This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] libstdc++ atomicity, x86 agnostic version

Benjamin Kosnik wrote:

Can't we follow Jakub's advice?


It's something to look into. However, if you see this code being inlined
in std::string in performance critical regions, I'd appreciate knowing
exactly where.

A testcase would still be nice....


Hi again.

What's currently present in the performance testsuite tests the exponential growth policy: basically there are *no* calls of _M_refcopy and _M_dispose.

It's needed something completely different, but basically *every real world* test on a multi processor machine or in a MT environment can only be negatively affected by not inlining the operations done at construction and destruction time.

I suggest, pre-emptively to work out Jakub's proposal. I volunteer to do that (hopefully with some help from him), I can allocate time to that during the next few weeks.

Paolo.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]