07:38 <jedi18[m]> There seem to be no segmented algorithm tests for generate? Shall I add one?

11:24 <rachitt_shah[m]> Maintainers, I see our docs are made in sphinx and readthedocs.

11:24 <rachitt_shah[m]> There are some open issues, however, they are from 2020, just wanted to check if I should proceed working on them.

12:08 <hkaiser> rachitt_shah[m]: yes, we're intersted in moving the docs forward

12:08 <hkaiser> jedi18[m]: yes, if there is no test, then adding one (or two) is a good idea

12:09 <rachitt_shah[m]> Got it!

12:09 <rachitt_shah[m]> I'm thinking of starting it off from here - https://github.com/STEllAR-GROUP/hpx/issues/4698, I would update you on the issue.

12:10 <hkaiser> sounds good

12:10 <hkaiser> rachitt_shah[m]: are you able to build the docs ?

12:10 <rachitt_shah[m]> Sure. Will ping here if I need a hand, would send a PR by Thursday/Friday.

12:10 <rachitt_shah[m]> Yes hkaiser

12:10 <hkaiser> ok

12:10 <hkaiser> good

12:11 <rachitt_shah[m]> Thank you! Will ping here if I need guidance, since I've just started exploring the organization and our project.

12:12 <hkaiser> sure

12:12 <rachitt_shah[m]> :D

12:12 <jedi18[m]> hkaiser: Oh ok sure! They seem to be missing the sentinel tests too so I'll add them too

12:13 <hkaiser> jedi18[m]: I don't think we support sentinels for the segmented algorithms

12:14 <mdiers[m]> hkaiser: Hi, https://github.com/STEllAR-GROUP/hpx/pull/5117 is updated and I would be happy if it would continue there now.

12:14 <jedi18[m]> No I don't mean for the segmented algortihms, generate and fill have sentinel overloads in the container algorithms but are missing sentinel tests, so I'll add those

12:15 <hkaiser> mdiers[m]: thanks for the test case, I'll see what I can do

12:15 <hkaiser> jedi18[m]: ahh, perfect

12:16 <jedi18[m]> I'm thinking of finishing up the disentangling of all the remaining segmented algorithms so I can then work on more interesting stuff

12:17 <hkaiser> jedi18[m]: that would be great, much appreciated

12:17 <hkaiser> I see we've got you hooked ;-)

12:20 <jedi18[m]> Thanks :D

12:20 <mdiers[m]> hkaiser: Thanks, I will also check it on an AWS cluster.

12:23 <srinivasyadav227> i am getting this error on windows "FindThreads only works if either C or CXX language is enabled", i installed from github

12:24 <hkaiser> srinivasyadav227: you'd need to be more specific - best is to create a ticket explaining what you tried, what you expected, and what actually happened

12:25 <hkaiser> then come back here to discuss

12:25 <srinivasyadav227> hkaiser: ok sure

12:32 <K-ballo> curious, both C and CXX are enabled by default

13:44 <gnikunj[m]> gonidelis[m]: yt?

13:49 <gnikunj[m]> hkaiser yt?

13:51 <gonidelis[m]> gnikunj[m]: here now

13:51 <gnikunj[m]> gonidelis[m] hkaiser any clue which node I should use for gpu benchmarking on rostam?

13:52 <gonidelis[m]> gnikunj[m]: haven't done any

13:52 <gonidelis[m]> so i am not quite sure

13:52 <gnikunj[m]> I see :/

13:53 <gnikunj[m]> I'll email Ali

13:53 <gonidelis[m]> do u run tests for the talk?

13:53 <gnikunj[m]> I need to run it for the talk and my meeting tomorrow

13:54 <gonidelis[m]> i see

14:28 <gonidelis[m]> what's this for ? https://github.com/gonidelis/hpx/blob/8f608a97ae4f765cb8fdd166b478097267d65c4b/libs/parallelism/algorithms/include/hpx/parallel/util/foreach_partitioner.hpp#L70

15:52 <gonidelis[m]> I am trying to run a simple `hpx::transform(par, ...)` and I get a `core dumped`. When I change `par` to `seq` it runs fine. It seems like our parallel transform does not work

15:53 <gonidelis[m]> I know who to blame for that...

15:54 <hkaiser> gnikunj[m]: https://github.com/STEllAR-GROUP/hpx/wiki/Running-HPX-on-Rostam

15:54 <gnikunj[m]> thanks!

15:54 <hkaiser> gonidelis[m]: show me the code

15:55 <gonidelis[m]> hkaiser: sec

15:56 <K-ballo> note that something working in seq does not imply it will work with par

15:57 <gonidelis[m]> https://gist.github.com/gonidelis/6d745e5c00b364205a17664de16d2f02

15:57 <hkaiser> K-ballo: it shouldn't segfault, though

15:57 <gonidelis[m]> K-ballo: sure, the call is trivial

15:57 <gonidelis[m]> exactly ^^

15:57 <K-ballo> segfault is a valid representation of UB

15:57 <hkaiser> lol

15:57 <gonidelis[m]> ub?

15:57 <K-ballo> if you promise the algorithms guarantees that don't hold

15:57 <K-ballo> you get undefined behavior

15:57 <hkaiser> as is launching rockets...

15:57 <gnikunj[m]> lol

15:57 <gonidelis[m]> und beh

15:57 <gonidelis[m]> ok

15:58 <gonidelis[m]> lol

15:58 <hkaiser> gonidelis[m]: give me the full code and the command line options

15:58 <gonidelis[m]> oh ok

15:58 <hkaiser> gonidelis[m]: otherwise I can't reproduce the issue

16:00 <gonidelis[m]> hkaiser: check same link

16:00 <gonidelis[m]> https://gist.github.com/gonidelis/6d745e5c00b364205a17664de16d2f02

16:01 <hkaiser> gonidelis[m]: ok, will have a look

18:37 <gnikunj[m]> building things on a gpu is a pain

19:34 <gnikunj[m]> can someone please help me with installing HPX (with CUDA) on rostam? I've been trying for 2h now ;_;

19:34 <gnikunj[m]> cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$HOME/install/hpx -DHPX_WITH_MALLOC=tcmalloc -DHPX_WITH_CUDA=ON -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_C_COMPILER=clang -DHPX_WITH_CUDA_CLANG=ON -DHPX_CUDA_CLANG_FLAGS=--cuda-gpu-arch=sm_70 -DCMAKE_CXX_STANDARD=17 -DHPX_USE_CMAKE_CXX_STANDARD=ON ..

19:34 <gnikunj[m]> I am using above command ^^

19:39 <hkaiser> gnikunj[m]: https://github.com/STEllAR-GROUP/hpx/blob/master/.jenkins/lsu/env-gcc-9-cuda-11.sh

19:39 <hkaiser> the networking=off is orthogonal

19:39 <gnikunj[m]> wait sm_35 lol

19:39 <gnikunj[m]> that's ancient

19:40 <hkaiser> this is do make sure the tests can run on older architectures (K40 etc.)

19:40 <gnikunj[m]> are we using it on the nodes with V100?

19:40 <hkaiser> newer archs should work as well

19:40 <gnikunj[m]> got it. Let me try.

19:53 <gnikunj[m]> hkaiser: it worked \o/