aserio changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar.cct.lsu.edu | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | Buildbot: http://rostam.cct.lsu.edu/ | Log: http://irclog.cct.lsu.edu/
<hkaiser> K-ballo: I thought I did fix the tests :/
<K-ballo> something's off, if they were truly broken they'd shown on circle
<hkaiser> nod
<hkaiser> you're right, I missed at least one
<K-ballo> what do I need to do for the tests to run?
<github> [hpx] hkaiser pushed 1 new commit to fixing_2730: https://git.io/vQ08W
<github> hpx/fixing_2730 d4ca218 Hartmut Kaiser: Started to make lcos::channel migratable
<K-ballo> copy dlls somewhere?
<hkaiser> K-ballo: you might have to copy the boost and hwloc dlls
<K-ballo> nod, where?
<K-ballo> Debug/bin maybe?
<hkaiser> build/debug/bin
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/vQ08E
<github> hpx/master ebfa638 Hartmut Kaiser: Fixing inclusive_scan_test
<hkaiser> K-ballo: ^^
<hkaiser> ahh, right
<K-ballo> oh crap I closed the doc
<K-ballo> saved it, how smart.. these ones fail for me:
<K-ballo> inclusive_scan_executors_v1_test_exe
<K-ballo> inclusive_scan_test_exe
<K-ballo> reduce_by_key_test_exe
<K-ballo> partitioned_vector_inclusive_scan_test_exe
<K-ballo> scan_different_inits_test
<K-ballo> inclusive-scan related
<hkaiser> ok, will look
<K-ballo> I wonder why they don't fail on circle, should try those again on top-of-master
<hkaiser> by default inclusive_scan compatibility is still on
<K-ballo> does that mean I turned it off?
<K-ballo> HPX_WITH_INCLUSIVE_SCAN_COMPATIBILITY:BOOL=ON
<K-ballo> btw typo "inclkusive_scan"
<K-ballo> I'll check the inclusive scan compatibility in more detail.. perhaps I broke the compatibility case
<hkaiser> K-ballo: where is the typo?
<hkaiser> ahh
<hkaiser> tks
<K-ballo> that defines.hpp of mine again...
<K-ballo> effectively turns of inclusive scan compatibility
<K-ballo> I should remove hpx from the global include paths
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/vQ04s
<github> hpx/master 12e7720 Hartmut Kaiser: More inclusive_scan test fixes...
<hkaiser> should be fine now
<K-ballo> 1> The following tests FAILED:
<K-ballo> 1> 163 - tests.unit.components.partitioned_vector_subview (Failed)
<K-ballo> 1> 457 - tests.unit.threads.thread_affinity (Failed)
<K-ballo> partitioned_vector_subview might be missing some more dlls?, it just says failed with no output
<github> [hpx] K-ballo opened pull request #2734: Add C++11 range utilities (master...range) https://git.io/vQ0Bz
bikineev has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
<hkaiser> K-ballo: I'll look tomorrow
<zao> K-ballo: There we go, swapped the guts of my computers out. I now have one that can actually compile :)
<hkaiser> K-ballo: should we ban boost.range now?
hkaiser has quit [Quit: bye]
<zao> Failing a fair number of regression tests, are you people building those?
<zao> *running
<K-ballo> the core functions, possibly
<K-ballo> I built the RUN_TESTS target, I expect that ran regression tests too, but now I'm having doubts
<zao> Seems like most of my failures are distributed.
<K-ballo> build with extra flags? mpi and such?
<zao> Stock build with Boost, hwloc and Python.
<zao> I wouldn't begin to know what "MPI" would be on a desktop with MSVC.
<zao> Assumedly OpenMPI or some other abomination.
<K-ballo> I'm running regressions again just in case
<zao> (first few hundred, this is sloooow)
<zao> I thought that HPX tests were especially horrible until I ran into something else using CTest at work, btw :D
<zao> 1>145/514 Test #145: tests.unit.components.distributed.tcp.migrate_component ...................................***Timeout 1500.01 sec
<zao> Not the best timeout, that.
<K-ballo> 1> 99% tests passed, 1 tests failed out of 108
<K-ballo> 1> The following tests FAILED:
<K-ballo> 1> 77 - tests.regressions.parallel.scan_different_inits (Failed)
<K-ballo> expected failure, since it didn't build
<K-ballo> that's a lot of regressions indeed
K-ballo has quit [Quit: K-ballo]
<zao> Is there actual output anywhere? Most probably not.
<zao> "C:\stellar\build-hpx\Testing\Temporary\LastTest.log"
<zao> Apparently.
<zao> Process 0 failed with an unexpected error code of -1073741819 (expected 0)
<zao> Welp. that's not overly helpful.
<zao> c000'0005, missing DLL probably.
EverYoung has joined #ste||ar
<zao> Oh no, that's just an access violation.
<zao> Trying to enter a critical section at 0x24.
<zao> Not going to go overly well, that.
<zao> Definitely not going to look through code that most probably never ever has worked right :(
<zao> Seems to be towards shutdown?
