path: root/tensorflow/core/kernels/xsmm_conv2d_test.cc

/* Copyright 2016 The TensorFlow Authors. All Rights Reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/

#include "include/libxsmm.h"
#include "tensorflow/core/framework/fake_input.h"
#include "tensorflow/core/graph/graph.h"
#include "tensorflow/core/graph/node_builder.h"
#include "tensorflow/core/kernels/conv_ops.h"
#include "tensorflow/core/kernels/ops_testutil.h"
#include "tensorflow/core/platform/test.h"

namespace tensorflow {
namespace {

typedef struct {
  int nImg;
  int nIfm;
  int nOfm;
  int ifhp;
  int ifwp;
  int ifh;
  int ifw;
  int ofhp;
  int ofwp;
  int ofh;
  int ofw;
  int pad_h;
  int pad_w;
  int pad_h_in;
  int pad_w_in;
  int pad_h_out;
  int pad_w_out;
  int kh;
  int kw;
  int stride_h;
  int stride_w;
} naive_conv_t;

LIBXSMM_INLINE void naive_copy_NCHW_to_NHWC(const float* nchw, Tensor& nhwc,
                                            int N, int H, int W, int C) {
  LIBXSMM_VLA_DECL(4, const float, input, nchw, C, H, W);
  int n, h, w, c;
  auto output = nhwc.flat<float>();
  for (n = 0; n < N; n++) {
    for (h = 0; h < H; h++) {
      for (w = 0; w < W; w++) {
        for (c = 0; c < C; c++) {
          output(n * H * W * C + h * W * C + w * C + c) =
              LIBXSMM_VLA_ACCESS(4, input, n, c, h, w, C, H, W);
        }
      }
    }
  }
}

LIBXSMM_INLINE void naive_copy_KCRS_to_RSCK(const float* kcrs, Tensor& rsck,
                                            int R, int S, int C, int K) {
  LIBXSMM_VLA_DECL(4, const float, input, kcrs, C, R, S);
  int r, s, c, k;
  auto output = rsck.flat<float>();

  for (r = 0; r < R; r++) {
    for (s = 0; s < S; s++) {
      for (c = 0; c < C; c++) {
        for (k = 0; k < K; k++) {
          output(r * S * C * K + s * C * K + c * K + k) =
              LIBXSMM_VLA_ACCESS(4, input, k, c, r, s, C, R, S);
        }
      }
    }
  }
}

LIBXSMM_INLINE void zero_buf(float* buf, long size) {
  int i;
  for (i = 0; i < size; ++i) {
    buf[i] = 0.0f;
  }
}

LIBXSMM_INLINE void copy_buf(Tensor& dst, float* src, long size) {
  long i;
  auto output = dst.flat<float>();
  for (i = 0; i < size; ++i) output(i) = src[i];
}

LIBXSMM_INLINE void init_buf(float* buf, long size, int initPos, int initOne) {
  int i;
  zero_buf(buf, size);
  for (i = 0; i < size; ++i) {
    buf[i] =
        (float)((initOne != 0)
                    ? 1.0
                    : ((initPos != 0) ? drand48() : (0.05 - drand48() / 10.0)));
  }
}

LIBXSMM_INLINE void naive_conv_fp(naive_conv_t* param, const float* input,
                                  float* output, const float* filter) {
  int nImg = param->nImg;
  int nIfm = param->nIfm;
  int nOfm = param->nOfm;
  int ifhp = param->ifhp;
  int ifwp = param->ifwp;
  int ofhp = param->ofhp;
  int ofwp = param->ofwp;
  int ifh = param->ifh;
  int ifw = param->ifw;
  int ofh = param->ofh;
  int ofw = param->ofw;
  int pad_h = param->pad_h;
  int pad_w = param->pad_w;
  int pad_h_in = param->pad_h_in;
  int pad_w_in = param->pad_w_in;
  int pad_h_out = param->pad_h_out;
  int pad_w_out = param->pad_w_out;
  int kh = param->kh;
  int kw = param->kw;
  int stride_h = param->stride_h;
  int stride_w = param->stride_w;
  /* loop counters */
  int img, ofm, ifm, oj, oi, ij, ii, kj, ki;

  LIBXSMM_VLA_DECL(4, float, output_t, output + (pad_w_out * ofwp + pad_h_out),
                   nOfm, ofhp, ofwp);
  LIBXSMM_VLA_DECL(4, const float, input_t,
                   input + (pad_w_in * ifwp + pad_h_in), nIfm, ifhp, ifwp);
  LIBXSMM_VLA_DECL(4, const float, filter_t, filter, nIfm, kh, kw);

  for (img = 0; img < nImg; ++img) {
    for (ofm = 0; ofm < nOfm; ++ofm) {
      for (ifm = 0; ifm < nIfm; ++ifm) {
        for (oj = 0; oj < ofh; ++oj) {
          ij = oj * stride_h - pad_h;
          for (oi = 0; oi < ofw; ++oi) {
            ii = oi * stride_w - pad_w;
            for (kj = 0; kj < kh; ++kj) {
              if (ij + kj < 0 || ij + kj >= ifh) continue;
              for (ki = 0; ki < kw; ++ki) {
                if (ii + ki < 0 || ii + ki >= ifw) continue;
                LIBXSMM_VLA_ACCESS(4, output_t, img, ofm, oj, oi, nOfm, ofhp,
                                   ofwp) +=
                    LIBXSMM_VLA_ACCESS(4, input_t, img, ifm, ij + kj, ii + ki,
                                       nIfm, ifhp, ifwp) *
                    LIBXSMM_VLA_ACCESS(4, filter_t, ofm, ifm, kj, ki, nIfm, kh,
                                       kw);
              }
            }
          }
        }
      }
    }
  }
}

void RunXsmmVsGeneric() {}

class XsmmConv2DTest : public OpsTestBase {
 protected:
  void MakeOp(int stride) {
    TF_CHECK_OK(NodeDefBuilder("xsmm", "Conv2D")
                    .Input(FakeInput(DT_FLOAT))
                    .Input(FakeInput(DT_FLOAT))
                    .Attr("strides", {1, stride, stride, 1})
                    .Attr("padding", "VALID")
                    .Finalize(node_def()));

    TF_ASSERT_OK(InitOp());
  }
};

TEST_F(XsmmConv2DTest, Basic) {
  MakeOp(1);

  // setup scoped allocator, which uses cpu_allocator() for this scope
  const libxsmm_tf_allocator<libxsmm_scratch_allocator> tf_allocator;

  int ifw = 14;   /* input width, "W" */
  int ifh = 14;   /* input height, "H" */
  int nImg = 32;  /* mini-batch size, "N" */
  int nIfm = 64;  /* number of input feature maps, "C" */
  int nOfm = 64;  /* number of output feature maps, "K" */
  int kh = 3;     /* filter height, "R" */
  int kw = 3;     /* filter width, "S" */
  int pad = 0;    /* padding in output */
  int stride = 1; /* stride when accessing inputs */

  int stride_w = stride;
  int stride_h = stride;
  int pad_h = pad;
  int pad_w = pad;

  int pad_h_in = pad_h;
  int pad_w_in = pad_w;

  int pad_h_out = 0;
  int pad_w_out = 0;

  /* deriving some values for naive code */
  int ofh = (ifh + 2 * pad_h - kh) / stride_h + 1;
  int ofw = (ifw + 2 * pad_w - kw) / stride_w + 1;
  int ifhp = ifh + 2 * pad_h_in;
  int ifwp = ifw + 2 * pad_w_in;
  int ofhp = ofh + 2 * pad_h_out;
  int ofwp = ofw + 2 * pad_w_out;

  // Initialization of Filter and Image

  /* allocate data */
  float* naive_input = (float*)libxsmm_aligned_scratch(
      nImg * nIfm * ifhp * ifwp * sizeof(float), 2097152);
  float* naive_output = (float*)libxsmm_aligned_scratch(
      nImg * nOfm * ofhp * ofwp * sizeof(float), 2097152);
  float* naive_filter = (float*)libxsmm_aligned_scratch(
      nOfm * nIfm * kh * kw * sizeof(float), 2097152);
  /* initialize data */
  init_buf(naive_input, nImg * nIfm * ifhp * ifwp, 0, 0);
  zero_buf(naive_output, nImg * nOfm * ofhp * ofwp);
  init_buf(naive_filter, nOfm * nIfm * kh * kw, 0, 0);

