00:07
<
hkaiser >
nice, thanks
02:09
jaafar has quit [Ping timeout: 250 seconds]
02:39
hkaiser has quit [Quit: bye]
02:58
K-ballo has quit [Quit: K-ballo]
05:50
jaafar has joined #ste||ar
05:51
jaafar has quit [Client Quit]
05:52
jaafar has joined #ste||ar
06:29
jaafar has quit [Quit: Konversation terminated!]
07:43
nikunj1997 has joined #ste||ar
07:45
nikunj97 has quit [Ping timeout: 258 seconds]
08:06
hkaiser has joined #ste||ar
08:23
david_pfander has joined #ste||ar
08:54
ste||ar-github has joined #ste||ar
08:54
ste||ar-github has left #ste||ar [#ste||ar]
09:26
<
zao >
(on GCC Ubuntu x64, that is)
09:42
ste||ar-github has joined #ste||ar
09:42
ste||ar-github has left #ste||ar [#ste||ar]
09:46
ste||ar-github has joined #ste||ar
09:46
ste||ar-github has left #ste||ar [#ste||ar]
09:47
ste||ar-github has joined #ste||ar
09:47
<
ste||ar-github >
hpx/master e426a5e Hartmut Kaiser: Merge pull request #3607 from STEllAR-GROUP/adding_component_paths...
09:47
ste||ar-github has left #ste||ar [#ste||ar]
10:01
quaz0r has quit [Ping timeout: 250 seconds]
10:03
quaz0r has joined #ste||ar
10:16
ste||ar-github has joined #ste||ar
10:16
<
ste||ar-github >
hpx/gitlab_runner c6cdc8f Thomas Heller: WIP: Adding support for gitlab runners
10:16
ste||ar-github has left #ste||ar [#ste||ar]
10:19
<
heller_ >
almost closing in to the gitlab runner stuff...
10:20
ste||ar-github has joined #ste||ar
10:20
<
ste||ar-github >
hpx/gitlab_runner c6be94a Thomas Heller: WIP: Adding support for gitlab runners
10:20
ste||ar-github has left #ste||ar [#ste||ar]
10:52
<
heller_ >
a full build from checkout to failure :P
10:52
<
heller_ >
in 23 minutes
11:06
ste||ar-github has joined #ste||ar
11:06
<
ste||ar-github >
hpx/gh-pages 35bd999 StellarBot: Updating Sphinx docs
11:06
ste||ar-github has left #ste||ar [#ste||ar]
11:35
K-ballo has joined #ste||ar
12:03
nikunj97 has joined #ste||ar
12:06
nikunj1997 has quit [Ping timeout: 250 seconds]
13:24
<
simbergm >
heller_: cool
13:24
<
simbergm >
so that's running on that whatever-it-was-called-thing with free slow instances and occasional access to bigger instances
13:25
<
heller_ >
simbergm: the cool thing here is, that gitlab runners are in addition able to run over ssh
13:25
<
heller_ >
yeah, during tests, the bigger instances were almost always available
13:27
<
simbergm >
what do you mean by run over ssh?
13:28
<
heller_ >
that we can target any machine we have access to
13:29
<
simbergm >
ah, as in not over http or something else annoying like that
13:32
<
heller_ >
so the runners on the cloud provider use docker-machine
13:32
<
heller_ >
to autoscale
13:33
<
heller_ >
also: I set up a small test repository to play with the setup
13:33
<
heller_ >
so, in theory, this would allow us, to add new build configurations with PRs, for example
13:33
<
heller_ >
and having that nicely integrated
13:34
<
heller_ >
and less messy ;)
13:34
<
heller_ >
this could replace circle-ci, buildbot and pycicle
13:35
<
heller_ >
all integrated into one tool, which we don't have to maintain, we'd only have to maintain the infrastructure
13:37
<
heller_ >
packet also offers servers with V100s and ARM64...
13:51
<
heller_ >
so, what do you think?
13:54
<
simbergm >
heller_: I think it sounds wonderful
13:56
<
simbergm >
do we know how much compute time we could get realistically? to build 20 open PRs with a few different configurations :/
13:56
<
simbergm >
although I guess your point is that we can use rostam as well ;)
13:56
<
simbergm >
does gitlab have an option to only run ci on demand?
