diff --git a/src/core/CL/CLKernels.h b/src/core/CL/CLKernels.h
index ac05168..7383dce 100644
--- a/src/core/CL/CLKernels.h
+++ b/src/core/CL/CLKernels.h
@@ -42,8 +42,6 @@
 #include "src/core/CL/kernels/CLComparisonKernel.h"
 #include "src/core/CL/kernels/CLConvertFullyConnectedWeightsKernel.h"
 #include "src/core/CL/kernels/CLConvolutionKernel.h"
-#include "src/core/CL/kernels/CLCopyKernel.h"
-#include "src/core/CL/kernels/CLCropKernel.h"
 #include "src/core/CL/kernels/CLDeconvolutionLayerUpsampleKernel.h"
 #include "src/core/CL/kernels/CLDeconvolutionReshapeOutputKernel.h"
 #include "src/core/CL/kernels/CLDepthConvertLayerKernel.h"
@@ -97,7 +95,6 @@
 #include "src/core/CL/kernels/CLMeanStdDevKernel.h"
 #include "src/core/CL/kernels/CLMeanStdDevNormalizationKernel.h"
 #include "src/core/CL/kernels/CLMedian3x3Kernel.h"
-#include "src/core/CL/kernels/CLMemsetKernel.h"
 #include "src/core/CL/kernels/CLMinMaxLayerKernel.h"
 #include "src/core/CL/kernels/CLMinMaxLocationKernel.h"
 #include "src/core/CL/kernels/CLNonLinearFilterKernel.h"
@@ -105,7 +102,6 @@
 #include "src/core/CL/kernels/CLNormalizationLayerKernel.h"
 #include "src/core/CL/kernels/CLNormalizePlanarYUVLayerKernel.h"
 #include "src/core/CL/kernels/CLPadLayerKernel.h"
-#include "src/core/CL/kernels/CLPermuteKernel.h"
 #include "src/core/CL/kernels/CLPixelWiseMultiplicationKernel.h"
 #include "src/core/CL/kernels/CLPriorBoxLayerKernel.h"
 #include "src/core/CL/kernels/CLQLSTMLayerNormalizationKernel.h"
@@ -116,7 +112,6 @@
 #include "src/core/CL/kernels/CLReductionOperationKernel.h"
 #include "src/core/CL/kernels/CLRemapKernel.h"
 #include "src/core/CL/kernels/CLReorgLayerKernel.h"
-#include "src/core/CL/kernels/CLReshapeLayerKernel.h"
 #include "src/core/CL/kernels/CLReverseKernel.h"
 #include "src/core/CL/kernels/CLScaleKernel.h"
 #include "src/core/CL/kernels/CLScharr3x3Kernel.h"
diff --git a/src/core/CL/kernels/CLCopyKernel.cpp b/src/core/CL/kernels/CLCopyKernel.cpp
deleted file mode 100644
index ca38b65..0000000
--- a/src/core/CL/kernels/CLCopyKernel.cpp
+++ /dev/null
@@ -1,169 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#include "src/core/CL/kernels/CLCopyKernel.h"
-
-#include "arm_compute/core/CL/CLHelpers.h"
-#include "arm_compute/core/CL/CLKernelLibrary.h"
-#include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/Utils.h"
-#include "src/core/helpers/AutoConfiguration.h"
-#include "src/core/helpers/WindowHelpers.h"
-#include "support/StringSupport.h"
-
-namespace arm_compute
-{
-namespace
-{
-Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output, Window *output_window = nullptr)
-{
-    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
-
-    // Validate output if initialized
-    if(output->total_size() != 0)
-    {
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(input, output);
-        if(output_window == nullptr)
-        {
-            ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(input->tensor_shape(), output->tensor_shape());
-        }
-        else
-        {
-            ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(input->tensor_shape(), output_window->shape());
-        }
-    }
-
-    return Status{};
-}
-
-} // namespace
-
-CLCopyKernel::CLCopyKernel()
-    : _input(nullptr), _output(nullptr), _output_window(), _has_output_window(false)
-{
-}
-
-void CLCopyKernel::configure(const ICLTensor *input, ICLTensor *output, Window *output_window)
-{
-    configure(CLKernelLibrary::get().get_compile_context(), input, output, output_window);
-}
-
-void CLCopyKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Window *output_window)
-{
-    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), output->info(), output_window));
-
-    auto padding_info = get_padding_info({ input, output });
-
-    _input  = input;
-    _output = output;
-
-    // Create kernel
-    CLBuildOptions build_opts;
-    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type()));
-
-    // Output auto inizialitation if not yet initialized
-    auto_init_if_empty(*(output->info()), *(input->info()));
-
-    // Configure window
-    const unsigned int vec_size_x = adjust_vec_size(16 / input->info()->element_size(), input->info()->dimension(0));
-
-    const Window win_config = calculate_max_window(*(input->info()), Steps(vec_size_x));
-
-    if(output_window != nullptr)
-    {
-        _has_output_window             = true;
-        _output_window                 = Window(*output_window);
-        const int  width_x             = output_window->num_iterations(0);
-        const int  vec_size_x_leftover = width_x % vec_size_x;
-        const bool multi_access_x      = width_x >= static_cast<int32_t>(vec_size_x);
-
-        if(multi_access_x)
-        {
-            _output_window.set(Window::DimX, Window::Dimension(output_window->x().start(), ceil_to_multiple(output_window->x().end(), vec_size_x), vec_size_x));
-        }
-
-        build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
-    }
-    else
-    {
-        const int width_x             = input->info()->tensor_shape().x();
-        const int vec_size_x_leftover = width_x % vec_size_x;
-
-        build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
-    }
-
-    build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
-
-    // Build kernel
-    _kernel = create_kernel(compile_context, "copy_tensor", build_opts.options());
-
-    // Validate and set the window
-    ICLKernel::configure_internal(win_config);
-
-    ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
-}
-
-Status CLCopyKernel::validate(const arm_compute::ITensorInfo *input, const arm_compute::ITensorInfo *output, Window *output_window)
-{
-    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output, output_window));
-
-    return Status{};
-}
-
-void CLCopyKernel::run(const Window &window, cl::CommandQueue &queue)
-{
-    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
-    ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
-
-    Window slice;
-
-    if(_has_output_window)
-    {
-        slice            = window.first_slice_window_3D();
-        Window out_slice = _output_window.first_slice_window_3D();
-        do
-        {
-            unsigned int idx = 0;
-            add_3D_tensor_argument(idx, _input, slice);
-            add_3D_tensor_argument(idx, _output, out_slice);
-            enqueue(queue, *this, slice, lws_hint());
-        }
-        while(window.slide_window_slice_3D(slice) && _output_window.slide_window_slice_3D(out_slice));
-    }
-    else
-    {
-        Window collapsed = window.collapse_if_possible(ICLKernel::window(), Window::DimZ);
-        slice            = collapsed.first_slice_window_3D();
-        do
-        {
-            unsigned int idx = 0;
-            add_3D_tensor_argument(idx, _input, slice);
-            add_3D_tensor_argument(idx, _output, slice);
-            enqueue(queue, *this, slice, lws_hint());
-        }
-        while(collapsed.slide_window_slice_3D(slice));
-    }
-}
-} // namespace arm_compute
diff --git a/src/core/CL/kernels/CLCopyKernel.h b/src/core/CL/kernels/CLCopyKernel.h
deleted file mode 100644
index 9a20b88..0000000
--- a/src/core/CL/kernels/CLCopyKernel.h
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCOPYKERNEL_H
-#define ARM_COMPUTE_CLCOPYKERNEL_H
-
-#include "arm_compute/core/Types.h"
-#include "src/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a copy between two tensors */
-class CLCopyKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCopyKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCopyKernel(const CLCopyKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCopyKernel &operator=(const CLCopyKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCopyKernel(CLCopyKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCopyKernel &operator=(CLCopyKernel &&) = default;
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  input         Source tensor. Data types supported: All.
-     * @param[out] output        Destination tensor. Data types supported: same as @p input.
-     * @param[in]  output_window (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, Window *output_window = nullptr);
-    /** Initialize the kernel's input, output.
-     *
-     * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor. Data types supported: All.
-     * @param[out] output          Destination tensor. Data types supported: same as @p input.
-     * @param[in]  output_window   (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Window *output_window = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLCopyKernel
-     *
-     * @param[in] input         Source tensor info. Data types supported: All.
-     * @param[in] output        Destination tensor info. Data types supported: same as @p input.
-     * @param[in] output_window (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, Window *output_window = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Window           _output_window;
-    bool             _has_output_window;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCOPYKERNEL_H */
diff --git a/src/core/CL/kernels/CLCropKernel.cpp b/src/core/CL/kernels/CLCropKernel.cpp
deleted file mode 100644
index 9cf15ff..0000000
--- a/src/core/CL/kernels/CLCropKernel.cpp
+++ /dev/null
@@ -1,134 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#include "src/core/CL/kernels/CLCropKernel.h"
-
-#include "arm_compute/core/CL/CLHelpers.h"
-#include "arm_compute/core/CL/CLKernelLibrary.h"
-#include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/IAccessWindow.h"
-#include "arm_compute/core/TensorInfo.h"
-#include "src/core/CPP/Validate.h"
-#include "src/core/helpers/WindowHelpers.h"
-
-#include "support/StringSupport.h"
-
-#include <map>
-
-namespace arm_compute
-{
-CLCropKernel::CLCropKernel()
-    : _input(nullptr), _output(nullptr), _start(), _batch_index(0), _extrapolation_value(0)
-{
-}
-
-void CLCropKernel::configure(const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value, Window *output_window)
-{
-    configure(CLKernelLibrary::get().get_compile_context(), input, output, start, end, batch_index, extrapolation_value, output_window);
-}
-
-void CLCropKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index,
-                             float extrapolation_value, Window *output_window)
-{
-    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_ERROR_THROW_ON(validate(input->info(), output->info(), start, end, batch_index, extrapolation_value, output_window));
-
-    _input               = input;
-    _output              = output;
-    _start               = start;
-    _batch_index         = batch_index;
-    _extrapolation_value = extrapolation_value;
-
-    const int vec_size_x = 4;
-    // Create and update the window (if needed)
-    Window win = calculate_max_window(*output->info());
-
-    if(output_window != nullptr)
-    {
