Re: [RFD] optimized versions of openssl
Michael Stone writes:
> On Wed, Sep 04, 2002 at 03:35:58PM +0200, Marcelo E. Magallon wrote:
> > The shared library is 179 kB. Why don't you just provide the optimized
> > versions in the same package? Are the any stability/correctness issues
>
> Now for the real overachiever, what would be really cool is if you
> hacked openssl to do *runtime* detection of which optimizations to use.
I suspect it would be "better" to do install-time selection; the
preferred way to choose is to benchmark the alternatives on a
realistic task. That isn't something most people would want to do when
they load the library, since it takes a while to do, and on a given
CPU, the results are unlikely to change over time.
Semi-static choice also makes it easier to diagnose which variant is
being used when there are problems and is probably easier to do than
run-time selection. (The overachiever variant of this is to set up
dpkg alternatives at install-time so that the system admin can use the
existing framework to override the automatic selection.)
Michael
Reply to: