When using protobuf , there comes the performance bottleneck.
Then I found the hot point is focus on
GeneratedMessageReflection::ListFields (which costs 30%+ CPU idle) by
gprofiler, which indicates that there is too much "new" and "delete" .
------------------------API of ListFields---------------------------------
void
GeneratedMessageReflection::ListFields(
const Message&
message,
vector<const FieldDescriptor*>* output) const;
---------------------------------------------------------------------------------
As the function called, the second arg : "output" is always a empty vector.
So it will cause new/delete for many times in case of plenty fields
So as title mentioned, Why don't we use deque or other container to avoid
too much new/delte, instead of using vector.
--
You received this message because you are subscribed to the Google Groups
"Protocol Buffers" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/protobuf/-/AhsFq2N6IxEJ.
To post to this group, send email to protobuf@googlegroups.com.
To unsubscribe from this group, send email to
protobuf+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/protobuf?hl=en.
