Add tensor stride

example/13_pool2d_fwd/pool2d_fwd_common.hpp

@@ -117,6 +117,7 @@ bool pool_test(bool do_verification,
         static_cast<OutDataType*>(out_device_buf.GetDeviceBuffer()),
         static_cast<IndexDataType*>(out_indices_device_buf.GetDeviceBuffer()),
         {N, C, Hi, Wi},
+        {C * Hi * Wi, 1, Wi * C, C},
         {Y, X},
         {N, C, Ho, Wo},
         window_strides,

example/48_pool3d_fwd/pool3d_fwd_common.hpp

@@ -125,6 +125,7 @@ bool pool3d_test(bool do_verification,
         static_cast<OutDataType*>(out_device_buf.GetDeviceBuffer()),
         static_cast<IndexDataType*>(out_indices_device_buf.GetDeviceBuffer()),
         {N, C, Di, Hi, Wi},
+        {Di * C * Hi * Wi, 1, C * Hi * Wi, Wi * C, C},
         {Z, Y, X},
         {N, C, Do, Ho, Wo},
         window_strides,

include/ck/tensor_operation/gpu/device/device_pool_fwd.hpp

@@ -21,6 +21,7 @@ struct DevicePoolFwd : public BaseOperator
                         void* p_out_dev,
                         void* p_out_indices_dev,
                         std::array<ck::index_t, InOutRank> input_lengths,
+                        std::array<ck::index_t, InOutRank> input_stride,
                         std::array<ck::index_t, WindowRank> window_lengths,
                         std::array<ck::index_t, InOutRank> output_lengths,
                         std::array<ck::index_t, WindowRank> window_strides,