Minor improvement of inference profiling

* Start inference profiling at the actual beginning
* Add profiling events for EnqueueInputs and EnqueueOutputs
* Add profiling event for working memory allocation
* Refactor Execute body to remove code duplication
* forward arguments to constructors rather than copy

Change-Id: Iacab85f0a02e88e2423885f86f97e4dba4037319
Signed-off-by: Derek Lamberti <derek.lamberti@arm.com>

1 files changed, 1 insertions, 0 deletions

diff --git a/src/armnn/Runtime.cpp b/src/armnn/Runtime.cpp
index 5692494836..28e2df22ab 100644
--- a/src/armnn/Runtime.cpp
+++ b/src/armnn/Runtime.cpp
@@ -308,6 +308,7 @@ Status Runtime::EnqueueWorkload(NetworkId networkId,
                                 const InputTensors& inputTensors,
                                 const OutputTensors& outputTensors)
 {
+    ARMNN_SCOPED_PROFILING_EVENT(Compute::Undefined, "EnqueueWorkload");
     LoadedNetwork* loadedNetwork = GetLoadedNetworkPtr(networkId);
 
     static thread_local NetworkId lastId = networkId;