Re: exception performance in modern gcc

[Ian Lance Taylor <iant at google dot com>]

Brendan Miller <catphive@catphive.net> writes:

> I understand that exceptions aren't meant to be thrown often, so it
> doesn't usually matter if exceptions are a little slow, but I'd like
> to understand the performance a little better. If anyone knows where I
> can find notes on how the exception propagation mechanism works at a
> fairly low level that would also be helpful. I'm trying to get a
> picture of where the costs come from, aside from calling destructors
> as the stack unwinds.

The exception unwinding mechanism used by gcc is described at
http://www.codesourcery.com/public/cxx-abi/abi-eh.html .  Ignore the
word "Itanium."

Ian