11:46 <nikunj> hkaiser: I've updated all the sections, and added a few doubts in the paper. I'll explain all the sections on call, go through the paper if you have time. Everything is pushed to github.

11:48 <hkaiser> nikunj: ok, thanks

14:15 <parsa> simbergm: hey, we're setting up a meeting with our gsod participant. diehlpk suggests 4 or 5pm CEST (9 or 10am CT) this Friday. would that time work for you?

14:16 <simbergm> parsa: yep, that should work just fine

14:16 <simbergm> thanks!

14:17 <parsa> ?

14:21 <K-ballo> hkaiser: yt?

14:31 <bluescarni> hello, I have a few beginner-level questions about HPX. Is this a good place to ask?

14:43 <bluescarni> I'll try to briefly describe my use case, apologies in advance if this is not the correct place

14:46 <bluescarni> I have an astrophysical simulation code which is currently running on a single machine, parallelized via TBB. It uses fairly simple constructs, the usual tbb::parallel_for and a few task_groups here and there whenever independent parts of the simulation can be run in parallel

14:48 <bluescarni> there is in particular a part of the simulation loop which takes by far the most amount of time, and which consists of the independent simulation of many smaller disjoint systems which is thus trivially parallelisable via tbb::parallel_for.

14:50 <bluescarni> the time it takes to simulate each one of these smaller systems can vary wildly, thus the self-balancing behaviour of tbb::parallel_for is very helpful

14:51 <bluescarni> is there an HPX construct that would allow to transform this multi-threaded parallel for into a distributed one, keeping the load balancing behaviour?

14:52 <bluescarni> I took a look at HPX's parallel for_each(), but I could not understand whether that can run in a distributed setup as well

14:53 <K-ballo> it can run in distributed with one of the distributed containers, but it won't auto load balance

14:54 <bluescarni> thanks for the reply! So it would be a matter of laying out my data in a distributed vector and then running a parallal for_each() on the vector's iterators?

14:54 <bluescarni> I could probably come up with a way of doing some manual load balancing, even if not perfect that would be nevertheless a big improvement

14:56 <K-ballo> stay on this channel for a while and you'll get some questions leading to more precise suggestions

14:58 <bluescarni> ok! I will have to go in a while, but I'll read the log and reply as soon as possible. Thanks a lot for the help

15:01 <nikunj> hkaiser: yt?

15:01 <hkaiser> nikunj: trying to connect to skype, sec

15:01 <nikunj> sure sure

18:22 <diehlpk_work> simbergm, Are you available this Friday for a short skype meeting with the Season of Docs student?

19:05 <simbergm> diehlpk_work: yepyep

19:16 <diehlpk_work> simbergm, 4 or 5 would work for you?

19:17 <simbergm> diehlpk_work: yeah, both are fine

19:18 <diehlpk_work> Ok, we will add you via Skype