replace cuDNN alphas and betas with coefficient values

Give cuDNN {0, 1} constants for controlling accumulation through the
alpha and beta coefficients.

diff --git a/include/caffe/util/cudnn.hpp b/include/caffe/util/cudnn.hpp
index f496362..b531dd5 100644 (file)
--- a/include/caffe/util/cudnn.hpp
+++ b/include/caffe/util/cudnn.hpp
@@ -50,10 +50,14 @@ template <typename Dtype> class dataType;
 template<> class dataType<float>  {
  public:
   static const cudnnDataType_t type = CUDNN_DATA_FLOAT;
+  static float oneval, zeroval;
+  static const void *one, *zero;
 };
 template<> class dataType<double> {
  public:
   static const cudnnDataType_t type = CUDNN_DATA_DOUBLE;
+  static double oneval, zeroval;
+  static const void *one, *zero;
 };
 
 template <typename Dtype>
@@ -102,9 +106,9 @@ inline void setConvolutionDesc(cudnnConvolutionDescriptor_t* conv,
 }
 
 template <typename Dtype>
-inline void createPoolingDesc(cudnnPoolingDescriptor_t* conv,
+inline void createPoolingDesc(cudnnPoolingDescriptor_t* pool_desc,
     PoolingParameter_PoolMethod poolmethod, cudnnPoolingMode_t* mode,
-    int h, int w, int stride_h, int stride_w) {
+    int h, int w, int pad_h, int pad_w, int stride_h, int stride_w) {
   switch (poolmethod) {
   case PoolingParameter_PoolMethod_MAX:
     *mode = CUDNN_POOLING_MAX;
@@ -115,9 +119,9 @@ inline void createPoolingDesc(cudnnPoolingDescriptor_t* conv,
   default:
     LOG(FATAL) << "Unknown pooling method.";
   }
-  CUDNN_CHECK(cudnnCreatePoolingDescriptor(conv));
-  CUDNN_CHECK(cudnnSetPooling2dDescriptor(*conv, *mode, h, w,
-        0, 0, stride_h, stride_w));
+  CUDNN_CHECK(cudnnCreatePoolingDescriptor(pool_desc));
+  CUDNN_CHECK(cudnnSetPooling2dDescriptor(*pool_desc, *mode, h, w,
+        pad_h, pad_w, stride_h, stride_w));
 }
 
 }  // namespace cudnn

diff --git a/src/caffe/layers/cudnn_conv_layer.cu b/src/caffe/layers/cudnn_conv_layer.cu
index 4a70c69..ed0d91d 100644 (file)
--- a/src/caffe/layers/cudnn_conv_layer.cu
+++ b/src/caffe/layers/cudnn_conv_layer.cu
@@ -21,9 +21,6 @@ void CuDNNConvolutionLayer<Dtype>::Forward_gpu(
 
     // Forward through cuDNN in parallel over groups.
     for (int g = 0; g < this->group_; g++) {
-      Dtype alpha = 1.0;
-      Dtype beta = 0.0;
-
       cudnnConvolutionFwdAlgo_t algo;
 
       // get the desired convolution algorithm
@@ -56,23 +53,21 @@ void CuDNNConvolutionLayer<Dtype>::Forward_gpu(
 
       // Filters.
       CUDNN_CHECK(cudnnConvolutionForward(handle_[g],
-            reinterpret_cast<void *>(&alpha),
+            cudnn::dataType<Dtype>::one,
             bottom_descs_[i], bottom_data + bottom_offset_ * g,
             filter_desc_, weight + weight_offset_ * g,
             conv_descs_[i],
             algo, workspace, workspaceSizeInBytes,
-            reinterpret_cast<void *>(&beta),
+            cudnn::dataType<Dtype>::zero,
             top_descs_[i], top_data + top_offset_ * g));
 
       // Bias.
       if (this->bias_term_) {
         const Dtype* bias_data = this->blobs_[1]->gpu_data();
-        Dtype alpha = 1.0;
-        Dtype beta = 1.0;
         CUDNN_CHECK(cudnnAddTensor(handle_[g], CUDNN_ADD_SAME_C,
-              reinterpret_cast<void *>(&alpha),
+              cudnn::dataType<Dtype>::one,
               bias_desc_, bias_data + bias_offset_ * g,
-              reinterpret_cast<void *>(&beta),
+              cudnn::dataType<Dtype>::one,
               top_descs_[i], top_data + top_offset_ * g));
       }
     }
@@ -105,25 +100,22 @@ void CuDNNConvolutionLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
     for (int g = 0; g < this->group_; g++) {
       // Gradient w.r.t. bias.
       if (this->bias_term_ && this->param_propagate_down_[1]) {
-        Dtype alpha = 1.0;
-        Dtype beta = 1.0;
         CUDNN_CHECK(cudnnConvolutionBackwardBias(handle_[0*this->group_ + g],
-              reinterpret_cast<void *>(&alpha),
+              cudnn::dataType<Dtype>::one,
               top_descs_[i],  top_diff + top_offset_ * g,
-              reinterpret_cast<void *>(&beta),
+              cudnn::dataType<Dtype>::one,
               bias_desc_, bias_diff + bias_offset_ * g));
       }
 
