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/
parsa has joined #ste||ar
<heller> hmm
<github> [hpx] hkaiser created performance_optimizations (+1 new commit): https://git.io/vFH4s
<github> hpx/performance_optimizations d0e95d2 Hartmut Kaiser: Speeding up accessing the resource partitioner and the topology info...
<Guest96225> [hpx] hkaiser opened pull request #3015: Speeding up accessing the resource partitioner and the topology info (master...performance_optimizations) https://git.io/vFH4G
jaafar_ has quit [Ping timeout: 268 seconds]
diehlpk has quit [Ping timeout: 248 seconds]
hkaiser has quit [Quit: bye]
K-ballo has quit [Quit: K-ballo]
parsa has quit [Quit: Zzzzzzzzzzzz]
<github> [hpx] StellarBot pushed 1 new commit to gh-pages: https://git.io/vFHrK
<github> hpx/gh-pages 373887c StellarBot: Updating docs
<github> [hpx] biddisco force-pushed performance_optimizations from d0e95d2 to 7afcce6: https://git.io/vFH6d
<github> hpx/performance_optimizations 7afcce6 Hartmut Kaiser: Speeding up accessing the resource partitioner and the topology info...
hkaiser has joined #ste||ar
<github> [hpx] hkaiser pushed 1 new commit to master: https://git.io/vFHXh
<github> hpx/master 05a2335 Hartmut Kaiser: Merge pull request #3008 from STEllAR-GROUP/remove-get_full_machine_mask...
<github> [hpx] hkaiser pushed 2 new commits to master: https://git.io/vFH1v
<github> hpx/master 0fc026b Mikael Simberg: Silence warning about casting away qualifiers in itt_notify.hpp
<github> hpx/master 33813a0 Hartmut Kaiser: Merge pull request #3012 from msimberg/silence-const-cast...
<hkaiser> nice
<hkaiser> do you have the full pdf?
<heller> Not right now
<hkaiser> k
<heller> Have to log on to the university network
<heller> Remind me tomorrow
<hkaiser> pls do ;)
<heller> Of to the swimming pool :p
<hkaiser> heller: I've got it
<heller> Good
<heller> hkaiser: those guys are actively working on a real time scheduler now
<hkaiser> nice
<hkaiser> yah, the conckusions are that the overheads are too high for their use cases
david_pfander has joined #ste||ar
david_pfander has quit [Client Quit]
<heller> hkaiser: difficult situation
<hkaiser> any idea what that guys means?
<heller> hkaiser: traditionally their workload is small due to small CPUs.
<hkaiser> right
<heller> Hen and egg problem
<heller> Real time for multi core is an unsolved problem
<heller> hkaiser: re reddit post. Seems like a premature optimization person
<hkaiser> yes
<heller> Looking for a solution for something he thinks is bad without even saying what he wants
<heller> Or currently has
<hkaiser> writing an answer saying exactly this ;)
K-ballo has joined #ste||ar
denisblank has joined #ste||ar
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
eschnett has quit [Quit: eschnett]
<github> [hpx] hkaiser pushed 1 new commit to fixing_dataflow: https://git.io/vFHb4
<github> hpx/fixing_dataflow b1565b4 Hartmut Kaiser: Adding atomic 'finished' flag....
<hkaiser> denisblank: yt?
parsa has quit [Quit: Zzzzzzzzzzzz]
<denisblank> hkaiser: yes
<zao> Ah, saw that you already found the PDF, bah.
<hkaiser> denisblank: bbiab, sorry gotta run now
<hkaiser> denisblank: see the PR
<denisblank> hkaiser: ok
eschnett has joined #ste||ar
<heller> wash: hpx takes a lock when scheduling tasks :p
<hkaiser> heller: only when scheduling threads
<hkaiser> denisblank: back now
jaafar_ has joined #ste||ar
eschnett has quit [Quit: eschnett]
parsa has joined #ste||ar
diehlpk has joined #ste||ar
<heller> hkaiser: sure, that's what I wrote. The OPs initial assessment that locks are bad thing is what's to blame
<heller> I'm still wondering about how to get rid of the thread_map_...
<heller> Incidentally, mikael stumbled over the very same question
<diehlpk> heller, Have you ever debugged over shh with circle-ci?
<diehlpk> I try to access the docker image but can not find the build folder?
Smasher has quit [Ping timeout: 246 seconds]
Smasher has joined #ste||ar
<hkaiser> heller: no, you wrote that a lock is acquired while scheduling a task
denisblank has quit [Quit: denisblank]
<diehlpk> hkaiser, Any idea about the circle-ci issue?
<hkaiser> diehlpk: what issue?
<diehlpk> It is quite strange that the hpx examples are running, but not PeriHPX
<diehlpk> terminate called after throwing an instance of 'hpx::detail::exception_with_info<hpx::exception>'
<diehlpk> what(): description is nullptr: HPX(bad_parameter)
<hkaiser> show me
<diehlpk> Do you have access to PeriHPX circle-ci https://circleci.com/gh/diehlpk/PeridynamicHPX/193
<hkaiser> hmm
<hkaiser> does it run locally for you?
<diehlpk> Yes, on Fedora and Ubuntu
<hkaiser> hmm, how do I build this?
<diehlpk> Check out the repo and build yaml-cpp
<hkaiser> k
<hkaiser> give me some time for this pls
<diehlpk> blaze and blaze-iterative is shipped within
<diehlpk> No problem
<diehlpk> It is not important
<diehlpk> By the way, you are now third author of the PeriHPX paper
<hkaiser> diehlpk: have you tried ssh'ing into the circleci instance?
<hkaiser> thanks
<zao> For us playing along at home, what repo is that?
<diehlpk> Yes, started the docker image from build and run the hpx example there and it is working
<diehlpk> Running my example results in this strange error
<hkaiser> I meant to directly ssh into the circleci instance as it builds things
<diehlpk> I was only able to run the docker image when the circle.yaml is finished
<diehlpk> zao, It is a private repo, but I can add you
<diehlpk> hkaiser, ssh alows you to access the node
<hkaiser> yes
<diehlpk> But docker let me load the image only when all things are finished
<hkaiser> at the top of the circleci site is a tab 'Debug via SSH'
<diehlpk> Yes
<diehlpk> I did this
<diehlpk> ssh + docker run -v $PWD:/hpx -w /hpx/ -i -t peridhpx:build_env /bin/bash
<diehlpk> In the docker image, I was running the fibonacci example and hello_world
<diehlpk> These run perfectly
<diehlpk> When running my own examples always this error shows up
<diehlpk> We had the same for HPXCL
<diehlpk> I assume that it maybe relates to the cloned containers from the hpx dev docker image
<diehlpk> One difference is that we hpx_main and in the other case we use hpx_init
<zao> diehlpk: I would be 'zao' on github, but like HK, I promise nothing :)
<hkaiser> diehlpk: well, run it in gdb and see where it blows up?