-        ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(win, *output_window);
-        win = *output_window;
-    }
-
-    const int  output_width_x = win.num_iterations(0);
-    const bool multi_access_x = output_width_x >= vec_size_x;
-    const bool remainder_x    = output_width_x % vec_size_x > 0;
-
-    if(multi_access_x)
-    {
-        win.set(Window::DimX,
-                Window::Dimension(win.x().start(), ceil_to_multiple(win.x().end(), vec_size_x), vec_size_x));
-    }
-    ICLKernel::configure_internal(win);
-
-    // Create kernel
-    CLBuildOptions build_opts;
-    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type()));
-    build_opts.add_option_if(multi_access_x, "-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
-    build_opts.add_option_if(multi_access_x && remainder_x, "-DLAST_ACCESSED_X=" + support::cpp11::to_string(std::max<int>(output_width_x - vec_size_x, 0)));
-    build_opts.add_option_if(start.x > end.x, "-DWIDTH_FLIPPED=");
-    build_opts.add_option_if(start.y > end.y, "-DHEIGHT_FLIPPED=");
-    _kernel = create_kernel(compile_context, "crop_tensor", build_opts.options());
-}
-
-Status CLCropKernel::validate(const ITensorInfo *input, const ITensorInfo *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value, Window *output_window)
-{
-    ARM_COMPUTE_UNUSED(extrapolation_value, output_window);
-    ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(input);
-    ARM_COMPUTE_RETURN_ERROR_ON(input->data_type() == DataType::UNKNOWN);
-    ARM_COMPUTE_RETURN_ERROR_ON_DATA_LAYOUT_NOT_IN(input, DataLayout::NHWC);
-    ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape().num_dimensions() > 4);
-    ARM_COMPUTE_RETURN_ERROR_ON(start.x < 0 || start.y < 0 || end.x < 0 || end.y < 0);
-    ARM_COMPUTE_RETURN_ERROR_ON(start.x >= static_cast<int32_t>(input->dimension(1)) || start.y >= static_cast<int32_t>(input->dimension(2))
-                                || end.x >= static_cast<int32_t>(input->dimension(1)) || end.y >= static_cast<int32_t>(input->dimension(2)));
-    ARM_COMPUTE_RETURN_ERROR_ON(batch_index >= input->dimension(3));
-    if(output_window != nullptr)
-    {
-        ARM_COMPUTE_RETURN_ERROR_ON(output_window->x().step() != 1);
-    }
-    if(output->total_size() > 0)
-    {
-        ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_NOT_IN(output, DataType::F32);
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(input, output);
-        ARM_COMPUTE_RETURN_ERROR_ON(output->num_dimensions() > 3);
-    }
-    return Status{};
-}
-
-void CLCropKernel::run(const Window &window, cl::CommandQueue &queue)
-{
-    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
-    ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
-
-    Window in_slice = Window();
-    in_slice.use_tensor_dimensions(_input->info()->tensor_shape());
-    in_slice.set(Window::DimX, Window::Dimension(in_slice.x().start(), ceil_to_multiple(in_slice.x().end(), window.x().step()), window.x().step()));
-    in_slice.set(3, Window::Dimension(_batch_index, _batch_index + 1, 1));
-
-    unsigned int idx = 0;
-    add_3D_tensor_argument(idx, _input, in_slice);
-    add_3D_tensor_argument(idx, _output, window);
-    add_argument(idx, _start.x);
-    add_argument(idx, _start.y);
-    enqueue(queue, *this, window, lws_hint());
-}
-} // namespace arm_compute
diff --git a/src/core/CL/kernels/CLCropKernel.h b/src/core/CL/kernels/CLCropKernel.h
deleted file mode 100644
index cbfada5..0000000
--- a/src/core/CL/kernels/CLCropKernel.h
+++ /dev/null
@@ -1,103 +0,0 @@
-/*
- * Copyright (c) 2019-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLCROPKERNEL_H
-#define ARM_COMPUTE_CLCROPKERNEL_H
-
-#include "arm_compute/core/Types.h"
-#include "src/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform a copy between two tensors */
-class CLCropKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLCropKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCropKernel(const CLCropKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers). */
-    CLCropKernel &operator=(const CLCropKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLCropKernel(CLCropKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLCropKernel &operator=(CLCropKernel &&) = default;
-    /** Configure kernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in]  input               Source tensor. Data type supported: All. Data layouts supported: NHWC.
-     * @param[out] output              Destination tensor. Data type supported: F32
-     * @param[in]  start               Coordinates of where to start cropping the image.
-     * @param[in]  end                 Coordinates of where to end cropping the image.
-     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in]  output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0, Window *output_window = nullptr);
-    /** Configure kernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in]  compile_context     The compile context to be used.
-     * @param[in]  input               Source tensor. Data type supported: All. Data layouts supported: NHWC.
-     * @param[out] output              Destination tensor. Data type supported: F32
-     * @param[in]  start               Coordinates of where to start cropping the image.
-     * @param[in]  end                 Coordinates of where to end cropping the image.
-     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in]  output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
-                   Window *output_window = nullptr);
-
-    /** Static function to check if given info will lead to a valid configuration of @ref CLStridedSliceKernel
-     *
-     * @note Supported tensor rank: up to 4
-     *
-     * @param[in] input               Source tensor info. Data type supported: All. Data layouts supported: NHWC.
-     * @param[in] output              Destination tensor info. Data type supported: F32
-     * @param[in] start               Coordinates of where to start cropping the image.
-     * @param[in] end                 Coordinates of where to end cropping the image.
-     * @param[in] batch_index         Fourth dimension index of the 3D image to crop in @p input.
-     * @param[in] extrapolation_value Value to be used for values outside of the image. Default is 0.
-     * @param[in] output_window       Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
-                           Window *output_window = nullptr);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor       *_output;
-    Coordinates2D    _start;
-    uint32_t         _batch_index;
-    float            _extrapolation_value;
-};
-} // namespace arm_compute
-#endif /*ARM_COMPUTE_CLCROPKERNEL_H */
diff --git a/src/core/CL/kernels/CLPermuteKernel.cpp b/src/core/CL/kernels/CLPermuteKernel.cpp
deleted file mode 100644
index 07d83bd..0000000
--- a/src/core/CL/kernels/CLPermuteKernel.cpp
+++ /dev/null
@@ -1,145 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#include "src/core/CL/kernels/CLPermuteKernel.h"
-#include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/utils/misc/ShapeCalculator.h"
-#include "src/core/helpers/AutoConfiguration.h"
-#include "src/core/helpers/WindowHelpers.h"
-#include "support/StringSupport.h"
-
-namespace arm_compute
-{
-CLPermuteKernel::CLPermuteKernel()
-    : _input(nullptr), _output(nullptr), _perm()
-{
-}
-namespace
-{
-TensorShape get_output_shape(const ITensorInfo *input, const PermutationVector &perm)
-{
-    TensorShape output_shape = input->tensor_shape();
-    permute(output_shape, perm);
-    return output_shape;
-}
-
-Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output, const PermutationVector &perm)
-{
-    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_RETURN_ERROR_ON(input->data_type() == DataType::UNKNOWN);
-    ARM_COMPUTE_RETURN_ERROR_ON_MSG(input->num_dimensions() < 1 || input->num_dimensions() > 4,
-                                    "Permutation upto 4-D input tensor is supported");
-    ARM_COMPUTE_RETURN_ERROR_ON_MSG(perm.num_dimensions() < 1 || perm.num_dimensions() > 4,
-                                    "Permutation vector size should be less than or equal to 4");
-    for(const auto &p : perm)
-    {
-        ARM_COMPUTE_RETURN_ERROR_ON_MSG(p >= perm.num_dimensions(), "Permutation vector has invalid values");
-    }
-
-    // Validate configured output
-    if(output->total_size() != 0)
-    {
-        const TensorShape output_shape = misc::shape_calculator::compute_permutation_output_shape(*input, perm);
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(output->tensor_shape(), output_shape);
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(input, output);
-        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
-    }
-    return Status{};
-}
-} // namespace
-
-void CLPermuteKernel::configure(const ICLTensor *input, ICLTensor *output, const PermutationVector &perm)
-{
-    configure(CLKernelLibrary::get().get_compile_context(), input, output, perm);
-}
-
-void CLPermuteKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PermutationVector &perm)
-{
-    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
-    auto              padding_info = get_padding_info({ input, output });
-    const TensorShape output_shape = get_output_shape(input->info(), perm);
-    // Output auto inizialitation if not yet initialized
-    auto_init_if_empty(*output->info(), input->info()->clone()->set_tensor_shape(output_shape));
-
-    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), output->info(), perm));
-
-    _input  = input;
-    _output = output;
-    _perm   = perm;
-
-    // Create kernel
-    CLBuildOptions build_opts;
-    build_opts.add_option("-DDATA_TYPE=" + get_cl_unsigned_type_from_element_size(data_size_from_type(input->info()->data_type())));
-    build_opts.add_option("-DDEPTH_IN=" + support::cpp11::to_string(input->info()->dimension(2)));
-    // New positions of  width(W), height(H), channel(C) and batch(D) based on permutation vector
-    build_opts.add_option("-DP1=" + support::cpp11::to_string((_perm.num_dimensions() >= 1) ? perm[0] : 0));
-    build_opts.add_option("-DP2=" + support::cpp11::to_string((_perm.num_dimensions() >= 2) ? perm[1] : 1));
-    build_opts.add_option("-DP3=" + support::cpp11::to_string((_perm.num_dimensions() >= 3) ? perm[2] : 2));
-    build_opts.add_option("-DP4=" + support::cpp11::to_string((_perm.num_dimensions() >= 4) ? perm[3] : 3));
-
-    _kernel = create_kernel(compile_context, "permute", build_opts.options());
-
-    // Configure  kernel window
-    Window win = calculate_max_window(*input->info(), Steps());
-
-    // The CLPermute doesn't need padding so update_window_and_padding() can be skipped
-    Coordinates coord;
-    coord.set_num_dimensions(output->info()->num_dimensions());
-    output->info()->set_valid_region(ValidRegion(coord, output->info()->tensor_shape()));
