https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105989
Iain Sandoe <iains at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Last reconfirmed| |2024-07-15
Status|UNCONFIRMED |NEW
--- Comment #12 from Iain Sandoe <iains at gcc dot gnu.org> ---
(In reply to Michal Jankovič from comment #11)
> (In reply to Iain Sandoe from comment #10)
> > actually I thought I explained the issue in email to Michal - that we need
> > to make some changes for correctness to these mechanisms and therefore that
> > we should defer the optimisation until those are done.
>
> Ok, I must have missed that or misremembered;
NP - and it's been unfortunate that it's taken some time to get some support to
revisit GCC coroutines.
> I remember you recommending I
> get some real-world stats on the memory savings to help make this higher
> priority, but I couldn't find the time.
Yeah I still think we need to do this - my plan was to try and instrument and
then see how it applies to folly's test cases; the general issue being that
compiler testcases and some of the smaller coroutines libraries would likely
not really show as much advantage since they do not have particularly complex
scoping or large frames.
