aboutsummaryrefslogtreecommitdiff
path: root/src/core/CL/kernels/CLAccumulateKernel.cpp
diff options
context:
space:
mode:
authorAnthony Barbier <anthony.barbier@arm.com>2017-09-04 18:44:23 +0100
committerAnthony Barbier <anthony.barbier@arm.com>2018-09-17 13:03:09 +0100
commit6ff3b19ee6120edf015fad8caab2991faa3070af (patch)
treea7a6dcd16dfd56d79fa1b56a313caeebcc939b68 /src/core/CL/kernels/CLAccumulateKernel.cpp
downloadComputeLibrary-6ff3b19ee6120edf015fad8caab2991faa3070af.tar.gz
COMPMID-344 Updated doxygen
Change-Id: I32f7b84daa560e460b77216add529c8fa8b327ae
Diffstat (limited to 'src/core/CL/kernels/CLAccumulateKernel.cpp')
-rw-r--r--src/core/CL/kernels/CLAccumulateKernel.cpp83
1 files changed, 83 insertions, 0 deletions
diff --git a/src/core/CL/kernels/CLAccumulateKernel.cpp b/src/core/CL/kernels/CLAccumulateKernel.cpp
new file mode 100644
index 0000000000..6333f04e71
--- /dev/null
+++ b/src/core/CL/kernels/CLAccumulateKernel.cpp
@@ -0,0 +1,83 @@
+/*
+ * Copyright (c) 2016, 2017 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 "arm_compute/core/CL/kernels/CLAccumulateKernel.h"
+
+#include "arm_compute/core/CL/CLKernelLibrary.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/CL/OpenCL.h"
+#include "arm_compute/core/Error.h"
+#include "arm_compute/core/IAccessWindow.h"
+#include "arm_compute/core/Types.h"
+#include "arm_compute/core/Validate.h"
+
+using namespace arm_compute;
+
+void CLAccumulateKernel::configure(const ICLTensor *input, ICLTensor *accum)
+{
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8);
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::S16);
+
+ // Create kernel
+ _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate"));
+
+ // Make sure _kernel is initialized before calling the parent's configure
+ constexpr unsigned int num_elems_processed_per_iteration = 16;
+ ICLSimple2DKernel::configure(input, accum, num_elems_processed_per_iteration);
+}
+
+void CLAccumulateWeightedKernel::configure(const ICLTensor *input, float alpha, ICLTensor *accum)
+{
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8);
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::U8);
+ ARM_COMPUTE_ERROR_ON(alpha < 0.0 || alpha > 1.0);
+
+ // Create kernel
+ _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate_weighted"));
+
+ // Set static kernel arguments
+ unsigned int idx = 2 * num_arguments_per_2D_tensor(); //Skip the input and output parameters
+ _kernel.setArg(idx++, alpha);
+
+ // Configure kernel window
+ constexpr unsigned int num_elems_processed_per_iteration = 16;
+ ICLSimple2DKernel::configure(input, accum, num_elems_processed_per_iteration);
+}
+
+void CLAccumulateSquaredKernel::configure(const ICLTensor *input, uint32_t shift, ICLTensor *accum)
+{
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::U8);
+ ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(accum, 1, DataType::S16);
+ ARM_COMPUTE_ERROR_ON(shift > 15);
+
+ // Create kernel
+ _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("accumulate_squared"));
+
+ // Set static kernel arguments
+ unsigned int idx = 2 * num_arguments_per_2D_tensor(); //Skip the input and output parameters
+ _kernel.setArg(idx++, shift);
+
+ // Configure kernel window
+ constexpr unsigned int num_elems_processed_per_iteration = 16;
+ ICLSimple2DKernel::configure(input, accum, num_elems_processed_per_iteration);
+}