First version of im2col kernel, unoptimized but working

2 files changed, 110 insertions, 1 deletions

diff --git a/src/kernels/levelx/im2col.opencl b/src/kernels/levelx/im2col.opencl
new file mode 100644
index 00000000..3f10881f
--- /dev/null
+++ b/src/kernels/levelx/im2col.opencl
@@ -0,0 +1,73 @@
+
+// =================================================================================================
+// This file is part of the CLBlast project. The project is licensed under Apache Version 2.0. This
+// project loosely follows the Google C++ styleguide and uses a tab-size of two spaces and a max-
+// width of 100 characters per line.
+//
+// Author(s):
+//   Cedric Nugteren <www.cedricnugteren.nl>
+//
+// This file contains the im2col kernel.
+//
+// =================================================================================================
+
+// Enables loading of this file using the C++ pre-processor's #include (C++11 standard raw string
+// literal). Comment-out this line for syntax-highlighting when developing.
+R"(
+
+// =================================================================================================
+
+#define WGS1 16
+#define WGS2 16
+
+__kernel __attribute__((reqd_work_group_size(WGS1, WGS2, 1)))
+void im2col(const int input_h, const int input_w,
+            const int output_h, const int output_w,
+            const int kernel_h, const int kernel_w,
+            const int pad_h, const int pad_w,
+            const int stride_h, const int stride_w,
+            const int dilation_h, const int dilation_w,
+            const __global real* restrict im_buffer, const int im_offset,
+            __global real* col_buffer, const int col_offset) {
+
+  // Thread IDs
+  const int h_id = get_global_id(0); // image height, max 'output_h'
+  const int w_id = get_global_id(1); // image width, max 'output_w'
+  const int c_id = get_global_id(2); // input channels
+  if (h_id < output_h && w_id < output_w) {
+
+    #pragma unroll
+    for (int kh_id = 0; kh_id < kernel_h; ++kh_id) { // kernel height
+      #pragma unroll
+      for (int kw_id = 0; kw_id < kernel_w; ++kw_id) { // kernel width
+
+        // Retrieves the input value
+        const int h_index = -pad_h + kh_id * dilation_h + stride_h * h_id;
+        const int w_index = -pad_w + kw_id * dilation_w + stride_w * w_id;
+        real val;
+        if (h_index >= 0 && h_index < input_h &&
+            w_index >= 0 && w_index < input_w) {
+          const int input_index = w_index + input_w * (h_index + input_h * c_id);
+          val = im_buffer[input_index + im_offset];
+        }
+        else {
+          SetToZero(val);
+        }
+
+        // Sets the output value
+        const int kernel_index = kw_id + kernel_w * kh_id;
+        const int patch_index = w_id + output_w * h_id;
+        const int output_index = patch_index + kernel_index * output_w * output_h +
+                                  c_id * output_w * output_h * kernel_h * kernel_w;
+        col_buffer[output_index + col_offset] = val;
+      }
+    }
+  }
+}
+
+// =================================================================================================
+
+// End of the C++11 raw string literal
+)"
+
+// =================================================================================================
diff --git a/src/routines/levelx/xim2col.cpp b/src/routines/levelx/xim2col.cpp
index 150220d6..10c9c10c 100644
--- a/src/routines/levelx/xim2col.cpp
+++ b/src/routines/levelx/xim2col.cpp
@@ -23,7 +23,7 @@ namespace clblast {
 template <typename T>
 Xim2col<T>::Xim2col(Queue &queue, EventPointer event, const std::string &name):
         Routine(queue, event, name, {}, PrecisionValue<T>(), {}, {
-#include "../../kernels/level3/level3.opencl"
+#include "../../kernels/levelx/im2col.opencl"
         }) {
 }
 
@@ -40,6 +40,42 @@ void Xim2col<T>::DoIm2col(const size_t channels, const size_t height, const size
 
   // Makes sure all dimensions are larger than zero
   if ((channels == 0) || (height == 0) || (width == 0)) { throw BLASError(StatusCode::kInvalidDimension); }
+
+  // Sets the output height and width
+  const auto size_h = height + 2 * pad_h;
+  const auto padding_h = dilation_h * (kernel_h - 1) + 1;
+  const auto output_h = (size_h >= padding_h) ? (size_h - padding_h) / stride_h + 1 : 1;
+  const auto size_w = width + 2 * pad_w;
+  const auto padding_w = dilation_w * (kernel_w - 1) + 1;
+  const auto output_w = (size_w >= padding_w) ? (size_w - padding_w) / stride_w + 1 : 1;
+
+  // Retrieves the Xcopy kernel from the compiled binary
+  auto kernel = Kernel(program_, "im2col");
+
+  // Sets the kernel arguments
+  kernel.SetArgument(0, static_cast<int>(height));
+  kernel.SetArgument(1, static_cast<int>(width));
+  kernel.SetArgument(2, static_cast<int>(output_h));
+  kernel.SetArgument(3, static_cast<int>(output_w));
+  kernel.SetArgument(4, static_cast<int>(kernel_h));
+  kernel.SetArgument(5, static_cast<int>(kernel_w));
+  kernel.SetArgument(6, static_cast<int>(pad_h));
+  kernel.SetArgument(7, static_cast<int>(pad_w));
+  kernel.SetArgument(8, static_cast<int>(stride_h));
+  kernel.SetArgument(9, static_cast<int>(stride_w));
+  kernel.SetArgument(10, static_cast<int>(dilation_h));
+  kernel.SetArgument(11, static_cast<int>(dilation_w));
+  kernel.SetArgument(12, im_buffer());
+  kernel.SetArgument(13, static_cast<int>(im_offset));
+  kernel.SetArgument(14, col_buffer());
+  kernel.SetArgument(15, static_cast<int>(col_offset));
+
+  // Launches the kernel
+  const auto h_ceiled = Ceil(output_h, 16);
+  const auto w_ceiled = Ceil(output_w, 16);
+  auto global = std::vector<size_t>{h_ceiled, w_ceiled, channels};
+  auto local = std::vector<size_t>{16, 16, 1};
+  RunKernel(kernel, queue_, device_, global, local, event_);
 }
 
 // =================================================================================================