idx: i32) -> vector<1xi32> {
+ // CHECK: %[[ret:.*]] = rocdl.raw.ptr.buffer.load %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : i32
+ // CHECK: %[[cast:.*]] = llvm.bitcast %[[ret]] : i32 to vector<1xi32>
+ // CHECK: return %[[cast]]
+ %0 = amdgpu.raw_buffer_load {boundsCheck = true} %buf[%idx] : memref<64xi32>, i32 -> vector<1xi32>
+ func.return %0 : vector<1xi32>
+}
+
// CHECK-LABEL: func @gpu_gcn_raw_buffer_load_2xi32
func.func @gpu_gcn_raw_buffer_load_2xi32(%buf: memref<64xi32>, %idx: i32) -> vector<2xi32> {
// CHECK: %[[ret:.*]] = rocdl.raw.ptr.buffer.load %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : vector<2xi32>
@@ -159,6 +168,14 @@ func.func @gpu_gcn_raw_buffer_store_i32(%value: i32, %buf: memref<64xi32>, %idx:
func.return
}
+// CHECK-LABEL: func @gpu_gcn_raw_buffer_store_1xf32
+func.func @gpu_gcn_raw_buffer_store_1xf32(%value: vector<1xf32>, %buf: memref<64xf32>, %idx: i32) {
+ // CHECK: %[[cast:.*]] = llvm.bitcast %{{.*}} : vector<1xf32> to f32
+ // CHECK: rocdl.raw.ptr.buffer.store %[[cast]], %{{.*}}, %{{.*}}, %{{.*}}, %{{.*}} : f32
+ amdgpu.raw_buffer_store {boundsCheck = true} %value -> %buf[%idx] : vector<1xf32> -> memref<64xf32>, i32
+ func.return
+}
+
// CHECK-LABEL: func @gpu_gcn_raw_buffer_store_2xi8
func.func @gpu_gcn_raw_buffer_store_2xi8(%value: vector<2xi8>, %buf: memref<64xi8>, %idx: i32) {
// CHECK: %[[cast:.*]] = llvm.bitcast %{{.*}} : vector<2xi8> to i16
|
krzysz00
approved these changes
Feb 23, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Buffer intrinsics doesn't support 1-element vectors, cast them to scalars.