diff --git a/compiler/src/iree/compiler/DispatchCreation/FormDispatchRegions.cpp b/compiler/src/iree/compiler/DispatchCreation/FormDispatchRegions.cpp
index 25163fa85c58..dba316994288 100644
--- a/compiler/src/iree/compiler/DispatchCreation/FormDispatchRegions.cpp
+++ b/compiler/src/iree/compiler/DispatchCreation/FormDispatchRegions.cpp
@@ -303,6 +303,10 @@ FusionGroup::getRootParallelLoopToOpMap(Operation *op) const {
       if (failed(composedMap) || (newMap && composedMap != newMap)) {
         return failure();
       }
+      if (composedMap.value().getNumResults() ==
+          composedMap.value().getNumOfZeroResults()) {
+        return failure();
+      }
       newMap = composedMap.value();
     }
   } else {
@@ -336,6 +340,8 @@ FusionGroup::getRootParallelLoopToOpMap(Operation *op) const {
       }
     }
   }
+
+  // Fail if there is no mapping or if there are no parallel loops in common.
   if (!newMap) {
     return failure();
   }
diff --git a/compiler/src/iree/compiler/DispatchCreation/test/form_dispatch_regions.mlir b/compiler/src/iree/compiler/DispatchCreation/test/form_dispatch_regions.mlir
index c9ac66f7c207..fcffa4520167 100644
--- a/compiler/src/iree/compiler/DispatchCreation/test/form_dispatch_regions.mlir
+++ b/compiler/src/iree/compiler/DispatchCreation/test/form_dispatch_regions.mlir
@@ -1875,3 +1875,38 @@ util.func public @no_fusion_use_from_above(%arg0 : tensor<?x?xf32>,
 //       CHECK:       tensor.extract %[[DISPATCH0]]
 //       CHECK:     flow.return %[[CONSUMER]]
 //       CHECK:   util.return %[[DISPATCH1]]
+
+// -----
+
+util.func public @dont_fuse_no_shared_parallel_loops(%arg0: tensor<16x16x24xf32>, %arg1: tensor<64x3x32xf32>) -> (tensor<64x3x32xf32>, tensor<32x64x3xf32>) {
+  %cst = arith.constant 0.000000e+00 : f32
+  %0 = tensor.empty() : tensor<f32>
+  %1 = tensor.empty() : tensor<64x3x32xf32>
+  %2 = tensor.empty() : tensor<32x64x3xf32>
+  %3 = linalg.fill ins(%cst : f32) outs(%0 : tensor<f32>) -> tensor<f32>
+  %4 = linalg.generic {indexing_maps = [affine_map<(d0, d1, d2) -> (d0, d1, d2)>, affine_map<(d0, d1, d2) -> ()>], iterator_types = ["reduction", "reduction", "reduction"]} ins(%arg0 : tensor<16x16x24xf32>) outs(%3 : tensor<f32>) {
+  ^bb0(%in: f32, %out: f32):
+    %6 = arith.mulf %in, %in : f32
+    %7 = arith.addf %out, %6 : f32
+    linalg.yield %7 : f32
+  } -> tensor<f32>
+  %5:2 = linalg.generic {indexing_maps = [affine_map<(d0, d1, d2) -> (d0, d1, d2)>, affine_map<(d0, d1, d2) -> ()>, affine_map<(d0, d1, d2) -> (d0, d1, d2)>, affine_map<(d0, d1, d2) -> (d2, d0, d1)>], iterator_types = ["parallel", "parallel", "parallel"]} ins(%arg1, %4 : tensor<64x3x32xf32>, tensor<f32>) outs(%1, %2 : tensor<64x3x32xf32>, tensor<32x64x3xf32>) {
+  ^bb0(%in: f32, %in_0: f32, %out: f32, %out_1: f32):
+    %6 = arith.addf %in, %in_0 : f32
+    linalg.yield %6, %6 : f32, f32
+  } -> (tensor<64x3x32xf32>, tensor<32x64x3xf32>)
+  util.return %5#0, %5#1 : tensor<64x3x32xf32>, tensor<32x64x3xf32>
+}
+
+// CHECK-LABEL: util.func public @dont_fuse_no_shared_parallel_loops(
+//  CHECK-SAME:     %[[ARG0:[a-zA-Z0-9]+]]: tensor<16x16x24xf32>
+//  CHECK-SAME:     %[[ARG1:[a-zA-Z0-9]+]]: tensor<64x3x32xf32>)
+//       CHECK:   %[[DISPATCH0:.+]] = flow.dispatch.region
+//       CHECK:   %[[REDUCTION:.+]] = linalg.generic
+//  CHECK-SAME:       ins(%[[ARG0]]
+//       CHECK:     flow.return %[[REDUCTION]]
+//       CHECK:   %[[DISPATCH1:.+]]:2 = flow.dispatch.region
+//       CHECK:     %[[GENERIC:.+]]:2 = linalg.generic
+//  CHECK-SAME:       ins(%[[ARG1]], %[[DISPATCH0]]
+//       CHECK:     flow.return %[[GENERIC]]#0, %[[GENERIC]]#1
+//       CHECK:   util.return %[[DISPATCH1]]#0, %[[DISPATCH1]]#1
diff --git a/tests/external/iree-test-suites/sharktank_models/benchmarks/sdxl/clip_rocm.json b/tests/external/iree-test-suites/sharktank_models/benchmarks/sdxl/clip_rocm.json
index d9c94118bcd1..1d3fbbd6a438 100644
--- a/tests/external/iree-test-suites/sharktank_models/benchmarks/sdxl/clip_rocm.json
+++ b/tests/external/iree-test-suites/sharktank_models/benchmarks/sdxl/clip_rocm.json
@@ -23,9 +23,9 @@
         "mi308": 7.45
     },
     "golden_dispatch": {
-        "mi250": 792,
-        "mi300": 792,
-        "mi308": 792
+        "mi250": 794,
+        "mi300": 794,
+        "mi308": 794
     },
     "golden_size": {
         "mi250": 460000,