K-ballo changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar.cct.lsu.edu | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | Buildbot: http://rostam.cct.lsu.edu/ | Log: http://irclog.cct.lsu.edu/
nanmiao has joined #ste||ar
hkaiser has joined #ste||ar
V|r has joined #ste||ar
Vir has quit [*.net *.split]
diehlpk_work has quit [Remote host closed the connection]
K-ballo has quit [Quit: K-ballo]
hkaiser has quit [Quit: bye]
nanmiao has quit [Quit: Connection closed]
<jedi18[m]>
Hi! Could someone help me understand what's wrong? I'm stuck at one place
<jedi18[m]>
It should return a minmax_element_result<local_iterator_type> right? Previously it was returning a std::pair<local_iterator_type,local_iterator_type> but since I've changed the implementation it now returns a minmax_element_result
<jedi18[m]>
So I guess the problem is with the min_max_result struct in result_types.hpp? Am I missing some operator overload required to make this work?
<jedi18[m]>
Nevermind, fixed it :D, I had to add the local_algorithm_result in dispatch for util::min_max_result
V|r is now known as Vir
<jedi18[m]>
How do I know for which file clang-format is failing? I can't find it when I run clang-format locally
<hkaiser>
all algorithms we have are based on three different partitioner implementations internally: foreach_partitioner, partitioner, and scan_partitioner
<ms[m]>
hkaiser: thanks for forwarding the email to the right people! :)
<hkaiser>
I think we have covered all that rely on the first, so any future work might involve touching not only the algorithm itself, but the (deep) implementation as well
<hkaiser>
I'd go with algorithms that are based on the partitioner, the scan_partitioner is more complex
<hkaiser>
ms[m]: np, I was the one suggesting to send it to Hal in the first place...
<jedi18[m]>
Oh ok thanks, how do I come to know which are based on the partitioner? By looking at the current implementations?
<ms[m]>
doesn't matter, seems to have ended up at a useful contact
<hkaiser>
jedi18[m]: look at the code ;-)
<jedi18[m]>
Ok thanks, will do!
<hkaiser>
ms[m]: do you have the wiki credentials?
<ms[m]>
hkaiser: nope
<ms[m]>
is this real...? I'm in in any case :P
<hkaiser>
yah
<ms[m]>
thanks!
<hkaiser>
ms[m]: that also gives you access to all internal meeting notes, documents, etc.
<ms[m]>
neat
<K-ballo>
make sure to strike those credentials from the public logs
<ms[m]>
it may be a bit too late for that... hkaiser can you do that with the irc logs?
<hkaiser>
nobody has the matching url ;-)
<hkaiser>
I can ask Alireza to delete it
<ms[m]>
security by obscurity!
shubham has joined #ste||ar
K-ballo has quit [Ping timeout: 265 seconds]
K-ballo has joined #ste||ar
jehelset has joined #ste||ar
<hkaiser>
ms[m]: the change fixed #4955, I pushed the fix
<ms[m]>
hkaiser: hmm, nice, thanks for checking!
<gonidelis[m]>
hey! I am trying to connect to my rostam account with a new machine. Could I just copy the keys under ~/.ssh from my old machine in order to skip two factor verification?
<ms[m]>
\
<ms[m]>
oops
<ms[m]>
gonidelis[m]: you can, but it's good practice to have separate keys on each machine
<gonidelis[m]>
huh
<gonidelis[m]>
so I will need two step verif at least for the first time
<jedi18[m]>
What do you recommend I work on next? Should I try adapting another algorithm or work in a different issue(to be accustomed to some other part the idea)?