       // Gradient w.r.t. weights.
       if (this->param_propagate_down_[0]) {
         const Dtype* bottom_data = bottom[i]->gpu_data();
-        Dtype alpha = 1.0;
-        Dtype beta = 1.0;
         CUDNN_CHECK(cudnnConvolutionBackwardFilter(handle_[1*this->group_ + g],
-              reinterpret_cast<void *>(&alpha),
+              cudnn::dataType<Dtype>::one,
               bottom_descs_[i], bottom_data + bottom_offset_ * g,
               top_descs_[i],    top_diff + top_offset_ * g,
-              conv_descs_[i], reinterpret_cast<void *>(&beta),
+              conv_descs_[i],
+              cudnn::dataType<Dtype>::one,
               filter_desc_, weight_diff + weight_offset_ * g));
       }
 
@@ -133,13 +125,12 @@ void CuDNNConvolutionLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
           weight = this->blobs_[0]->gpu_data();
         }
         Dtype* bottom_diff = bottom[i]->mutable_gpu_diff();
-        Dtype alpha = 1.0;
-        Dtype beta = 0.0;
         CUDNN_CHECK(cudnnConvolutionBackwardData(handle_[2*this->group_ + g],
-              reinterpret_cast<void *>(&alpha),
+              cudnn::dataType<Dtype>::one,
               filter_desc_, weight + weight_offset_ * g,
-              top_descs_[i],    top_diff + top_offset_ * g,
-              conv_descs_[i], reinterpret_cast<void *>(&beta),
+              top_descs_[i], top_diff + top_offset_ * g,
+              conv_descs_[i],
+              cudnn::dataType<Dtype>::zero,
               bottom_descs_[i], bottom_diff + bottom_offset_ * g));
       }
     }

diff --git a/src/caffe/layers/cudnn_pooling_layer.cu b/src/caffe/layers/cudnn_pooling_layer.cu
index be7c4a8..a952b85 100644 (file)
--- a/src/caffe/layers/cudnn_pooling_layer.cu
+++ b/src/caffe/layers/cudnn_pooling_layer.cu
@@ -14,14 +14,10 @@ void CuDNNPoolingLayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
     const vector<Blob<Dtype>*>& top) {
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* top_data = top[0]->mutable_gpu_data();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnPoolingForward(handle_, pooling_desc_,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         top_desc_, top_data));
 }
 
@@ -35,15 +31,11 @@ void CuDNNPoolingLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
   const Dtype* top_data = top[0]->gpu_data();
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* bottom_diff = bottom[0]->mutable_gpu_diff();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnPoolingBackward(handle_, pooling_desc_,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         top_desc_, top_data, top_desc_, top_diff,
         bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         bottom_desc_, bottom_diff));
 }
 

diff --git a/src/caffe/layers/cudnn_relu_layer.cu b/src/caffe/layers/cudnn_relu_layer.cu
index b9d0870..21d1485 100644 (file)
--- a/src/caffe/layers/cudnn_relu_layer.cu
+++ b/src/caffe/layers/cudnn_relu_layer.cu
@@ -17,15 +17,11 @@ void CuDNNReLULayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
 
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* top_data = top[0]->mutable_gpu_data();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationForward(this->handle_,
         CUDNN_ACTIVATION_RELU,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->top_desc_, top_data));
 }
 
@@ -46,16 +42,12 @@ void CuDNNReLULayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
   const Dtype* top_diff = top[0]->gpu_diff();
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* bottom_diff = bottom[0]->mutable_gpu_diff();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationBackward(this->handle_,
         CUDNN_ACTIVATION_RELU,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->top_desc_, top_data, this->top_desc_, top_diff,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->bottom_desc_, bottom_diff));
 }
 

diff --git a/src/caffe/layers/cudnn_sigmoid_layer.cu b/src/caffe/layers/cudnn_sigmoid_layer.cu
index 9bb9150..7a06cf7 100644 (file)
--- a/src/caffe/layers/cudnn_sigmoid_layer.cu
+++ b/src/caffe/layers/cudnn_sigmoid_layer.cu
@@ -12,15 +12,11 @@ void CuDNNSigmoidLayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
     const vector<Blob<Dtype>*>& top) {
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* top_data = top[0]->mutable_gpu_data();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationForward(this->handle_,
         CUDNN_ACTIVATION_SIGMOID,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->top_desc_, top_data));
 }
 
