06:45 <ste||ar-github> [hpx] biddisco pushed 1 new commit to future_overhead_timing: https://github.com/STEllAR-GROUP/hpx/commit/94b4dae20e48797c2ecdb0ab5f0e0e971a6d1736

06:45 <ste||ar-github> hpx/future_overhead_timing 94b4dae John Biddiscombe: Add print utility for DartMeasurentTime to lightweight_test...

09:05 <ste||ar-github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://github.com/STEllAR-GROUP/hpx/commit/0449e3719dc10f51ff00d3a14ef65508a27a8840

09:05 <ste||ar-github> hpx/gh-pages 0449e37 StellarBot: Updating docs

12:52 <ste||ar-github> [hpx] biddisco pushed 3 new commits to future_overhead_timing: https://github.com/STEllAR-GROUP/hpx/compare/94b4dae20e48...21c3ec55d8a6

12:52 <ste||ar-github> hpx/future_overhead_timing 9a588ac John Biddiscombe: Add tests.performance.local to unit testing

12:52 <ste||ar-github> hpx/future_overhead_timing 4cc0620 John Biddiscombe: Add CDash compatible time output to some benchmarks

12:52 <ste||ar-github> hpx/future_overhead_timing 21c3ec5 John Biddiscombe: Exclude benchmarks from CTest that are not HPX threaded, or currently failing

15:20 <zao> heller: I'm still messing around with the topology for degenerate machines. For N PUs, would it make more sense from a HPX perspective to have N fake cores with one PU each, or 1 fake core with N PUs? There is no information about hyperthreads anywhere, it's just a flat bunch of PUs, one per hardware thread.

15:20 <zao> Both ways will be a bit silly, just looking for the one that will make HPX algos suck less.

15:20 <zao> Leaning toward N single-PU cores.

15:28 <K-ballo> toss a coin

15:29 <K-ballo> where's hkaiser these days?

15:36 <jbjnr__> I was wondering the same

15:37 <jbjnr__> zao: N single PU cores would be the place to start

15:38 <zao> Punching the partitioner now, that ends up with empty pools.

15:38 <zao> Thus far I've found out that the remote GDB functionality of both VC++, VS Code and Qt Creator doesn't work :D

15:39 <zao> Also, I can't for the life of me get logging working.

15:44 <ste||ar-github> [hpx] biddisco force-pushed future_overhead_timing from 21c3ec5 to b147e77: https://github.com/STEllAR-GROUP/hpx/commits/future_overhead_timing

15:44 <ste||ar-github> hpx/future_overhead_timing 65a6e1a John Biddiscombe: Add print utility for DartMeasurentTime to lightweight_test...

15:44 <ste||ar-github> hpx/future_overhead_timing 3130529 John Biddiscombe: Add tests.performance.local to unit testing

15:44 <ste||ar-github> hpx/future_overhead_timing 370d7f1 John Biddiscombe: Add CDash compatible time output to some benchmarks

16:06 <zao> New and exciting failures... "mmap() failed to allocate thread stack" and "assertion 'mode_ & mode_allow_oversubscription || num_threads == cfg_.num_threads_' failed: HPX(assertion_failure)"

16:06 <zao> Guessing I'm inconsistent somewhere, there's so many friendly helper functions in topology :)

17:43 <jbjnr__> zao: I get those mmap stack failed errors no some of the benchmarks

17:44 <jbjnr__> ^no = on

17:47 <zao> It would help if I had a proper development environment and not a GDB that crashes on accessing our thread_specific_ptr:s :D

18:12 <ste||ar-github> [hpx] biddisco force-pushed future_overhead_timing from b147e77 to cd86d5d: https://github.com/STEllAR-GROUP/hpx/commits/future_overhead_timing

18:12 <ste||ar-github> hpx/future_overhead_timing 297784d John Biddiscombe: Add CDash compatible time output to some benchmarks

18:12 <ste||ar-github> hpx/future_overhead_timing cd86d5d John Biddiscombe: Exclude benchmarks from CTest that are not HPX threaded, or currently failing

18:31 <ste||ar-github> [hpx] biddisco force-pushed future_overhead_timing from cd86d5d to a4687de: https://github.com/STEllAR-GROUP/hpx/commits/future_overhead_timing

18:31 <ste||ar-github> hpx/future_overhead_timing 88e03d1 John Biddiscombe: Add tests.performance.local to unit testing

18:31 <ste||ar-github> hpx/future_overhead_timing 17dda8d John Biddiscombe: Add print utility for DartMeasurentTime to lightweight_test...

18:31 <ste||ar-github> hpx/future_overhead_timing ebc7d16 John Biddiscombe: Add CDash compatible time output to some benchmarks

20:12 <ste||ar-github> [hpx] biddisco force-pushed future_overhead_timing from a4687de to dc3e85e: https://github.com/STEllAR-GROUP/hpx/commits/future_overhead_timing

20:12 <ste||ar-github> hpx/future_overhead_timing 3c9c0bb John Biddiscombe: Add CDash compatible time output to some benchmarks

20:12 <ste||ar-github> hpx/future_overhead_timing dc3e85e John Biddiscombe: Exclude benchmarks from CTest that are not HPX threaded, or currently failing