diff --git a/compiler/include/byteir/Dialect/mhlo/Analysis/ShapeAnalysis.h b/compiler/include/byteir/Dialect/mhlo/Analysis/ShapeAnalysis.h
index 613b1cd91..f156d690c 100644
--- a/compiler/include/byteir/Dialect/mhlo/Analysis/ShapeAnalysis.h
+++ b/compiler/include/byteir/Dialect/mhlo/Analysis/ShapeAnalysis.h
@@ -144,7 +144,8 @@ class MhloShapeAnalysisBase : public ShapeAnalysis<ShapeKnowledgeType> {
                           wrapperShapeValueKnowledges);
 
     return inferFunc(op->getContext(), op->getLoc(), range,
-                     op->getAttrDictionary(), op->getPropertiesStorage(), op->getRegions(), results);
+                     op->getAttrDictionary(), op->getPropertiesStorage(),
+                     op->getRegions(), results);
   }
 };
 
diff --git a/compiler/test/Dialect/SCF/forallCollapsing.mlir b/compiler/test/Dialect/SCF/forallCollapsing.mlir
index 5ce12e26d..d71152ea6 100644
--- a/compiler/test/Dialect/SCF/forallCollapsing.mlir
+++ b/compiler/test/Dialect/SCF/forallCollapsing.mlir
@@ -25,9 +25,9 @@ func.func @Elementwise(%arg0: memref<32x1024x?x30xf32>) -> memref<32768x?x30xf32
 	%dim = memref.dim %arg0, %c2 : memref<32x1024x?x30xf32>
 	%alloc = memref.alloc(%dim) : memref<32768x?x30xf32>
 	scf.forall (%arg1, %arg2, %arg3) in (32768, %dim, 30) {
-		%subview = memref.subview %collapse_shape[%arg1, %arg2, %arg3] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>
-		%subview_0 = memref.subview %alloc[%arg1, %arg2, %arg3] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>
-		linalg.generic {indexing_maps = [#map, #map], iterator_types = ["parallel", "parallel", "parallel"]} ins(%subview : memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>) outs(%subview_0 : memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>) attrs =  {__byteir_gpu_tile_elementwise_0} {
+		%subview = memref.subview %collapse_shape[%arg1, %arg2, %arg3] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>
+		%subview_0 = memref.subview %alloc[%arg1, %arg2, %arg3] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>
+		linalg.generic {indexing_maps = [#map, #map], iterator_types = ["parallel", "parallel", "parallel"]} ins(%subview : memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>) outs(%subview_0 : memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>) attrs =  {__byteir_gpu_tile_elementwise_0} {
 		^bb0(%in: f32, %out: f32):
 			%0 = arith.mulf %in, %in : f32
 			linalg.yield %0 : f32
@@ -49,6 +49,6 @@ func.func @Elementwise(%arg0: memref<32x1024x?x30xf32>) -> memref<32768x?x30xf32
 // CHECK-NEXT: %[[V2:.*]] = arith.divsi %arg1, %[[C30]] : index
 // CHECK-NEXT: %[[V3:.*]] = arith.remsi %[[V2]], %[[DIM]] : index
 // CHECK-NEXT: %[[V4:.*]] = arith.divsi %[[V2]], %[[DIM]] : index
-// CHECK-NEXT: %[[SUBVIEW:.*]] = memref.subview %[[COLLAPSE]][%[[V4]], %[[V3]], %[[V1]]] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>
-// CHECK-NEXT: %[[SUBVIEW_0:.*]] = memref.subview %[[ALLOC]][%[[V4]], %[[V3]], %[[V1]]] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>
-// CHCCK-NEXT: linalg.generic {indexing_maps = [#map, #map], iterator_types = ["parallel", "parallel", "parallel"]} ins(%[[SUBVIEW]] : memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>) outs(%[[SUBVIEW_0]] : memref<1x1x1xf32, strided<[300, 30, 1], offset: ?>>) attrs =  {__byteir_gpu_tile_elementwise_0} {
+// CHECK-NEXT: %[[SUBVIEW:.*]] = memref.subview %[[COLLAPSE]][%[[V4]], %[[V3]], %[[V1]]] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>
+// CHECK-NEXT: %[[SUBVIEW_0:.*]] = memref.subview %[[ALLOC]][%[[V4]], %[[V3]], %[[V1]]] [1, 1, 1] [1, 1, 1] : memref<32768x?x30xf32> to memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>
+// CHCCK-NEXT: linalg.generic {indexing_maps = [#map, #map], iterator_types = ["parallel", "parallel", "parallel"]} ins(%[[SUBVIEW]] : memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>) outs(%[[SUBVIEW_0]] : memref<1x1x1xf32, strided<[?, 30, 1], offset: ?>>) attrs =  {__byteir_gpu_tile_elementwise_0} {