Nicolas Williams wrote: > But it's also important to consider whether the common uses are in MT or > single-threaded programs. Why even bother measuring the cost for > single-threaded programs if they are not the target, or if no > consequential single-threaded programs exist that might use this > library?
There is not such thing as a non MT program post Solaris 10. All processes potentially have more than one thread. Please do not attempt to regress the huge amount of work that went into unifying the process model for Solaris 10 by providing confusing by shipping multiple libraries. This is not about measurements it is about the architecture of reference and for Solaris 10 onwards it is an always MT world and as a resut of that libraries are now free to create threads as part of their implementation even if those threads aren't exposed to the original application. For this to be possible there can be only one acceptable version of a library and that is the MT one - even for an application that thinks it is single threaded. -- Darren J Moffat