-
-    ICLKernel::configure_internal(win);
-    ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
-}
-
-Status CLPermuteKernel::validate(const ITensorInfo *input, const ITensorInfo *output, const PermutationVector &perm)
-{
-    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output, perm));
-
-    return Status{};
-}
-
-void CLPermuteKernel::run(const Window &window, cl::CommandQueue &queue)
-{
-    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
-    ARM_COMPUTE_ERROR_ON_MISMATCHING_WINDOWS(ICLKernel::window(), window);
-
-    Window slice_in = window.first_slice_window_4D().collapse(ICLKernel::window(), 2, 4);
-
-    // Setup output slice
-    Window slice_out(slice_in);
-    slice_out.set(Window::DimX, Window::Dimension(0, 0, 0));
-    slice_out.set(Window::DimY, Window::Dimension(0, 0, 0));
-    slice_out.set(Window::DimZ, Window::Dimension(0, 0, 0));
-    slice_out.set(3, Window::Dimension(0, 0, 0));
-
-    do
-    {
-        unsigned int idx = 0;
-        add_4D_tensor_argument(idx, _input, slice_in);
-        add_4D_tensor_argument(idx, _output, slice_out);
-        enqueue(queue, *this, slice_in, lws_hint());
-    }
-    while(window.slide_window_slice_4D(slice_in) && window.slide_window_slice_4D(slice_out));
-}
-} // namespace arm_compute
\ No newline at end of file
diff --git a/src/core/CL/kernels/CLPermuteKernel.h b/src/core/CL/kernels/CLPermuteKernel.h
deleted file mode 100644
index d1bb875..0000000
--- a/src/core/CL/kernels/CLPermuteKernel.h
+++ /dev/null
@@ -1,90 +0,0 @@
-/*
- * Copyright (c) 2018-2020 Arm Limited.
- *
- * SPDX-License-Identifier: MIT
- *
- * Permission is hereby granted, free of charge, to any person obtaining a copy
- * of this software and associated documentation files (the "Software"), to
- * deal in the Software without restriction, including without limitation the
- * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
- * sell copies of the Software, and to permit persons to whom the Software is
- * furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in all
- * copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
- * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-#ifndef ARM_COMPUTE_CLPERMUTEKERNEL_H
-#define ARM_COMPUTE_CLPERMUTEKERNEL_H
-
-#include "src/core/CL/ICLKernel.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** OpenCL kernel to perform tensor permutation.
- *
- * Permutes given a permutation vector
- */
-class CLPermuteKernel : public ICLKernel
-{
-public:
-    /** Default constructor */
-    CLPermuteKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPermuteKernel(const CLPermuteKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLPermuteKernel &operator=(const CLPermuteKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLPermuteKernel(CLPermuteKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLPermuteKernel &operator=(CLPermuteKernel &&) = default;
-    /** Set the input and output of the kernel.
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] input  The input tensor to permute. Data types supported: All.
-     * @param[in] output The output tensor. Data types supported: Same as @p input
-     * @param[in] perm   Permutation vector
-     */
-    void configure(const ICLTensor *input, ICLTensor *output, const PermutationVector &perm);
-    /** Set the input and output of the kernel.
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] compile_context The compile context to be used.
-     * @param[in] input           The input tensor to permute. Data types supported: All.
-     * @param[in] output          The output tensor. Data types supported: Same as @p input
-     * @param[in] perm            Permutation vector
-     */
-    void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, const PermutationVector &perm);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLPermuteKernel
-     *
-     * @note Arbitrary permutation vectors are supported with rank not greater than 4
-     *
-     * @param[in] input  First tensor input info. Data types supported: All.
-     * @param[in] output Output tensor info. Data types supported: same as @p input.
-     * @param[in] perm   Permutation vector
-     *
-     * @return a status
-     */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output, const PermutationVector &perm);
-
-    // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
-
-private:
-    const ICLTensor *_input;
-    ICLTensor        *_output;
-    PermutationVector _perm;
-};
-} // arm_compute
-#endif /*ARM_COMPUTE_CLPERMUTEKERNEL_H */
diff --git a/src/core/gpu/cl/kernels/ClCopyKernel.cpp b/src/core/gpu/cl/kernels/ClCopyKernel.cpp
new file mode 100644
index 0000000..d6c87f8
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClCopyKernel.cpp
@@ -0,0 +1,170 @@
+/*
+ * Copyright (c) 2018-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "src/core/gpu/cl/kernels/ClCopyKernel.h"
+
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLKernelLibrary.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/Helpers.h"
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Utils.h"
+#include "arm_compute/core/Validate.h"
+#include "src/core/CL/CLValidate.h"
+#include "src/core/helpers/AutoConfiguration.h"
+#include "src/core/helpers/WindowHelpers.h"
+#include "support/Cast.h"
+#include "support/StringSupport.h"
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+namespace
+{
+Status validate_arguments(const ITensorInfo *src, const ITensorInfo *dst, Window *dst_window = nullptr)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, dst);
+
+    // Validate dst if initialized
+    if(dst->total_size() != 0)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, dst);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(src, dst);
+        if(dst_window == nullptr)
+        {
+            ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(src->tensor_shape(), dst->tensor_shape());
+        }
+        else
+        {
+            ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(src->tensor_shape(), dst_window->shape());
+        }
+    }
+
+    return Status{};
+}
+
+} // namespace
+
+void ClCopyKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Window *dst_window)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, dst, dst_window));
+
+    auto padding_info = get_padding_info({ src, dst });
+
+    // Create kernel
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(src->data_type()));
+
+    // Output auto inizialitation if not yet initialized
+    auto_init_if_empty(*dst, *src);
+
+    // Configure window
+    const unsigned int vec_size_x = adjust_vec_size(16 / src->element_size(), src->dimension(0));
+
+    const Window win_config = calculate_max_window(*src, Steps(vec_size_x));
+
+    if(dst_window != nullptr)
+    {
+        _has_dst_window                = true;
+        _dst_window                    = Window(*dst_window);
+        const int  width_x             = dst_window->num_iterations(0);
+        const int  vec_size_x_leftover = width_x % vec_size_x;
+        const bool multi_access_x      = width_x >= static_cast<int32_t>(vec_size_x);
+
+        if(multi_access_x)
+        {
+            _dst_window.set(Window::DimX, Window::Dimension(dst_window->x().start(), ceil_to_multiple(dst_window->x().end(), vec_size_x), vec_size_x));
+        }
+
+        build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
+    }
+    else
+    {
+        const int width_x             = src->tensor_shape().x();
+        const int vec_size_x_leftover = width_x % vec_size_x;
+
+        build_opts.add_option("-DVEC_SIZE_LEFTOVER=" + support::cpp11::to_string(vec_size_x_leftover));
+    }
+
+    build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
+
+    // Build kernel
+    _kernel = create_kernel(compile_context, "copy_tensor", build_opts.options());
+
+    // Validate and set the window
+    ICLKernel::configure_internal(win_config);
+
+    ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
+}
+
+Status ClCopyKernel::validate(const arm_compute::ITensorInfo *src, const arm_compute::ITensorInfo *dst, Window *dst_window)
+{
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, dst, dst_window));
+
+    return Status{};
+}
+
+void ClCopyKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
+{
+    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
+    ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
+
+    const auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
+    auto       dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
+
+    Window slice;
+
+    if(_has_dst_window)
+    {
+        slice            = window.first_slice_window_3D();
+        Window out_slice = _dst_window.first_slice_window_3D();
+        do
+        {
+            unsigned int idx = 0;
+            add_3D_tensor_argument(idx, src, slice);
+            add_3D_tensor_argument(idx, dst, out_slice);
+            enqueue(queue, *this, slice, lws_hint());
+        }
+        while(window.slide_window_slice_3D(slice) && _dst_window.slide_window_slice_3D(out_slice));
+    }
+    else
+    {
+        Window collapsed = window.collapse_if_possible(ICLKernel::window(), Window::DimZ);
+        slice            = collapsed.first_slice_window_3D();
+        do
+        {
+            unsigned int idx = 0;
+            add_3D_tensor_argument(idx, src, slice);
+            add_3D_tensor_argument(idx, dst, slice);
+            enqueue(queue, *this, slice, lws_hint());
+        }
+        while(collapsed.slide_window_slice_3D(slice));
+    }
+}
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
diff --git a/src/core/gpu/cl/kernels/ClCopyKernel.h b/src/core/gpu/cl/kernels/ClCopyKernel.h
new file mode 100644
index 0000000..9d5457d
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClCopyKernel.h
@@ -0,0 +1,71 @@
+/*
+ * Copyright (c) 2018-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef ARM_COMPUTE_CL_COPY_KERNEL_H
+#define ARM_COMPUTE_CL_COPY_KERNEL_H
+
+#include "src/core/common/Macros.h"
+#include "src/core/gpu/cl/ClCompileContext.h"
+#include "src/core/gpu/cl/IClKernel.h"
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+/** OpenCL kernel to perform a copy between two tensors */
+class ClCopyKernel : public ICLKernel
+{
+public:
+    ClCopyKernel() = default;
+    ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClCopyKernel);
+    /** Initialize the kernel's src, dst.
+     *
+     * @param[in]  compile_context The compile context to be used.
+     * @param[in]  src             Source tensor info. Data types supported: All.
+     * @param[out] dst             Destination tensor info. Data types supported: same as @p src.
+     * @param[in]  dst_window      (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
+     */
+    void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Window *dst_window = nullptr);
