Add image to column kernel (#867)

* Add image to column kernel

* Add instances, tests, profiler, example

* Add client example

* Several fixes of image to column

* Fix variable name in device_image_to_column_impl

* Several fixes of image to column profiler

* Fix num_btype calculation

* Make new mesaurements for correct bytes calculation

library/src/tensor_operation_instance/gpu/image_to_column/CMakeLists.txt

@@ -0,0 +1,5 @@
+add_instance_library(device_image_to_column_instance
+   device_image_to_column_nhwc_1d_instance.cpp
+   device_image_to_column_nhwc_2d_instance.cpp
+   device_image_to_column_nhwc_3d_instance.cpp
+)

library/src/tensor_operation_instance/gpu/image_to_column/device_image_to_column_nhwc_1d_instance.cpp

@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/gpu/image_to_column/device_image_to_column_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_image_to_column_nhwc_1d_bf16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<1, GNWC, BF16, BF16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_bf16_instances<1, GNWC>{});
+}
+
+void add_device_image_to_column_nhwc_1d_f16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<1, GNWC, F16, F16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f16_instances<1, GNWC>{});
+}
+
+void add_device_image_to_column_nhwc_1d_f32_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<1, GNWC, F32, F32>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f32_instances<1, GNWC>{});
+}
+
+void add_device_image_to_column_nhwc_1d_i8_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<1, GNWC, int8_t, int8_t>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_i8_instances<1, GNWC>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/image_to_column/device_image_to_column_nhwc_2d_instance.cpp

@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/gpu/image_to_column/device_image_to_column_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_image_to_column_nhwc_2d_bf16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<2, GNHWC, BF16, BF16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_bf16_instances<2, GNHWC>{});
+}
+
+void add_device_image_to_column_nhwc_2d_f16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<2, GNHWC, F16, F16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f16_instances<2, GNHWC>{});
+}
+
+void add_device_image_to_column_nhwc_2d_f32_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<2, GNHWC, F32, F32>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f32_instances<2, GNHWC>{});
+}
+
+void add_device_image_to_column_nhwc_2d_i8_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<2, GNHWC, int8_t, int8_t>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_i8_instances<2, GNHWC>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck

library/src/tensor_operation_instance/gpu/image_to_column/device_image_to_column_nhwc_3d_instance.cpp

@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: MIT
+// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/gpu/image_to_column/device_image_to_column_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+void add_device_image_to_column_nhwc_3d_bf16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<3, GNDHWC, BF16, BF16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_bf16_instances<3, GNDHWC>{});
+}
+
+void add_device_image_to_column_nhwc_3d_f16_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<3, GNDHWC, F16, F16>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f16_instances<3, GNDHWC>{});
+}
+
+void add_device_image_to_column_nhwc_3d_f32_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<3, GNDHWC, F32, F32>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_f32_instances<3, GNDHWC>{});
+}
+
+void add_device_image_to_column_nhwc_3d_i8_instances(
+    std::vector<std::unique_ptr<DeviceImageToColumn<3, GNDHWC, int8_t, int8_t>>>& instances)
+{
+    add_device_operation_instances(instances, device_image_to_column_i8_instances<3, GNDHWC>{});
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck