Add thread cluster lengths to kernel name.

include/ck/tensor_operation/gpu/device/impl/device_grouped_gemm_xdl_splitk_cshuffle_tile_loop.hpp

@@ -732,6 +732,7 @@ struct DeviceGroupedGemmXdlSplitKCShuffle : public DeviceGroupedGemmSplitK<ALayo
             << BBlockTransferSrcScalarPerVector << ", "
             << CShuffleMXdlPerWavePerShuffle << ", "
             << CShuffleNXdlPerWavePerShuffle << ", "
+            << ABlockTransferThreadClusterLengths_K0_M_K1{} << ", "
             << getGemmSpecializationString(GemmSpec)
             << ">";
         // clang-format on

include/ck/utility/sequence.hpp

@@ -897,3 +897,13 @@ template <index_t NSize, index_t I>
 using uniform_sequence_gen_t = typename uniform_sequence_gen<NSize, I>::type;
 
 } // namespace ck
+
+template <ck::index_t... Is>
+std::ostream& operator<<(std::ostream& os, const ck::Sequence<Is...>)
+{
+    using S = ck::Sequence<Is...>;
+    os << "{";
+    ck::static_for<0, S::Size(), 1>{}([&](auto i) { os << S::At(i).value << ", "; });
+    os << "}";
+    return os;
+}