  Tensor image(DT_FLOAT, {nImg, ifhp, ifwp, nIfm});

  Tensor filter(DT_FLOAT, {kh, kw, nIfm, nOfm});

  naive_copy_NCHW_to_NHWC(naive_input, image, nImg, ifhp, ifwp, nIfm);
  naive_copy_KCRS_to_RSCK(naive_filter, filter, kh, kw, nIfm, nOfm);

  // Run naive convolution

  naive_conv_t naive_param;

  naive_param.nImg = nImg;
  naive_param.nIfm = nIfm;
  naive_param.nOfm = nOfm;
  naive_param.ifhp = ifhp;
  naive_param.ifwp = ifwp;
  naive_param.ofhp = ofhp;
  naive_param.ofwp = ofwp;
  naive_param.ifh = ifh;
  naive_param.ifw = ifw;
  naive_param.ofh = ofh;
  naive_param.ofw = ofw;
  naive_param.pad_h = pad_h;
  naive_param.pad_w = pad_w;
  naive_param.pad_h_in = pad_h_in;
  naive_param.pad_w_in = pad_w_in;
  naive_param.pad_h_out = pad_h_out;
  naive_param.pad_w_out = pad_w_out;
  naive_param.kh = kh;
  naive_param.kw = kw;
  naive_param.stride_h = stride_h;
  naive_param.stride_w = stride_w;

  naive_conv_fp(&naive_param, naive_input, naive_output, naive_filter);

  AddInputFromArray<float>(image.shape(), image.flat<float>());
  AddInputFromArray<float>(filter.shape(), filter.flat<float>());

  // Run Op (TF)
  TF_ASSERT_OK(RunOpKernel());

  // Check the output.
  Tensor expected(DT_FLOAT, {nImg, ofhp, ofwp, nOfm});
  naive_copy_NCHW_to_NHWC(naive_output, expected, nImg, ofhp, ofwp, nOfm);

  test::ExpectTensorNear<float>(expected, *GetOutput(0), 1e-5);
  libxsmm_free(naive_input);
  libxsmm_free(naive_output);
  libxsmm_free(naive_filter);
}

/*


TEST(XsmmConv2DTest, Basic) {

    auto num_threads =
        ctx->device()->tensorflow_cpu_worker_threads()->num_threads;
    // See libxsmm_dnn.h for this struct definition.
    libxsmm_dnn_conv_desc desc;
    desc.N = batch;
    desc.C = in_depth;
    desc.H = input_rows;
    desc.W = input_cols;
    desc.K = out_depth;
    desc.R = filter_rows;
    desc.S = filter_cols;
    desc.u = stride_rows;
    desc.v = stride_cols;
    desc.pad_h = pad_rows;
    desc.pad_w = pad_cols;
    desc.pad_h_in = pad_rows;  // libxsmm supports only physical padding for now
    desc.pad_w_in = pad_cols;  // libxsmm supports only physical padding for now
    desc.pad_h_out = 0;
    desc.pad_w_out = 0;
    desc.threads = num_threads;
    desc.algo = LIBXSMM_DNN_CONV_ALGO_DIRECT;
    desc.buffer_format = LIBXSMM_DNN_TENSOR_FORMAT_NHWC;
    desc.filter_format =
LIBXSMM_DNN_TENSOR_FORMAT_LIBXSMM;//LIBXSMM_DNN_TENSOR_FORMAT_RSCK;
    desc.fuse_ops = LIBXSMM_DNN_CONV_FUSE_NONE;
    desc.options = LIBXSMM_DNN_CONV_OPTION_NONE;
    desc.datatype = LIBXSMM_DNN_DATATYPE_F32;

    if (!CanUseXsmmConv2D(desc, data_format)) {
      return false;
    }

    auto input_ptr = input.template flat<float>().data();
    auto filter_ptr = filter.template flat<float>().data();
    auto output_ptr = output->template flat<float>().data();

    bool success = functor::XsmmFwdConv2D<CPUDevice, float>()(
        ctx, desc, input_ptr, filter_ptr, output_ptr);
    return success;







}
*/
}  // namespace
}  // namespace tensorflow