+    /** Static function to check if given info will lead to a valid configuration of @ref ClCopyKernel
+     *
+     * @param[in] src        Source tensor info. Data types supported: All.
+     * @param[in] dst        Destination tensor info. Data types supported: same as @p src.
+     * @param[in] dst_window (Optional) Window to be used in case only copying into part of a tensor. Default is nullptr.
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *src, const ITensorInfo *dst, Window *dst_window = nullptr);
+
+    // Inherited methods overridden:
+    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    Window _dst_window{};
+    bool   _has_dst_window{};
+};
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
+#endif /*ARM_COMPUTE_CL_COPY_KERNEL_H */
diff --git a/src/core/gpu/cl/kernels/ClCropKernel.cpp b/src/core/gpu/cl/kernels/ClCropKernel.cpp
new file mode 100644
index 0000000..20e94b2
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClCropKernel.cpp
@@ -0,0 +1,137 @@
+/*
+ * Copyright (c) 2019-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "src/core/gpu/cl/kernels/ClCropKernel.h"
+
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLKernelLibrary.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/IAccessWindow.h"
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Utils.h"
+#include "src/core/CPP/Validate.h"
+#include "src/core/helpers/WindowHelpers.h"
+#include "support/Cast.h"
+#include "support/StringSupport.h"
+
+#include <map>
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+void ClCropKernel::configure(const ITensorInfo *src, ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value, Window *dst_window)
+{
+    configure(CLKernelLibrary::get().get_compile_context(), src, dst, start, end, batch_index, extrapolation_value, dst_window);
+}
+
+void ClCropKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index,
+                             float extrapolation_value, Window *dst_window)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_ERROR_THROW_ON(validate(src, dst, start, end, batch_index, extrapolation_value, dst_window));
+
+    _start               = start;
+    _batch_index         = batch_index;
+    _extrapolation_value = extrapolation_value;
+
+    const int vec_size_x = 4;
+    // Create and update the window (if needed)
+    Window win = calculate_max_window(*dst);
+
+    if(dst_window != nullptr)
+    {
+        ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(win, *dst_window);
+        win = *dst_window;
+    }
+
+    const int  dst_width_x    = win.num_iterations(0);
+    const bool multi_access_x = dst_width_x >= vec_size_x;
+    const bool remainder_x    = dst_width_x % vec_size_x > 0;
+
+    if(multi_access_x)
+    {
+        win.set(Window::DimX,
+                Window::Dimension(win.x().start(), ceil_to_multiple(win.x().end(), vec_size_x), vec_size_x));
+    }
+    ICLKernel::configure_internal(win);
+
+    // Create kernel
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(src->data_type()));
+    build_opts.add_option_if(multi_access_x, "-DVEC_SIZE=" + support::cpp11::to_string(vec_size_x));
+    build_opts.add_option_if(multi_access_x && remainder_x, "-DLAST_ACCESSED_X=" + support::cpp11::to_string(std::max<int>(dst_width_x - vec_size_x, 0)));
+    build_opts.add_option_if(start.x > end.x, "-DWIDTH_FLIPPED=");
+    build_opts.add_option_if(start.y > end.y, "-DHEIGHT_FLIPPED=");
+    _kernel = create_kernel(compile_context, "crop_tensor", build_opts.options());
+}
+
+Status ClCropKernel::validate(const ITensorInfo *src, const ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value, Window *dst_window)
+{
+    ARM_COMPUTE_UNUSED(extrapolation_value, dst_window);
+    ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(src);
+    ARM_COMPUTE_RETURN_ERROR_ON(src->data_type() == DataType::UNKNOWN);
+    ARM_COMPUTE_RETURN_ERROR_ON_DATA_LAYOUT_NOT_IN(src, DataLayout::NHWC);
+    ARM_COMPUTE_RETURN_ERROR_ON(src->tensor_shape().num_dimensions() > 4);
+    ARM_COMPUTE_RETURN_ERROR_ON(start.x < 0 || start.y < 0 || end.x < 0 || end.y < 0);
+    ARM_COMPUTE_RETURN_ERROR_ON(start.x >= static_cast<int32_t>(src->dimension(1)) || start.y >= static_cast<int32_t>(src->dimension(2))
+                                || end.x >= static_cast<int32_t>(src->dimension(1)) || end.y >= static_cast<int32_t>(src->dimension(2)));
+    ARM_COMPUTE_RETURN_ERROR_ON(batch_index >= src->dimension(3));
+    if(dst_window != nullptr)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON(dst_window->x().step() != 1);
+    }
+    if(dst->total_size() > 0)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_NOT_IN(dst, DataType::F32);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(src, dst);
+        ARM_COMPUTE_RETURN_ERROR_ON(dst->num_dimensions() > 3);
+    }
+    return Status{};
+}
+
+void ClCropKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
+{
+    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
+    ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
+
+    const auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
+    auto       dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
+
+    Window in_slice = Window();
+    in_slice.use_tensor_dimensions(src->info()->tensor_shape());
+    in_slice.set(Window::DimX, Window::Dimension(in_slice.x().start(), ceil_to_multiple(in_slice.x().end(), window.x().step()), window.x().step()));
+    in_slice.set(3, Window::Dimension(_batch_index, _batch_index + 1, 1));
+
+    unsigned int idx = 0;
+    add_3D_tensor_argument(idx, src, in_slice);
+    add_3D_tensor_argument(idx, dst, window);
+    add_argument(idx, _start.x);
+    add_argument(idx, _start.y);
+    enqueue(queue, *this, window, lws_hint());
+}
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
diff --git a/src/core/gpu/cl/kernels/ClCropKernel.h b/src/core/gpu/cl/kernels/ClCropKernel.h
new file mode 100644
index 0000000..92f94d4
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClCropKernel.h
@@ -0,0 +1,98 @@
+/*
+ * Copyright (c) 2019-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef ARM_COMPUTE_CLCROPKERNEL_H
+#define ARM_COMPUTE_CLCROPKERNEL_H
+
+#include "src/core/common/Macros.h"
+#include "src/core/gpu/cl/ClCompileContext.h"
+#include "src/core/gpu/cl/IClKernel.h"
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+/** OpenCL kernel to perform a copy between two tensors */
+class ClCropKernel : public ICLKernel
+{
+public:
+    ClCropKernel() = default;
+    ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClCropKernel);
+    /** Configure kernel
+     *
+     * @note Supported tensor rank: up to 4
+     *
+     * @param[in]  src                 Source tensor info. Data type supported: All. Data layouts supported: NHWC.
+     * @param[out] dst                 Destination tensor info. Data type supported: F32
+     * @param[in]  start               Coordinates of where to start cropping the image.
+     * @param[in]  end                 Coordinates of where to end cropping the image.
+     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p src.
+     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
+     * @param[in]  dst_window          Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
+     */
+    void configure(const ITensorInfo *src, ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0, Window *dst_window = nullptr);
+    /** Configure kernel
+     *
+     * @note Supported tensor rank: up to 4
+     *
+     * @param[in]  compile_context     The compile context to be used.
+     * @param[in]  src                 Source tensor info. Data type supported: All. Data layouts supported: NHWC.
+     * @param[out] dst                 Destination tensor info. Data type supported: F32
+     * @param[in]  start               Coordinates of where to start cropping the image.
+     * @param[in]  end                 Coordinates of where to end cropping the image.
+     * @param[in]  batch_index         Fourth dimension index of the 3D image to crop in @p src.
+     * @param[in]  extrapolation_value Value to be used for values outside of the image. Default is 0.
+     * @param[in]  dst_window          Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
+     */
+    void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
+                   Window *dst_window = nullptr);
+
+    /** Static function to check if given info will lead to a valid configuration of @ref CLStridedSliceKernel
+     *
+     * @note Supported tensor rank: up to 4
+     *
+     * @param[in] src                 Source tensor info. Data type supported: All. Data layouts supported: NHWC.
+     * @param[in] dst                 Destination tensor info. Data type supported: F32
+     * @param[in] start               Coordinates of where to start cropping the image.
+     * @param[in] end                 Coordinates of where to end cropping the image.
+     * @param[in] batch_index         Fourth dimension index of the 3D image to crop in @p src.
+     * @param[in] extrapolation_value Value to be used for values outside of the image. Default is 0.
+     * @param[in] dst_window          Output window to be used in case cropped image is being copied into a tensor. Default is nullptr.
+     */
+    static Status validate(const ITensorInfo *src, const ITensorInfo *dst, Coordinates2D start, Coordinates2D end, uint32_t batch_index, float extrapolation_value = 0,
+                           Window *dst_window = nullptr);
+
+    // Inherited methods overridden:
+    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    Coordinates2D _start{};
+    uint32_t      _batch_index{};
+    float         _extrapolation_value{};
+};
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
+#endif /*ARM_COMPUTE_CLCROPKERNEL_H */
diff --git a/src/core/CL/kernels/CLMemsetKernel.cpp b/src/core/gpu/cl/kernels/ClFillKernel.cpp
similarity index 67%
rename from src/core/CL/kernels/CLMemsetKernel.cpp
rename to src/core/gpu/cl/kernels/ClFillKernel.cpp
index 2543b07..b194ee5 100644
--- a/src/core/CL/kernels/CLMemsetKernel.cpp
+++ b/src/core/gpu/cl/kernels/ClFillKernel.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018-2020 Arm Limited.
+ * Copyright (c) 2018-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -21,40 +21,46 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-#include "src/core/CL/kernels/CLMemsetKernel.h"
+#include "src/core/gpu/cl/kernels/ClFillKernel.h"
+
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLKernelLibrary.h"
 #include "arm_compute/core/CL/ICLTensor.h"
-#include "arm_compute/core/utils/misc/ShapeCalculator.h"
+#include "arm_compute/core/Helpers.h"
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Utils.h"
+#include "arm_compute/core/Validate.h"
+#include "src/core/CL/CLValidate.h"
+#include "src/core/helpers/AutoConfiguration.h"
 #include "src/core/helpers/WindowHelpers.h"
+#include "support/Cast.h"
 #include "support/StringSupport.h"
 
