diff --git a/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_cooperative.hpp b/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_cooperative.hpp
index fefbf831..1c314aad 100644
--- a/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_cooperative.hpp
+++ b/include/cutlass/gemm/kernel/sm90_gemm_array_tma_warpspecialized_cooperative.hpp
@@ -465,6 +465,9 @@ public:
 
     TileScheduler scheduler{params.scheduler};
     auto work_tile_info = scheduler.get_current_work();
+    if (not work_tile_info.is_valid()) {
+      return;
+    }
 
     // Optionally append 1s until problem shape is rank-4 in case it is only rank-3 (MNK)
     auto problem_shape_MNKL = append<4>(params.problem_shape.get_problem_shape(work_tile_info.L_idx), Int<1>{});
@@ -623,7 +626,9 @@ public:
           // Get next work tile
           work_tile_info = fetch_next_work(work_tile_info, scheduler);
           if constexpr (IsGroupedGemmKernel) {
-            problem_shape_MNKL = append<4>(params.problem_shape.get_problem_shape(work_tile_info.L_idx), Int<1>{});
+            if (work_tile_info.is_valid()) {
+              problem_shape_MNKL = append<4>(params.problem_shape.get_problem_shape(work_tile_info.L_idx), Int<1>{});
+            }
           }
         } // Scheduler work fetch loop
 
@@ -702,7 +707,9 @@ public:
         // Get next work tile
         work_tile_info = fetch_next_work(work_tile_info, scheduler);
         if constexpr (IsGroupedGemmKernel) {
-          problem_shape_MNKL = append<4>(params.problem_shape.get_problem_shape(work_tile_info.L_idx), Int<1>{});
+          if (work_tile_info.is_valid()) {
+            problem_shape_MNKL = append<4>(params.problem_shape.get_problem_shape(work_tile_info.L_idx), Int<1>{});
+          }
         }
       } // Scheduler work fetch loop