On Sat, Sep 26, 2009 at 8:50 PM, Brian Wood <woodbrian77@gmail.com> wrote:

Daniel F. Savarese writes:

...

To the OP, I would advise you not take statements such as the above as  gospel without performing measurements yourself.  My experience[1] is directly contrary to the above.  I find boost::serialization execution time to be negligible (for binary archives) and entirely suitable for high performance networking applications even if that was not the creator's original objective (a credit to Robert Ramey's skill).

With respect for the library and Ramey, I'd like to point out that results obtained with the Boost Serialization library are frequently less efficient than those from the C++ Middleware Writer -- http://webEbenezer.net/comparison.html.

Additionally, the C++ Middleware Writer automates the job of writing serialization functions for structs/classes unless it is directed not to do so.

I actually do find Boost.Serialization to be very efficient, just the hobbling point for it is the registry it uses, that slows the whole thing down. I just found out about the Google events library thing a few days ago, I think that could actually make for a very efficient low-overhead network serializer, but for saving/loading and such, I still use Boost.Serialization due to its power. :)