EverYoung has quit [Ping timeout: 276 seconds]
taeguk has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
pree has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
mcopik has quit [Ping timeout: 240 seconds]
bikineev has joined #ste||ar
taeguk has quit [Ping timeout: 260 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
pree has quit [Ping timeout: 268 seconds]
mcopik has joined #ste||ar
<github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://git.io/vQ0i4
<github> hpx/gh-pages 6c74330 StellarBot: Updating docs
K-ballo has joined #ste||ar
<zao> So, anyway, seems like TCP PP in tests is quite hosed and can't shut down on my Windows box.
<zao> Not sure if the tests do actual work or not, but it's a bit suspicious that they run pretty much universally for 17.7x seconds.
<zao> Still, shouldn't _crash_ on teardown.
parsa has joined #ste||ar
hkaiser has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
<github> [hpx] K-ballo force-pushed range from ba4fc66 to ac7597b: https://git.io/voasb
<github> hpx/range 1449669 Agustin K-ballo Berge: Add C++11 range utilities
<github> hpx/range 77c90af Agustin K-ballo Berge: Cleanup acquire traits, replace boost::range with util/range in their implementation, patch broken whens
<github> hpx/range 7382b8f Agustin K-ballo Berge: Replace core boost::range with util/range, remove redundant parallel traits
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
hkaiser has quit [Read error: Connection reset by peer]
bikineev has joined #ste||ar
hkaiser has joined #ste||ar
<hkaiser> zao it's not clear to me why so many tests fail for you :/
parsa has quit [Quit: Zzzzzzzzzzzz]
hkaiser has quit [Ping timeout: 260 seconds]
hkaiser has joined #ste||ar
bikineev has quit [Remote host closed the connection]
<hkaiser> K-ballo: I can confirm the partitioned_vector_subview_test failure
<hkaiser> K-ballo: this does not happen on linux because of the missing iterator debugging, I'll alert Antoine
Matombo has joined #ste||ar
pree has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
pree has joined #ste||ar
pree has quit [Ping timeout: 240 seconds]
mbremer has joined #ste||ar
pree has joined #ste||ar
denis_blank has joined #ste||ar
parsa has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
<zao> hkaiser: I attached a post-mortem debugger to one of the tests, had some stack traces linked before.
<zao> Seems like the tests do something (nothing?) for 17.7s and then tear down, crashing on null pointers when destroying some receivers.
<zao> So I'm suspecting that there's an underlying problem somewhere in shutdown, and that there's something (on my machine?) that blocks the tests from working properly in the first place.
<zao> (I've set my PATH to include the Boost and HWLOC installations I used)
<zao> No firewall present, no output from those tests, just the access violations.
<zao> Would I get more info in a Debug build, mayhaps?
* zao tries for the heck of it
bikineev has joined #ste||ar
hkaiser has quit [Read error: Connection reset by peer]
<K-ballo> my runs where Debug, perhaps slower enough not to fail/crash ?
<zao> Mine were RWDI.
<zao> Also going to do a rebuild of it, just in the remote case of a miscompile. I don't trust that CPU.
<zao> K-ballo: Took around two hours to swap the motherboard out to the Skylake one, so now I'm on fewer cores, but way more safe.
pree has quit [Quit: AaBbCc]
<zao> Debug tests seem to pass thus far.
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
<zao> Now this is handy, test suite popping modal UI when it fails :D
<zao> I guess there's some benefit to being able to inspect a test as it fails, but I hope it's mitigated on test runners :)
<K-ballo> ugh, we don't disable those on tests?!?
<K-ballo> I ran using the RUN_TESTS target and I did not get the popup, it must be setting the env var
<zao> This was run_tests in VS GUI.
<zao> Didn't get any of it back when running the RWDI tests, but proc:ed once with the Debug tests.
<zao> Only failure thus far.
<github> [hpx] K-ballo created iterator_range (+1 new commit): https://git.io/vQEfO
<github> hpx/iterator_range 6d880e5 Agustin K-ballo Berge: Implement Ranges-TS aware iterator_range utility
<zao> So what's all this range stuff about?
<K-ballo> dropping Boost.Range, switching to C++11 std::begin/end
<K-ballo> (using a boost::swap style wrapper for core, to avoid the using...adl dance)
<zao> Heh, tests.unit.threads.thread also failed.
<zao> I like being pulled out of my fullscreen entertainment by crash dialogs :)
<zao> K-ballo: Ooh, nice.
<zao> Heh, Retry button in the assert dialog doesn't actually let you attach a debugger, the test harness swallows it.
<zao> Debug tests from yesterday, a bit more healthy than RWDI.
<K-ballo> that's not a good sign though :/
EverYoung has joined #ste||ar
<zao> More green than I've ever seen from HPX, heh.
EverYoung has quit [Ping timeout: 246 seconds]
hkaiser has joined #ste||ar
<github> [hpx] K-ballo force-pushed iterator_range from 6d880e5 to 02a8b96: https://git.io/vQEUm
<github> hpx/iterator_range 02a8b96 Agustin K-ballo Berge: Implement Ranges-TS aware iterator_range utility
mbremer has quit [Ping timeout: 260 seconds]
denis_blank has quit [Quit: denis_blank]
Matombo has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 276 seconds]
<github> [hpx] K-ballo force-pushed iterator_range from 02a8b96 to bddbee6: https://git.io/vQEUm
<github> hpx/iterator_range bddbee6 Agustin K-ballo Berge: Implement Ranges-TS aware iterator_range utility
EverYoung has joined #ste||ar