00:32 <zao> I wonder if nik ever read my message about tls callbacks, no acknowledge

00:32 <zao> Guess I should not have bothered.

00:32 <hkaiser> what nik?

00:33 <zao> Unj

00:33 <zao> Can’t complete their nickname as they have quit in the meantime.

00:33 <hkaiser> k

00:34 <hkaiser> I'd poke him again, he was around earlier

00:34 <zao> Guess they don’t read mentions when they’ve been logged into irc

00:35 <zao> I’m out travelling anyway, just felt I had a lead they might want to look into, even if it was for a negative result.

00:40 <hkaiser> zao: I'll give him a headsup

09:05 <github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://git.io/vhWj6

09:05 <github> hpx/gh-pages a164f08 StellarBot: Updating docs

13:06 <hkaiser> jbjnr_: yt?

15:25 <diehlpk> hkaiser, Is the storm project still active?

15:25 <diehlpk> We were thinking to interview HPX and Storm as the application using HPX

15:36 <jakub_golinowski> M-ms, yt?

15:39 <hkaiser> diehlpk: yah, storm is still ongoing

15:40 <M-ms> jakub_golinowski: yep, here

15:41 <jakub_golinowski> M-ms, you were right, in some places opencv is splitting the range. This was the case in the hpx_load_image example where image is converted to grey_scale.

15:42 <jakub_golinowski> the split is decided by the float nstripes

15:43 <jakub_golinowski> For now I am not sure why this is the case so I just ignore the split introduced by opencv and access the wholeRange (which is stored in context object) and with such a backe hpx_load_image is working properly wihtout any changes for both sequeantial, TBB and HPX

15:44 <jakub_golinowski> I have committed the new version of opencv to my fork of opencv

15:45 <M-ms> Ok, nice

15:45 <M-ms> But what does that mean in terms of parallelism?

15:45 <jakub_golinowski> Just want to make sure: the call is tomorrow at 4pm Zurich time?

15:46 <M-ms> Thanks, I'll try it out tomorrow (my own laptop decided to give up, back at work tomorrow)

15:46 <M-ms> Yep, I'm flexible though, preferably within normal working hours

15:46 <M-ms> Especially as Patrick is not joining

15:48 <diehlpk> M-ms, jakub_golinowski In two weeks I am back from my leave and we can try to find a new time

15:49 <jakub_golinowski> M-ms, 4pm would be good, I have a private room booked in the library for that time tomorrow so it should be quiet

15:50 <M-ms> Did you have a look at what kind of ranges you get for each (sub) for loop (so what's e.g. in the lambda in the mandelbrot example)? Do you get some kind of chunking with the tbb backend? And does for the hpx backend there is basically no chunking (at the moment)?

15:50 <M-ms> Good, let's stick with 4 pm then

15:51 <M-ms> diehlpk: ok, no problem

15:51 <jakub_golinowski> M-ms, so for now I was looking at the ranges for the hpx_load_image example. The log of my findings is here: https://docs.google.com/document/d/1VtI8jfzx92YiRW-o_D-_S9oNqvpw2I8udsyQyaGTFEc/edit#

16:01 <M-ms> Thanks, that looks good

16:02 <M-ms> I also think now would be a good time to ask on the opencv channel about that nstripes parameter as it seems to be quite fundamental to how the backends are used

16:05 <jakub_golinowski> M-ms, I agree - I will compose a message to opencv today

16:08 <M-ms> jakub_golinowski: perfect (please ping me as well over on the opencv channel)

16:08 <jakub_golinowski> ok

17:57 <jakub_golinowski> M-ms, yt?

18:11 <M-ms> jakub_golinowski: here now

18:11 <jakub_golinowski> I wrote down the message to opencv

18:13 <jakub_golinowski> M-ms, I added you to gdoc

18:20 <M-ms> jakub_golinowski: looks good in general, but where were you planning on sending that? I got the impression there was no (active) mailing list

18:21 <jakub_golinowski> M-ms, this is what I wanted to consult with you

18:22 <jakub_golinowski> I can just copy paste it to the irc but am not sure if this is a good idea

18:26 <M-ms> Hmm, no, I would try to keep it a bit shorter. There might not even be someone online at the moment who can help. So I would say introduce yourself shortly and ask one question at a time, see if someone has an answer. Then you might want to open a WIP pull request to get early feedback from the opencv side (easier for them to review that way).

18:29 <jakub_golinowski> ok

18:30 <jakub_golinowski> what is WIP pull request?

18:31 <M-ms> Ah, WIP is work in progress

18:32 <M-ms> Just used to mark that it's not ready to be merged but still keeping it visible for others to see the progress

18:33 <jakub_golinowski> ok

23:15 <parsa> hkaiser: microsoft's acquired github… is that how things are supposed to work?

23:18 <K-ballo> hah, so that's what people were talking about gitlab

23:56 <hkaiser> parsa: github or gitlab?

23:56 <parsa> github

23:56 <hkaiser> ok

23:57 <parsa> https://www.bloomberg.com/news/articles/2018-06-03/microsoft-is-said-to-have-agreed-to-acquire-coding-site-github

23:59 <hkaiser> let's see how that pans out ...