<simbergm>
hkaiser: do I take your ?️ on the CUDA PR as an approval? :)
<simbergm>
I think that's the last PR I'd like to have in for the first rc
<hkaiser>
:D yes
<hkaiser>
how about the cmake one (Debug/Release)?
<hkaiser>
that one is for one of your people
mdiers_ has quit [Ping timeout: 265 seconds]
<hkaiser>
simbergm: ^^
<simbergm>
hkaiser: if I figure it out quickly ;) it's more of a convenience thing though so it can wait for the next RC as well
<simbergm>
anything else you'd like to have in the rc still?
<hkaiser>
the intel fix?
<hkaiser>
not too urgent either, however
<simbergm>
yep... I think we can wait with that one and see if instaM has time to try it out first
mdiers_ has joined #ste||ar
<simbergm>
hkaiser: do you mind me pushing the modified check to the build type branch?
<hkaiser>
simbergm: fine by me
<simbergm>
thanks
<simbergm>
hkaiser: it seems to fail because HPX_WITH_MALLOC is set earlier as a non-cache variable
<hkaiser>
uhh
<simbergm>
the question is: do you expect all of our cache variables to be cache variables for dependent projects as well?
<simbergm>
do you expect to be able to change them?
<hkaiser>
good point
<hkaiser>
you probably will not have to change them
<simbergm>
we could explicitly forward them as cache variables if that's what you need, but I'm not sure if they should be
<hkaiser>
no, I don't think this is needed
<simbergm>
things will probably break generally if users change them...
<hkaiser>
right
<simbergm>
but I see your cmake gui shows all the variables even if they're not cache variables
<hkaiser>
yes
<simbergm>
was the screenshot from a dependent project or hpx itself?
<hkaiser>
it does not make a difference between cache and no-cache
<hkaiser>
that was hpx itself
<simbergm>
ccmake apparently only shows cache variables
<simbergm>
ah
<hkaiser>
let me check
<hkaiser>
simbergm: yah, it shows cached variables only
<hkaiser>
simbergm: let's clean all of this HPX_WITH_MALLOC business for dependent projects, then
<simbergm>
I think HPX_WITH_MALLOC was a special case because it was separately set as a cache variable in hpx_setup_allocator, but it doesn't need to be set there since it's already forwarded
<simbergm>
ugh, HPX_SetupAllocator is included twice even
<simbergm>
I'll do some cleanup
<hkaiser>
thanks
<simbergm>
HPX_DISABLE_CUSTOM_ALLOCATOR? is that a supported thing? disabling the allocator for dependent projects?
<hkaiser>
I can remove this 'flyby' from that branch
<simbergm>
I doubt that works
<hkaiser>
right
<hkaiser>
probably not needed either
<simbergm>
either way... I can create a new branch cleaning up the allocator options
<simbergm>
actually, I can remove it from that PR, I'll add the new check while I do it (didn't get to that point yet...)
<hkaiser>
ok, and I'll remove the change from the other PR
<hkaiser>
ok, fine - works for me ;-)
<simbergm>
hkaiser: ^ unless you already did it ;)
<hkaiser>
I have not touched it
<simbergm>
?️
<simbergm>
hkaiser: pushed
<hkaiser>
simbergm: thanks!
<hkaiser>
simbergm: LGTM
<zao>
hkaiser: I might be able to test at home with older Intel (2018.3.222). Would that help anything, or be largely superfluous?
<zao>
We've got even older ones at work, but I don't have all the deps I'd need compiled up there.
<hkaiser>
zao: don't worry about it too much - people interested in using Intel can always use the latest version
<hkaiser>
earlier versions were broken in different ways, afair, so let's not wake sleeping dogs ;-)
<zao>
Alright :)
<hkaiser>
zao: and thanks again for helping with this
<hkaiser>
zao: have we ever sent you a t-shirt, btw?
<zao>
Yup, it's in my rotation :)
<hkaiser>
ok, good
aserio has joined #ste||ar
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar
aserio has quit [Ping timeout: 250 seconds]
aserio has joined #ste||ar
aserio has quit [Ping timeout: 250 seconds]
aserio has joined #ste||ar
aserio has quit [Ping timeout: 276 seconds]
aserio has joined #ste||ar
hkaiser has quit [Read error: Connection reset by peer]
Coldblackice has joined #ste||ar
Coldblackice has quit [Ping timeout: 276 seconds]
Coldblackice has joined #ste||ar
Coldblackice_ has joined #ste||ar
Coldblackice has quit [Ping timeout: 240 seconds]
Coldblackice_ is now known as Coldblackice
hkaiser has joined #ste||ar
hkaiser has quit [Quit: bye]
khuck has joined #ste||ar
<khuck>
aserio: are we meeting today?
<khuck>
Chris and I are on the line... was there an assumption that because we just met in person we don't need to meet? :)
<khuck>
interestingly. the LSU CCT room connected automagically
<khuck>
but the room is empty.
<diehlpk_work>
Do we know if we 1d stencil example for distributed cases is working?
<diehlpk_work>
With 3 and 4 nodes they just run for ever until slurm kills the jobs after 25 minutes
aserio has quit [Ping timeout: 245 seconds]
<diehlpk_work>
Using 5 nodes works again
<diehlpk_work>
6 does not
aserio has joined #ste||ar
aserio has quit [Quit: aserio]
aserio has joined #ste||ar
hkaiser has joined #ste||ar
aserio has quit [Ping timeout: 276 seconds]
aserio has joined #ste||ar
aserio has quit [Quit: aserio]
jbjnr has quit [Read error: Connection reset by peer]