@@ -36,16 +32,12 @@ void CuDNNSigmoidLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
   const Dtype* top_diff = top[0]->gpu_diff();
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* bottom_diff = bottom[0]->mutable_gpu_diff();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationBackward(this->handle_,
         CUDNN_ACTIVATION_SIGMOID,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->top_desc_, top_data, this->top_desc_, top_diff,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->bottom_desc_, bottom_diff));
 }
 

diff --git a/src/caffe/layers/cudnn_softmax_layer.cu b/src/caffe/layers/cudnn_softmax_layer.cu
index 59c304f..a9e2fce 100644 (file)
--- a/src/caffe/layers/cudnn_softmax_layer.cu
+++ b/src/caffe/layers/cudnn_softmax_layer.cu
@@ -16,15 +16,11 @@ void CuDNNSoftmaxLayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
     const vector<Blob<Dtype>*>& top) {
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* top_data = top[0]->mutable_gpu_data();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnSoftmaxForward(handle_, CUDNN_SOFTMAX_ACCURATE,
         CUDNN_SOFTMAX_MODE_CHANNEL,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         top_desc_, top_data));
 }
 
@@ -37,14 +33,11 @@ void CuDNNSoftmaxLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
     const Dtype* bottom_data = bottom[0]->gpu_data();
     Dtype* bottom_diff = bottom[0]->mutable_gpu_diff();
 
-    Dtype alpha = 1.0;
-    Dtype beta = 0.0;
-
     CUDNN_CHECK(cudnnSoftmaxBackward(handle_, CUDNN_SOFTMAX_ACCURATE,
           CUDNN_SOFTMAX_MODE_CHANNEL,
-          reinterpret_cast<void *>(&alpha),
+          cudnn::dataType<Dtype>::one,
           top_desc_, top_data, top_desc_, top_diff,
-          reinterpret_cast<void *>(&beta),
+          cudnn::dataType<Dtype>::zero,
           bottom_desc_, bottom_diff));
   }
 }

diff --git a/src/caffe/layers/cudnn_tanh_layer.cu b/src/caffe/layers/cudnn_tanh_layer.cu
index e008b0d..d287f6f 100644 (file)
--- a/src/caffe/layers/cudnn_tanh_layer.cu
+++ b/src/caffe/layers/cudnn_tanh_layer.cu
@@ -12,15 +12,11 @@ void CuDNNTanHLayer<Dtype>::Forward_gpu(const vector<Blob<Dtype>*>& bottom,
     const vector<Blob<Dtype>*>& top) {
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* top_data = top[0]->mutable_gpu_data();
-
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationForward(this->handle_,
         CUDNN_ACTIVATION_TANH,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->top_desc_, top_data));
 }
 
@@ -37,15 +33,12 @@ void CuDNNTanHLayer<Dtype>::Backward_gpu(const vector<Blob<Dtype>*>& top,
   const Dtype* bottom_data = bottom[0]->gpu_data();
   Dtype* bottom_diff = bottom[0]->mutable_gpu_diff();
 
-  Dtype alpha = 1.0;
-  Dtype beta = 0.0;
-
   CUDNN_CHECK(cudnnActivationBackward(this->handle_,
         CUDNN_ACTIVATION_TANH,
-        reinterpret_cast<void *>(&alpha),
+        cudnn::dataType<Dtype>::one,
         this->top_desc_, top_data, this->top_desc_, top_diff,
         this->bottom_desc_, bottom_data,
-        reinterpret_cast<void *>(&beta),
+        cudnn::dataType<Dtype>::zero,
         this->bottom_desc_, bottom_diff));
 }
 

diff --git a/src/caffe/util/cudnn.cpp b/src/caffe/util/cudnn.cpp
new file mode 100644 (file)
index 0000000..1772f00
--- /dev/null
+++ b/src/caffe/util/cudnn.cpp
@@ -0,0 +1,23 @@
+#ifdef USE_CUDNN
+#include "caffe/util/cudnn.hpp"
+
+namespace caffe {
+namespace cudnn {
+
+float dataType<float>::oneval = 1.0;
+float dataType<float>::zeroval = 0.0;
+const void* dataType<float>::one =
+    static_cast<void *>(&dataType<float>::oneval);
+const void* dataType<float>::zero =
+    static_cast<void *>(&dataType<float>::zeroval);
+
+double dataType<double>::oneval = 1.0;
+double dataType<double>::zeroval = 0.0;
+const void* dataType<double>::one =
+    static_cast<void *>(&dataType<double>::oneval);
+const void* dataType<double>::zero =
+    static_cast<void *>(&dataType<double>::zeroval);
+
+}  // namespace cudnn
+}  // namespace caffe
+#endif