 namespace arm_compute
 {
-CLMemsetKernel::CLMemsetKernel()
-    : ICLKernel(), _tensor(nullptr), _full_window()
+namespace opencl
 {
-}
-
-void CLMemsetKernel::configure(ICLTensor        *tensor,
-                               const PixelValue &constant_value,
-                               Window           *window)
+namespace kernels
+{
+void ClFillKernel::configure(ITensorInfo      *tensor,
+                             const PixelValue &constant_value,
+                             Window           *window)
 {
     configure(CLKernelLibrary::get().get_compile_context(), tensor, constant_value, window);
 }
 
-void CLMemsetKernel::configure(const CLCompileContext &compile_context, ICLTensor *tensor,
-                               const PixelValue &constant_value,
-                               Window           *window)
+void ClFillKernel::configure(const CLCompileContext &compile_context, ITensorInfo *tensor,
+                             const PixelValue &constant_value,
+                             Window           *window)
 {
     ARM_COMPUTE_ERROR_ON_NULLPTR(tensor);
-    ARM_COMPUTE_ERROR_THROW_ON(validate(tensor->info(), constant_value, window));
+    ARM_COMPUTE_ERROR_THROW_ON(validate(tensor, constant_value, window));
 
-    _tensor = tensor;
-
-    const DataType data_type  = tensor->info()->data_type();
-    const int      vec_size_x = 16 / tensor->info()->element_size();
+    const DataType data_type  = tensor->data_type();
+    const int      vec_size_x = 16 / tensor->element_size();
 