<hkaiser>
jedi18[m]: I'd start thinking about the proposal and then select the next topic based on questions that arise from that
<jedi18[m]>
hkaiser: Ok I'll do that, can I share parts of my proposal from time to time and ask feedback on it?
nanmiao has joined #ste||ar
<hkaiser>
absolutely
<hkaiser>
you can always discuss things here
<ms[m]>
psa: daint is down for maintenance today, in case someone is wondering why cscs jenkins jobs are not running
<hkaiser>
ms[m]: ok, thanks
<hkaiser>
ms[m]: what I wanted to ask is how much effort it would be to set up a Phylanx jenkins on rostam
<ms[m]>
I think they're updating some modules as well, so the configs might need some changes once they're done
<hkaiser>
possibly using the docker images we already use on circleci
<ms[m]>
hkaiser: plain jenkins, not so much work
<ms[m]>
docker, no clue
<hkaiser>
ms[m]: I admit to hav enever looked at your jenkins setup
<ms[m]>
actually, I suppose docker would be fine as long as it's handled in the jenkins scripts
<hkaiser>
ms[m]: would we need a separate jenkins instance?
<ms[m]>
the config is quite horrible in general, but I think the setup we have right now is okay enough for running on slurm
<ms[m]>
so copy-pasting the hpx config should work just fine as a starting point
<ms[m]>
no, it can run on the same one
<hkaiser>
and connect to a different github project?
<ms[m]>
yeah, that's no problem
<hkaiser>
ok, thanks
<ms[m]>
it needs a separate project or whatever they call it in the jenkins instance, but it's definitely possible
<ms[m]>
most of the config can luckily be done in the repo rather than on the jenkins instance
<ms[m]>
gregor was also at some point interested in setting up a jenkins project, if someone from phylanx needs a short intro to jenkins I could do it for everyone interested in one go (not a very thorough intro, but enough to get things going)
<hkaiser>
ms[m]: I'd be interested
<ms[m]>
hkaiser: 👍️
<hkaiser>
btw: #5244: you're on the roll!
<ms[m]>
just let me know on what sort of schedule you'd like to set things up and I can try to prepare something
<hkaiser>
well, circleci has throttled their resources to a point that we can't use it for Phylanx anymore - so the sooner, the better
<ms[m]>
hkaiser: right, so let's try to do it sometime this week then
<ms[m]>
could you talk to alireza to set up jenkins accounts for anyone who wants/needs it (I guess you already have one)?
<zao>
Back when I tried building a test runner for HPX, I used Singularity to build and run tests in isolation I believe.
<zao>
Less intrusive than Docker, that's for sure :)
<ms[m]>
zao: what are the main benefits?
<zao>
ms[m]: No daemon needed, can run as an unprivileged user, has network isolation and other handy features.
<zao>
Can pull Docker containers from docker hub into their own image format.
<ms[m]>
zao: ah, yeah, that's definitely nicer
<zao>
It's Docker but usable in batch HPC, pretty much.
<hkaiser>
zao: yah, I think we have it available on rostam
<hkaiser>
ms[m]: I'll talk to Alireza - you said Greogor needs it as well?
<ms[m]>
he was at least interested earlier
<ms[m]>
he may already have an account though
<hkaiser>
ok
jehelset has quit [Remote host closed the connection]
shubham has quit [Quit: Connection closed for inactivity]
nanmiao has quit [Quit: Connection closed]
diehlpk_work has joined #ste||ar
diehlpk_work has quit [Remote host closed the connection]
<hkaiser>
nanmiao: I have limited things to one-argument functions as this is what shad does anyways (I think) - let me know if you need a variadic implementation
<nanmiao>
hkaiser Thank you!
nanmiao100 has joined #ste||ar
nanmiao100 has quit [Client Quit]
nanmiao has quit [Ping timeout: 240 seconds]
nanmiao has joined #ste||ar
nanmiao has quit [Quit: Connection closed]
nanmiao has joined #ste||ar
ct-clmsn has joined #ste||ar
ct-clmsn has quit [Client Quit]
nanmiao has quit [Quit: Connection closed]
bita has quit [Ping timeout: 246 seconds]
hkaiser has quit [Read error: Connection reset by peer]