Merge branch 'gpu-perf' into 'master'

Change the default gpu perf hint to normal but benchmark with high.

See merge request !368

mace/benchmark/benchmark_model.cc

@@ -215,8 +215,8 @@ DEFINE_bool(show_flops, true, "whether to estimate the model's FLOPs");
 DEFINE_int32(warmup_runs, 1, "how many runs to initialize model");
 DEFINE_string(model_data_file, "",
               "model data file name, used when EMBED_MODEL_DATA set to 0");
-DEFINE_int32(gpu_perf_hint, 2, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
-DEFINE_int32(gpu_priority_hint, 1, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
+DEFINE_int32(gpu_perf_hint, 0, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
+DEFINE_int32(gpu_priority_hint, 0, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
 DEFINE_int32(omp_num_threads, 4, "num of openmp threads");
 DEFINE_int32(cpu_power_option, 0,
              "0:DEFAULT/1:HIGH_PERFORMANCE/2:BATTERY_SAVE");

mace/core/runtime/opencl/opencl_runtime.cc

@@ -179,7 +179,7 @@ void OpenCLProfilingTimer::ClearTiming() {
   accumulated_micros_ = 0;
 }
 
-GPUPerfHint OpenCLRuntime::kGPUPerfHint = GPUPerfHint::PERF_DEFAULT;
+GPUPerfHint OpenCLRuntime::kGPUPerfHint = GPUPerfHint::PERF_NORMAL;
 GPUPriorityHint OpenCLRuntime::kGPUPriorityHint =
     GPUPriorityHint::PRIORITY_DEFAULT;
 

mace/python/tools/caffe_converter_lib.py

@@ -450,18 +450,6 @@ class CaffeConverter(object):
             final_op.output_shape_map[final_op.layer.top[0]] = output_shape
             self.resolved_ops.add(activation_op.name)
 
-        if op_def.type in ("Conv2D", "FusedConv2D") and \
-                output_shape[2] == 1 and \
-                ((input_format == 'NCHW' and output_shape[3] == 1) or
-                 (input_format == 'NHWC' and output_shape[1] == 1)):
-            print "convert op %s from CONV to FC" % op.name
-            op_def.type = 'FC'
-            filter_shape = weight_data.shape
-            new_shape = [filter_shape[0],
-                         filter_shape[1] * filter_shape[2] * filter_shape[3],
-                         1, 1]
-            weight_data.reshape(new_shape)
-
         op_def.output.extend([final_op.name + ':0'])
         self.add_output_shape(op_def, output_shape)
         self.net_def.op.extend([op_def])

mace/python/tools/tf_converter_lib.py

@@ -363,17 +363,15 @@ class TFConverter(object):
         op_def.name = op.name
         if op.type == 'DepthwiseConv2dNative':
             op_def.type = 'DepthwiseConv2d'
-            if self.device == 'neon':
-                self.transpose_filter_tensor[get_input_tensor(
-                    op, 1).name] = (3, 2, 0, 1)
         else:
             op_def.type = op.type
-            if self.device == 'neon':
-                self.transpose_filter_tensor[get_input_tensor(
-                    op, 1).name] = (3, 2, 0, 1)
-            else:
-                self.transpose_filter_tensor[get_input_tensor(
-                    op, 1).name] = (0, 1, 3, 2)
+
+        if self.device == 'neon':
+            self.transpose_filter_tensor[get_input_tensor(
+                op, 1).name] = (3, 2, 0, 1)
+        elif op.type == 'Conv2D':
+            self.transpose_filter_tensor[get_input_tensor(
+                op, 1).name] = (0, 1, 3, 2)
         if self.device == 'gpu':
             op_def.input.extend([op.inputs[0].name])
             if op_def.type == 'DepthwiseConv2d':
@@ -402,21 +400,71 @@ class TFConverter(object):
         final_op = op
         self.resolved_ops[op.name] = 1
 