     // Create and update the window (if needed)
-    _full_window = calculate_max_window(*tensor->info());
+    _full_window = calculate_max_window(*tensor);
     Window win   = _full_window;
     if(window != nullptr)
     {
@@ -81,7 +87,7 @@
     _kernel = create_kernel(compile_context, "memset", build_opts.options());
 }
 
-Status CLMemsetKernel::validate(const ITensorInfo *tensor, const PixelValue &constant_value, Window *window)
+Status ClFillKernel::validate(const ITensorInfo *tensor, const PixelValue &constant_value, Window *window)
 {
     ARM_COMPUTE_UNUSED(tensor);
     ARM_COMPUTE_UNUSED(constant_value);
@@ -92,11 +98,13 @@
     return Status{};
 }
 
-void CLMemsetKernel::run(const Window &window, cl::CommandQueue &queue)
+void ClFillKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
 {
     ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
     ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
 
+    const auto tensor = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
+
     // Collapse all the batches on the third
     Window collapsed = window.collapse_if_possible(_full_window, Window::DimZ);
     Window slice     = collapsed.first_slice_window_3D();
@@ -104,9 +112,11 @@
     do
     {
         unsigned int idx = 0;
-        add_3D_tensor_argument(idx, _tensor, slice);
+        add_3D_tensor_argument(idx, tensor, slice);
         enqueue(queue, *this, slice, lws_hint());
     }
     while(collapsed.slide_window_slice_3D(slice));
 }
+} // namespace kernels
+} // namespace opencl
 } // namespace arm_compute
diff --git a/src/core/CL/kernels/CLMemsetKernel.h b/src/core/gpu/cl/kernels/ClFillKernel.h
similarity index 61%
rename from src/core/CL/kernels/CLMemsetKernel.h
rename to src/core/gpu/cl/kernels/ClFillKernel.h
index dc103f5..136c124 100644
--- a/src/core/CL/kernels/CLMemsetKernel.h
+++ b/src/core/gpu/cl/kernels/ClFillKernel.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018-2020 Arm Limited.
+ * Copyright (c) 2018-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -21,50 +21,41 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-#ifndef ARM_COMPUTE_CLMEMSETKERNEL_H
-#define ARM_COMPUTE_CLMEMSETKERNEL_H
+#ifndef ARM_COMPUTE_CL_FILL_KERNEL_H
+#define ARM_COMPUTE_CL_FILL_KERNEL_H
 
-#include "arm_compute/core/PixelValue.h"
-#include "arm_compute/core/Types.h"
-#include "src/core/CL/ICLKernel.h"
+#include "src/core/common/Macros.h"
+#include "src/core/gpu/cl/ClCompileContext.h"
+#include "src/core/gpu/cl/IClKernel.h"
 
 namespace arm_compute
 {
-class ICLTensor;
-
+namespace opencl
+{
+namespace kernels
+{
 /** Interface for filling the planes of a tensor */
-class CLMemsetKernel : public ICLKernel
+class ClFillKernel : public ICLKernel
 {
 public:
-    /** Default constructor */
-    CLMemsetKernel();
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMemsetKernel(const CLMemsetKernel &) = delete;
-    /** Prevent instances of this class from being copied (As this class contains pointers) */
-    CLMemsetKernel &operator=(const CLMemsetKernel &) = delete;
-    /** Allow instances of this class to be moved */
-    CLMemsetKernel(CLMemsetKernel &&) = default;
-    /** Allow instances of this class to be moved */
-    CLMemsetKernel &operator=(CLMemsetKernel &&) = default;
-    /** Default destructor */
-    ~CLMemsetKernel() = default;
-
+    ClFillKernel() = default;
+    ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClFillKernel);
     /** Initialise the kernel's tensor and filling value
      *
-     * @param[in,out] tensor         Input tensor to fill. Supported data types: All.
+     * @param[in,out] tensor         Input tensor info. Supported data types: All.
      * @param[in]     constant_value The value used to fill the planes of the tensor
      * @param[in]     window         Window to be used in case setting only part of a tensor. Default is nullptr.
      */
-    void configure(ICLTensor *tensor, const PixelValue &constant_value, Window *window = nullptr);
+    void configure(ITensorInfo *tensor, const PixelValue &constant_value, Window *window = nullptr);
     /** Initialise the kernel's tensor and filling value
      *
      * @param[in]     compile_context The compile context to be used.
-     * @param[in,out] tensor          Input tensor to fill. Supported data types: All.
+     * @param[in,out] tensor          Input tensor info. Supported data types: All.
      * @param[in]     constant_value  The value used to fill the planes of the tensor
      * @param[in]     window          Window to be used in case setting only part of a tensor. Default is nullptr.
      */
-    void configure(const CLCompileContext &compile_context, ICLTensor *tensor, const PixelValue &constant_value, Window *window = nullptr);
-    /** Static function to check if given info will lead to a valid configuration of @ref CLMemsetKernel
+    void configure(const CLCompileContext &compile_context, ITensorInfo *tensor, const PixelValue &constant_value, Window *window = nullptr);
+    /** Static function to check if given info will lead to a valid configuration of @ref ClFillKernel
      *
      * @param[in] tensor         Source tensor info. Data types supported: All.
      * @param[in] constant_value The value used to fill the planes of the tensor
@@ -75,11 +66,12 @@
     static Status validate(const ITensorInfo *tensor, const PixelValue &constant_value, Window *window = nullptr);
 
