00:01 <hkaiser> K-ballo: I think v9

08:50 <gonidelis[m]> Any ideas why the compilation stops at `performance_counter_set.cpp: In member function ‘std::vector<hpx::performance_counters::counter_value> hpx::performance_counters::performance_counter_set::get_counter_values(hpx::launch::sync_policy, bool, hpx::error_code&) const’:

08:50 <gonidelis[m]> ` ?? I am up to date with upstream master

12:13 <K-ballo> gonidelis[m]: is there an error message with that?

13:35 <Yorlik> Is it possible to keep this line of code:

13:35 <Yorlik> static inline size_t hpx_worker_thread_count = hpx::get_num_worker_threads( );

13:35 <Yorlik> Or would I have to give up the member being static?

13:35 <Yorlik> I get an error the HPX runtime isn't initialized yet

13:35 <hkaiser> see

13:36 <hkaiser> make a function local static

13:36 <hkaiser> make it*

13:36 <hkaiser> if you need to be one in the first place

13:37 <hkaiser> but this might introduce other problems...

13:37 <hkaiser> Yorlik: so I'd suggest to just call hpx::get_num_worker_threads() whenever you need it

13:38 <Yorlik> I'm working on optimizing small object allocations with a custom per thread pooled allocator for messages and their parameters.

13:38 <Yorlik> But with an option to delete from another thread

13:39 <Yorlik> But I can work around this.

13:59 <nikunj97> hkaiser, meeting?

13:59 <hkaiser> nikunj97: sorry, tied up with all kind of chores right now

14:00 <nikunj97> ohh ok. Let's do it some other time then.

14:00 <nikunj97> Monday 9am?

14:00 <hkaiser> Yorlik: if you can prove that hpx::get_num_worker_threads( ) is causing perf issues we'll find another way. until then, just use it

14:00 <Yorlik> It's just setup code. Not an issue

14:01 <hkaiser> nikunj97: 10am would be better

14:01 <hkaiser> Yorlik: ok

14:01 <nikunj97> hkaiser, Monday 10am. Confirmed.

14:02 <hkaiser> nikunj97: https://lsu.zoom.us/j/3340410194

14:03 <nikunj97> hkaiser, saved. See you on Monday!

14:03 <hkaiser> thanks

17:03 <K-ballo> are these known to fail

17:03 <K-ballo> 1181 - tests.unit.modules.segmented_algorithms.distributed.tcp.partitioned_vector_adjacent_find1 (Timeout)

17:03 <K-ballo> 1202 - tests.unit.modules.segmented_algorithms.distributed.tcp.partitioned_vector_transform_reduce_binary2 (Timeout)

17:03 <K-ballo> 1190 - tests.unit.modules.segmented_algorithms.distributed.tcp.partitioned_vector_iter (Timeout)

17:08 <hkaiser> yes

17:08 <hkaiser> there are more

17:08 <hkaiser> those are on my list to investigate

17:08 <hkaiser> the migration tests fail nnow and then as well

20:45 <hkaiser_> gonidelis[m]: where do you see that?

20:48 <gonidelis[m]> I type `make -j4 tests.unit.modules.iterator_support`

20:49 <gonidelis[m]> and the compilation stops at `[ 70%] Built target tests.unit.modules.iterator_support.range`

20:49 <gonidelis[m]> hkaiser_:

21:02 <hkaiser_> gonidelis[m]: starnge

21:03 <hkaiser_> strange

21:07 <gonidelis[m]> ok I 'll try to fix it tomorrow with a little help from rori maybe

21:07 <gonidelis[m]> maybe I messed up sth in my build

21:07 <gonidelis[m]> was just asking if upstream master compiles fine for all of you ;)

21:09 <hkaiser_> sure it does ;-)

21:09 <gonidelis[m]> ok great...i 'll check my problem then