-        # convert global conv to fc
+        if len(self.tf_graph.get(op.name, [])) == 1 and \
+           self.tf_graph[op.name][0].type == 'BiasAdd':
+            bias_add_op = self.tf_graph[op.name][0]
+            if self.device == 'gpu':
+                output_name = self.add_buffer_to_image(
+                    get_input_tensor(bias_add_op, 1).name, "ARGUMENT")
+                op_def.input.extend([output_name])
+            else:
+                op_def.input.extend([get_input_tensor(bias_add_op, 1).name])
+            final_op = bias_add_op
+            self.resolved_ops[bias_add_op.name] = 1
+
+        if len(self.tf_graph.get(final_op.name, [])) == 1 and \
+           self.tf_graph[final_op.name][0].type in activation_name_map:
+            activation_op = self.tf_graph[final_op.name][0]
+            if op_def.type == "Conv2D":
+                op_def.type = "FusedConv2D"
+            fused_act_arg = op_def.arg.add()
+            fused_act_arg.name = 'activation'
+            fused_act_arg.s = activation_name_map[activation_op.type]
+            if activation_op.type == 'Relu6':
+                max_limit_arg = op_def.arg.add()
+                max_limit_arg.name = 'max_limit'
+                max_limit_arg.f = 6
+            final_op = activation_op
+            self.resolved_ops[activation_op.name] = 1
+
+        op_def.output.extend([output.name for output in final_op.outputs])
+        self.add_output_shape(final_op.outputs, op_def)
+        self.net_def.op.extend([op_def])
+
+    def check_conv_to_fc(self, op):
+        if self.device != 'neon' or op.type != "Conv2D":
+            return False
         filter_shape = get_input_tensor(op, 1).shape.as_list()
         input_shape = get_input_tensor(op, 0).shape.as_list()
-        if op_def.type == "Conv2D" and input_shape[1] == filter_shape[0] and \
-                input_shape[2] == filter_shape[1] and \
-                (op.get_attr('padding') == 'VALID' or filter_shape[0] == 1 and
-                 filter_shape[1] == 1):
-            print "convert op %s from CONV to FC" % op.name
-            op_def.type = 'FC'
-            self.reshape_tensor[get_input_tensor(op, 1).name] = \
-                [filter_shape[3],
-                 filter_shape[2] * filter_shape[1] * filter_shape[0], 1, 1]
+        return input_shape[1] == filter_shape[0] \
+            and input_shape[2] == filter_shape[1] \
+            and (op.get_attr('padding') == 'VALID' or filter_shape[0] == 1
+                 and filter_shape[1] == 1)
+
+    def convert_global_conv_to_fc(self, op):
+        op_def = mace_pb2.OperatorDef()
+        arg = op_def.arg.add()
+        arg.name = 'T'
+        arg.i = self.dt
+        op_def.name = op.name
+        op_def.type = 'FC'
+        self.transpose_filter_tensor[get_input_tensor(op, 1).name] = \
+            (3, 2, 0, 1)
+        filter_shape = get_input_tensor(op, 1).shape.as_list()
+        self.reshape_tensor[get_input_tensor(op, 1).name] = \
+            [filter_shape[3],
+             filter_shape[2] * filter_shape[1] * filter_shape[0], 1, 1]
+        op_def.input.extend(
+            [get_input_tensor(op, i).name for i in range(len(op.inputs))])
+
+        data_format_arg = op_def.arg.add()
+        data_format_arg.name = 'data_format'
+        data_format_arg.s = 'NCHW'
+        final_op = op
+        self.resolved_ops[op.name] = 1
 
         if len(self.tf_graph.get(op.name, [])) == 1 and \
-                self.tf_graph[op.name][0].type == 'BiasAdd':
+           self.tf_graph[op.name][0].type == 'BiasAdd':
             bias_add_op = self.tf_graph[op.name][0]
             if self.device == 'gpu':
                 output_name = self.add_buffer_to_image(
@@ -428,10 +476,8 @@ class TFConverter(object):
             self.resolved_ops[bias_add_op.name] = 1
 
         if len(self.tf_graph.get(final_op.name, [])) == 1 and \
-                self.tf_graph[final_op.name][0].type in activation_name_map:
+           self.tf_graph[final_op.name][0].type in activation_name_map:
             activation_op = self.tf_graph[final_op.name][0]
-            if op_def.type == "Conv2D":
-                op_def.type = "FusedConv2D"
             fused_act_arg = op_def.arg.add()
             fused_act_arg.name = 'activation'
             fused_act_arg.s = activation_name_map[activation_op.type]
@@ -985,6 +1031,8 @@ class TFConverter(object):
                 self.convert_reshape(op)
             elif self.is_atrous_conv2d(op):
                 self.convert_atrous_conv2d(op)
+            elif self.check_conv_to_fc(op):
+                self.convert_global_conv_to_fc(op)
             elif op.type == 'Conv2D' or op.type == 'DepthwiseConv2dNative':
                 if self.check_winograd_conv(op):
                     self.convert_winograd_conv(op)

mace/tools/validation/mace_run.cc

@@ -188,8 +188,8 @@ DEFINE_string(device, "OPENCL", "CPU/NEON/OPENCL/HEXAGON");
 DEFINE_int32(round, 1, "round");
 DEFINE_int32(restart_round, 1, "restart round");
 DEFINE_int32(malloc_check_cycle, -1, "malloc debug check cycle, -1 to disable");
-DEFINE_int32(gpu_perf_hint, 2, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
-DEFINE_int32(gpu_priority_hint, 1, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
+DEFINE_int32(gpu_perf_hint, 0, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
+DEFINE_int32(gpu_priority_hint, 0, "0:DEFAULT/1:LOW/2:NORMAL/3:HIGH");
 DEFINE_int32(omp_num_threads, 4, "num of openmp threads");
 DEFINE_int32(cpu_power_option,
              0,