     // Inherited methods overridden:
-    void run(const Window &window, cl::CommandQueue &queue) override;
+    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
 
 private:
-    ICLTensor *_tensor;
-    Window     _full_window;
+    Window _full_window{};
 };
+} // namespace kernels
+} // namespace opencl
 } // namespace arm_compute
 #endif /*ARM_COMPUTE_CLMEMSETRKERNEL_H */
diff --git a/src/core/gpu/cl/kernels/ClPermuteKernel.cpp b/src/core/gpu/cl/kernels/ClPermuteKernel.cpp
new file mode 100644
index 0000000..992c2a8
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClPermuteKernel.cpp
@@ -0,0 +1,157 @@
+/*
+ * Copyright (c) 2018-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "src/core/gpu/cl/kernels/ClPermuteKernel.h"
+
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLKernelLibrary.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/Helpers.h"
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Utils.h"
+#include "arm_compute/core/Validate.h"
+#include "arm_compute/core/utils/misc/ShapeCalculator.h"
+#include "src/core/CL/CLValidate.h"
+#include "src/core/helpers/AutoConfiguration.h"
+#include "src/core/helpers/WindowHelpers.h"
+#include "support/Cast.h"
+#include "support/StringSupport.h"
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+namespace
+{
+TensorShape get_dst_shape(const ITensorInfo *src, const PermutationVector &perm)
+{
+    TensorShape dst_shape = src->tensor_shape();
+    permute(dst_shape, perm);
+    return dst_shape;
+}
+
+Status validate_arguments(const ITensorInfo *src, const ITensorInfo *dst, const PermutationVector &perm)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_RETURN_ERROR_ON(src->data_type() == DataType::UNKNOWN);
+    ARM_COMPUTE_RETURN_ERROR_ON_MSG(src->num_dimensions() < 1 || src->num_dimensions() > 4,
+                                    "Permutation upto 4-D src tensor is supported");
+    ARM_COMPUTE_RETURN_ERROR_ON_MSG(perm.num_dimensions() < 1 || perm.num_dimensions() > 4,
+                                    "Permutation vector size should be less than or equal to 4");
+    for(const auto &p : perm)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON_MSG(p >= perm.num_dimensions(), "Permutation vector has invalid values");
+    }
+
+    // Validate configured dst
+    if(dst->total_size() != 0)
+    {
+        const TensorShape dst_shape = misc::shape_calculator::compute_permutation_output_shape(*src, perm);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(dst->tensor_shape(), dst_shape);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(src, dst);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, dst);
+    }
+    return Status{};
+}
+} // namespace
+
+void ClPermuteKernel::configure(const ITensorInfo *src, ITensorInfo *dst, const PermutationVector &perm)
+{
+    configure(CLKernelLibrary::get().get_compile_context(), src, dst, perm);
+}
+
+void ClPermuteKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, const PermutationVector &perm)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst);
+    auto              padding_info = get_padding_info({ src, dst });
+    const TensorShape dst_shape    = get_dst_shape(src, perm);
+    // Output auto inizialitation if not yet initialized
+    auto_init_if_empty(*dst, src->clone()->set_tensor_shape(dst_shape));
+
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, dst, perm));
+
+    _perm = perm;
+
+    // Create kernel
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_unsigned_type_from_element_size(data_size_from_type(src->data_type())));
+    build_opts.add_option("-DDEPTH_IN=" + support::cpp11::to_string(src->dimension(2)));
+    // New positions of  width(W), height(H), channel(C) and batch(D) based on permutation vector
+    build_opts.add_option("-DP1=" + support::cpp11::to_string((_perm.num_dimensions() >= 1) ? perm[0] : 0));
+    build_opts.add_option("-DP2=" + support::cpp11::to_string((_perm.num_dimensions() >= 2) ? perm[1] : 1));
+    build_opts.add_option("-DP3=" + support::cpp11::to_string((_perm.num_dimensions() >= 3) ? perm[2] : 2));
+    build_opts.add_option("-DP4=" + support::cpp11::to_string((_perm.num_dimensions() >= 4) ? perm[3] : 3));
+
+    _kernel = create_kernel(compile_context, "permute", build_opts.options());
+
+    // Configure  kernel window
+    Window win = calculate_max_window(*src, Steps());
+
+    // The CLPermute doesn't need padding so update_window_and_padding() can be skipped
+    Coordinates coord;
+    coord.set_num_dimensions(dst->num_dimensions());
+    dst->set_valid_region(ValidRegion(coord, dst->tensor_shape()));
+
+    ICLKernel::configure_internal(win);
+    ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
+}
+
+Status ClPermuteKernel::validate(const ITensorInfo *src, const ITensorInfo *dst, const PermutationVector &perm)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, dst, perm));
+
+    return Status{};
+}
+
+void ClPermuteKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
+{
+    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
+    ARM_COMPUTE_ERROR_ON_MISMATCHING_WINDOWS(ICLKernel::window(), window);
+
+    const auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
+    auto       dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
+
+    Window slice_in = window.first_slice_window_4D().collapse(ICLKernel::window(), 2, 4);
+
+    // Setup dst slice
+    Window slice_out(slice_in);
+    slice_out.set(Window::DimX, Window::Dimension(0, 0, 0));
+    slice_out.set(Window::DimY, Window::Dimension(0, 0, 0));
+    slice_out.set(Window::DimZ, Window::Dimension(0, 0, 0));
+    slice_out.set(3, Window::Dimension(0, 0, 0));
+
+    do
+    {
+        unsigned int idx = 0;
+        add_4D_tensor_argument(idx, src, slice_in);
+        add_4D_tensor_argument(idx, dst, slice_out);
+        enqueue(queue, *this, slice_in, lws_hint());
+    }
+    while(window.slide_window_slice_4D(slice_in) && window.slide_window_slice_4D(slice_out));
+}
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
\ No newline at end of file
diff --git a/src/core/gpu/cl/kernels/ClPermuteKernel.h b/src/core/gpu/cl/kernels/ClPermuteKernel.h
new file mode 100644
index 0000000..4cc7249
--- /dev/null
+++ b/src/core/gpu/cl/kernels/ClPermuteKernel.h
@@ -0,0 +1,86 @@
+/*
+ * Copyright (c) 2018-2021 Arm Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef ARM_COMPUTE_CL_PERMUTE_KERNEL_H
+#define ARM_COMPUTE_CL_PERMUTE_KERNEL_H
+
+#include "src/core/common/Macros.h"
+#include "src/core/gpu/cl/ClCompileContext.h"
+#include "src/core/gpu/cl/IClKernel.h"
+
+namespace arm_compute
+{
+namespace opencl
+{
+namespace kernels
+{
+/** OpenCL kernel to perform tensor permutation.
+ *
+ * Permutes given a permutation vector
+ */
+class ClPermuteKernel : public ICLKernel
+{
+public:
+    ClPermuteKernel() = default;
+    ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClPermuteKernel);
+    /** Set the src and dst of the kernel.
+     *
+     * @note Arbitrary permutation vectors are supported with rank not greater than 4
+     *
+     * @param[in] src  The src tensor info. Data types supported: All.
+     * @param[in] dst  The dst tensor info. Data types supported: Same as @p src
+     * @param[in] perm Permutation vector
+     */
+    void configure(const ITensorInfo *src, ITensorInfo *dst, const PermutationVector &perm);
+    /** Set the src and dst of the kernel.
+     *
+     * @note Arbitrary permutation vectors are supported with rank not greater than 4
+     *
+     * @param[in] compile_context The compile context to be used.
+     * @param[in] src             The src tensor info. Data types supported: All.
+     * @param[in] dst             The dst tensor info. Data types supported: Same as @p src
+     * @param[in] perm            Permutation vector
+     */
+    void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst, const PermutationVector &perm);
+    /** Static function to check if given info will lead to a valid configuration of @ref ClPermuteKernel
+     *
+     * @note Arbitrary permutation vectors are supported with rank not greater than 4
+     *
+     * @param[in] src  The src tensor info. Data types supported: All.
+     * @param[in] dst  The dst tensor info. Data types supported: same as @p src.
+     * @param[in] perm Permutation vector
+     *
+     * @return a status
+     */
+    static Status validate(const ITensorInfo *src, const ITensorInfo *dst, const PermutationVector &perm);
+
+    // Inherited methods overridden:
+    void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
+
+private:
+    PermutationVector _perm{};
+};
+} // namespace kernels
+} // namespace opencl
+} // namespace arm_compute
+#endif /*ARM_COMPUTE_CL_PERMUTE_KERNEL_H */
diff --git a/src/core/CL/kernels/CLReshapeLayerKernel.cpp b/src/core/gpu/cl/kernels/ClReshapeKernel.cpp
similarity index 60%
rename from src/core/CL/kernels/CLReshapeLayerKernel.cpp
rename to src/core/gpu/cl/kernels/ClReshapeKernel.cpp
index 58d7843..4da3fa0 100644
--- a/src/core/CL/kernels/CLReshapeLayerKernel.cpp
+++ b/src/core/gpu/cl/kernels/ClReshapeKernel.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017-2020 Arm Limited.
+ * Copyright (c) 2017-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -21,7 +21,7 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-#include "src/core/CL/kernels/CLReshapeLayerKernel.h"
+#include "src/core/gpu/cl/kernels/ClReshapeKernel.h"
 