13:57
<
zao >
So how does this fit into my vapourware test soaker? ;)
14:00
<
heller_ >
zao: you can define cron jobs to run pipelines ;)
14:00
<
heller_ >
simbergm: weee!
14:01
<
heller_ >
simbergm: regarding compute time: Yes, we can use rostam and emmy and piz daint and what not ;)
14:01
<
heller_ >
simbergm: let's start small and see how we can expand...
14:01
<
heller_ >
simbergm: with on demand, you mean not automatically per PR, or when you have a branch, you want to check if it works?
14:02
<
simbergm >
heller_: goody, I have the same concerns as hkaiser about community being on github etc, but you were planning on having a mirror to start with? I have to say I do quite like gitlab
14:03
<
heller_ >
I like gitlab very much as well
14:03
<
heller_ >
simbergm: currently, it is setup to mirror github
14:03
<
simbergm >
yeah, essentially only build once someone tells it to build to avoid unnecessary builds (e.g. WIP PRs)
14:03
<
heller_ >
yes, you can skip WIP PRs
14:03
<
heller_ >
you can even tag them accordingly in the gitlab UI ;)
14:04
<
simbergm >
or only build one configuration to start with and then build all once it's clean as a last check
14:04
<
simbergm >
for example
14:04
<
heller_ >
simbergm: I think it would already be quite an improvement if we just replace circle-ci as a first step
14:05
<
simbergm >
I guess it can't report to github? if it can I'm sold
14:10
<
heller_ >
it already does ;)
14:10
<
heller_ >
look at the PR
14:10
<
heller_ >
"ci/gitlab/gitlab_runner — Pipeline failed on GitLab Details"
14:10
<
heller_ >
first line for the checks
14:12
<
simbergm >
oh... well then :)
14:13
<
heller_ >
"just" need to finalize the yml...
14:14
<
heller_ >
btw: gitlab even comes with its own docker registry ;)
14:23
aserio has joined #ste||ar
14:31
<
hkaiser >
heller_, simbergm: I'm still not fond of moving away the main user-facing repo from github
14:31
<
hkaiser >
if we can do everything through the (automatic) mirror, let's do it
14:31
<
heller_ >
hkaiser: I didn't suggest to move everything away
14:31
<
heller_ >
just the CI part
14:31
<
hkaiser >
fine by me
14:35
<
heller_ >
hkaiser: however, as far as I can tell, gitlab is gaining more and more traction
14:35
<
heller_ >
but I couldn't care less where the repo is actually hosted...
14:37
<
heller_ >
however, I do think that gitlab offers a better UI than github
14:38
hkaiser has quit [Quit: bye]
14:40
<
heller_ >
simbergm: can we get rid of the rostam docs builder?
14:48
jaafar has joined #ste||ar
14:50
dmarce1 has quit [Ping timeout: 250 seconds]
14:53
<
simbergm >
heller_ definitely! Can you do it? Can I do it?
14:53
<
heller_ >
simbergm: I can do it, just wanted to check
14:58
<
heller_ >
sorry for the disruption ... didn't see that there is still a build ongoing :(
15:04
dmarce1 has joined #ste||ar
15:05
hkaiser has joined #ste||ar
15:45
akheir has joined #ste||ar
15:48
<
diehlpk >
jbjnr_, Are you interested to become a co-organizer for GSoC this year again?
15:48
aserio has quit [Ping timeout: 250 seconds]
16:18
aserio has joined #ste||ar
17:03
david_pfander has quit [Ping timeout: 244 seconds]
18:01
aserio has quit [Ping timeout: 258 seconds]
19:12
aserio has joined #ste||ar
19:50
<
diehlpk >
heller_, Do you still have contact to Arne?
19:50
<
diehlpk >
We should ask him if he is still interested in GSoC.
20:33
bibek has quit [Quit: Konversation terminated!]
20:35
bibek has joined #ste||ar
21:17
hkaiser has quit [Quit: bye]
22:19
hkaiser has joined #ste||ar
22:35
akheir has quit [Remote host closed the connection]
22:40
aserio has quit [Quit: aserio]
23:37
maxwellr96 has joined #ste||ar