diff --git a/include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_whole_k_prefetch_trload.hpp b/include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_whole_k_prefetch_trload.hpp
index a6056bcc99..e3ef73332b 100644
--- a/include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_whole_k_prefetch_trload.hpp
+++ b/include/ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_whole_k_prefetch_trload.hpp
@@ -678,17 +678,6 @@ struct BlockFmhaPipelineQRKSVSWholeKPrefetchTrLoad
 
             __builtin_amdgcn_sched_barrier(0x00000001);
 
-            if constexpr(!kPreloadWholeNextIterationK)
-            {
-                if(seqlen_k_curr < seqlen_k_end)
-                {
-                    k_tiles[I0] = load_tile(k_dram_window);
-                    move_tile_window(k_dram_window, {kN0Sub, 0});
-                };
-            }
-
-            __builtin_amdgcn_sched_barrier(0x00000001);
-
             // STAGE 3, Gemm_1 ( O = P@V )
             static_for<0, k1_loops, 1>{}([&](auto i_k1) {
                 if constexpr(i_k1 < k1_loops - NumPrefetchV)
@@ -697,6 +686,18 @@ struct BlockFmhaPipelineQRKSVSWholeKPrefetchTrLoad
                     move_tile_window(v_dram_window, {kK1, 0});
                 };
 
+                if constexpr(i_k1 == k1_loops - NumPrefetchV)
+                {
+                    if constexpr(!kPreloadWholeNextIterationK)
+                    {
+                        if(seqlen_k_curr < seqlen_k_end)
+                        {
+                            k_tiles[I0] = load_tile(k_dram_window);
+                            move_tile_window(k_dram_window, {kN0Sub, 0});
+                        };
+                    }
+                };
+
                 block_sync_lds();
                 gemm_1(
                     o_acc,