 #include "arm_compute/core/CL/CLHelpers.h"
 #include "arm_compute/core/CL/CLKernelLibrary.h"
@@ -38,73 +38,77 @@
 
 #include <string>
 
-/** [CLReshapeLayerKernel Kernel] **/
+/** [ClReshapeKernel Kernel] **/
 namespace arm_compute
 {
+namespace opencl
+{
+namespace kernels
+{
 namespace
 {
-Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output)
+Status validate_arguments(const ITensorInfo *src, const ITensorInfo *dst)
 {
-    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(input);
-    ARM_COMPUTE_RETURN_ERROR_ON(input->data_type() == DataType::UNKNOWN);
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(src);
+    ARM_COMPUTE_RETURN_ERROR_ON(src->data_type() == DataType::UNKNOWN);
 
-    ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
-    ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(input, output);
-    ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape().total_size() != output->tensor_shape().total_size());
+    ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(src, dst);
+    ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(src, dst);
+    ARM_COMPUTE_RETURN_ERROR_ON(src->tensor_shape().total_size() != dst->tensor_shape().total_size());
 
     return Status{};
 }
 } // namespace
 
-void CLReshapeLayerKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output)
+void ClReshapeKernel::configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst)
 {
-    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
-    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input, output));
+    ARM_COMPUTE_ERROR_ON_NULLPTR(src, dst);
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(src, dst));
 
-    auto padding_info = get_padding_info({ input, output });
+    auto padding_info = get_padding_info({ src, dst });
 
     // Create kernel
-    std::set<std::string> build_opts = { "-DDATA_TYPE=" + get_cl_unsigned_type_from_element_size(input->element_size()) };
+    std::set<std::string> build_opts = { "-DDATA_TYPE=" + get_cl_unsigned_type_from_element_size(src->element_size()) };
     _kernel                          = create_kernel(compile_context, "reshape_layer", build_opts);
 
     // Add static arguments
-    const cl_int2 input_shape =
+    const cl_int2 src_shape =
     {
         {
-            static_cast<cl_int>(input->tensor_shape()[0]),
-            static_cast<cl_int>(input->tensor_shape()[1])
+            static_cast<cl_int>(src->tensor_shape()[0]),
+            static_cast<cl_int>(src->tensor_shape()[1])
         }
     };
-    const cl_int2 output_shape =
+    const cl_int2 dst_shape =
     {
         {
-            static_cast<cl_int>(output->tensor_shape()[0]),
-            static_cast<cl_int>(output->tensor_shape()[1])
+            static_cast<cl_int>(dst->tensor_shape()[0]),
+            static_cast<cl_int>(dst->tensor_shape()[1])
         }
     };
-    unsigned int idx = 2 * num_arguments_per_3D_tensor(); // Skip the input and output parameters
-    _kernel.setArg<cl_int2>(idx++, input_shape);
-    _kernel.setArg<cl_int2>(idx++, output_shape);
+    unsigned int idx = 2 * num_arguments_per_3D_tensor(); // Skip the src and dst parameters
+    _kernel.setArg<cl_int2>(idx++, src_shape);
+    _kernel.setArg<cl_int2>(idx++, dst_shape);
 
     // Configure kernel window
-    Window win = calculate_max_window(*input);
+    Window win = calculate_max_window(*src);
 
-    // Set the output valid region
-    output->set_valid_region(ValidRegion(Coordinates(), output->tensor_shape()));
+    // Set the dst valid region
+    dst->set_valid_region(ValidRegion(Coordinates(), dst->tensor_shape()));
     ICLKernel::configure_internal(win);
 
     ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
 }
 
-Status CLReshapeLayerKernel::validate(const ITensorInfo *input, const ITensorInfo *output)
+Status ClReshapeKernel::validate(const ITensorInfo *src, const ITensorInfo *dst)
 {
-    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output));
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(src, dst));
 
     return Status{};
 }
 
-void CLReshapeLayerKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
+void ClReshapeKernel::run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue)
 {
     ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
     ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(IKernel::window(), window);
@@ -115,11 +119,13 @@
     const auto src = utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_SRC));
     auto       dst = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_DST));
 
-    // Set inputs
+    // Set srcs
     unsigned int idx = 0;
     add_3D_tensor_argument(idx, src, window_collapsed);
     add_3D_tensor_argument(idx, dst, window_collapsed);
     enqueue(queue, *this, slice, lws_hint());
 }
+} // namespace kernels
+} // namespace opencl
 } // namespace arm_compute
-/** [CLReshapeLayerKernel Kernel] **/
+/** [ClReshapeKernel Kernel] **/
diff --git a/src/core/CL/kernels/CLReshapeLayerKernel.h b/src/core/gpu/cl/kernels/ClReshapeKernel.h
similarity index 61%
rename from src/core/CL/kernels/CLReshapeLayerKernel.h
rename to src/core/gpu/cl/kernels/ClReshapeKernel.h
index 902c446..ee835c0 100644
--- a/src/core/CL/kernels/CLReshapeLayerKernel.h
+++ b/src/core/gpu/cl/kernels/ClReshapeKernel.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2017-2020 Arm Limited.
+ * Copyright (c) 2017-2021 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -21,39 +21,46 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-#ifndef ARM_COMPUTE_CLRESHAPELAYERKERNEL_H
-#define ARM_COMPUTE_CLRESHAPELAYERKERNEL_H
+#ifndef ARM_COMPUTE_CL_RESHAPE_KERNEL_H
+#define ARM_COMPUTE_CL_RESHAPE_KERNEL_H
 
-#include "arm_compute/core/Types.h"
-#include "src/core/CL/ICLKernel.h"
+#include "src/core/common/Macros.h"
+#include "src/core/gpu/cl/ClCompileContext.h"
+#include "src/core/gpu/cl/IClKernel.h"
 
 namespace arm_compute
 {
-class ICLTensor;
-
+namespace opencl
+{
+namespace kernels
+{
 /** Interface for the kernel to perform tensor reshaping */
-class CLReshapeLayerKernel : public ICLKernel
+class ClReshapeKernel : public ICLKernel
 {
 public:
-    /** Set the input and output of the kernel
+    ClReshapeKernel() = default;
+    ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE(ClReshapeKernel);
+    /** Set the src and dst of the kernel
      *
      * @param[in]  compile_context The compile context to be used.
-     * @param[in]  input           Source tensor info. Data type supported: All.
-     * @param[out] output          Destination tensor info. Data type supported: Same as @p input
+     * @param[in]  src             Source tensor info. Data type supported: All.
+     * @param[out] dst             Destination tensor info. Data type supported: Same as @p src
      */
-    void configure(const CLCompileContext &compile_context, const ITensorInfo *input, ITensorInfo *output);
+    void configure(const CLCompileContext &compile_context, const ITensorInfo *src, ITensorInfo *dst);
 
-    /** Static function to check if given info will lead to a valid configuration of @ref CLReshapeLayerKernel
+    /** Static function to check if given info will lead to a valid configuration of @ref ClReshapeKernel
      *
-     * @param[in] input  Source tensor info. Data type supported: All
-     * @param[in] output Destination tensor info. Data type supported: Same as @p input
+     * @param[in] src Source tensor info. Data type supported: All
+     * @param[in] dst Destination tensor info. Data type supported: Same as @p src
      *
      * @return a status
      */
-    static Status validate(const ITensorInfo *input, const ITensorInfo *output);
+    static Status validate(const ITensorInfo *src, const ITensorInfo *dst);
 
     // Inherited methods overridden:
     void run_op(ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override;
 };
+} // namespace opencl
+} // namespace kernels
 } // namespace arm_compute
-#endif /*ARM_COMPUTE_CLRESHAPELAYERKERNEL_H */
+#endif /*ARM_COMPUTE_CL_RESHAPE_KERNEL_H */