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/
Matombo has quit [Remote host closed the connection]
pree has joined #ste||ar
pree has quit [Ping timeout: 272 seconds]
eschnett has joined #ste||ar
EverYoung has quit [Ping timeout: 240 seconds]
K-ballo has quit [Quit: K-ballo]
denis_blank has quit [Quit: denis_blank]
hkaiser has quit [Quit: bye]
patg has quit [Quit: See you later]
bikineev has joined #ste||ar
Matombo has joined #ste||ar
ajaivgeorge has quit [Ping timeout: 240 seconds]
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
bikineev has quit [Read error: Connection reset by peer]
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
bikineev has quit [Ping timeout: 240 seconds]
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
K-ballo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
hkaiser has joined #ste||ar
<hkaiser>
heller_: ?
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
bikineev has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
<heller_>
hkaiser: hey
<hkaiser>
heller_: hey
<hkaiser>
could you explain to me pls how the parcel_await can be fixed?
<heller_>
hkaiser: no idea yet :/
<heller_>
I hope I can get to it in about an hour or two
<hkaiser>
ok
<heller_>
sorry ... the allscale review preperation is taking all my time at the moment...
<heller_>
I think I fixed the last blockers there now ...
<hkaiser>
k
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
pree has joined #ste||ar
bikineev_ has joined #ste||ar
bikineev has quit [Ping timeout: 255 seconds]
pree has quit [Read error: Connection reset by peer]
pree has joined #ste||ar
denis_blank has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
<denis_blank>
Hi, yesterday I've sent a mail to the hpx-users mailing list and it wasn't distributed yet. Is the mailing list moderated somehow?
<zao>
If memory serves me right, it's moderate-on-first-message or something.
<denis_blank>
Ok, thanks
* zao
eyes hkaiser
pree has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
eschnett has quit [Quit: eschnett]
pree has joined #ste||ar
<hkaiser>
denis_blank: I have not seen a moderation request, could you resend that email please?
<hkaiser>
heller_: parcel_await is utterly broken in many more ways if more than one parcel is being processed
<hkaiser>
this requires a complete redesign
<hkaiser>
denis_blank: what email address have you used?
pree has quit [Read error: Connection reset by peer]
<denis_blank>
hkaiser it was denis.blank@outlook.com
<hkaiser>
denis_blank: I meant what email you used for the list?
<denis_blank>
hpx-users@stellar.cct.lsu.edu
<hkaiser>
ok, pls send it one more time, then
<heller_>
hkaiser: can you be a bit more specific please?
<hkaiser>
for instance, the preprocess has_futures() function sets the promise to ready after the first parcel (if that has none), but what of parcels after that actually will have parcels?
<hkaiser>
will have futures*
<hkaiser>
then the promise gets set more than once, is that desired?
<hkaiser>
I'm not even sure what will happen in that case
<denis_blank>
hkaiser I just resent it, and it was distributed...
<hkaiser>
nod
<K-ballo>
denis_blank: in your github link to wait_all, if you click on "Blame" you'll see a list of the latest commits for each line
<heller_>
at the time, I couldn't figure out a different way to achieve that
<denis_blank>
K-ballo: but why would we prefer a sequential future request behaviour over a parallel one?
<hkaiser>
that fix-point iteration over the parcels is not necessary as we don't need to maintain the original sequence of parcels
<K-ballo>
denis_blank: for reasons that someone will explain in detail on the list
<hkaiser>
ready parcels could be sent immediately
<K-ballo>
denis_blank: I just wanted to show you how you can dig information on rationale.. we are not very good at using git properly, but the goal is that git-blame can be used to understand the Why behind what it's being done
<denis_blank>
K-ballo: Thanks, I will take a look into it
pree has joined #ste||ar
<heller_>
hkaiser: the main reason for not sending ready parcels immediately is mostly to keep the parcel coalescing in place.
<hkaiser>
it is not kept in place anyways as each parcel gets put separately anyways
<hkaiser>
that's another of those 'broken in many ways' points
<heller_>
hkaiser: parcel awaiting happens in the parcelport, not the parcel handler. It might make sense to have it in the parcel handler indeed, before calling put_parcel on the PP or message handler. This will require all parcelports to have the same archive flags though. We discussed the "put separately" issue already and agreed that it should not be as is currently
<hkaiser>
nod
<hkaiser>
ok
<hkaiser>
just trying to raise the urge to fix this mess ;)
<heller_>
sure
patg has joined #ste||ar
hkaiser has quit [Quit: bye]
pree has quit [Read error: Connection reset by peer]
<ABresting>
over this link, Bryce and I have setup a repo for any questions I have
<ABresting>
this way I can document all those and get overall knowledge
<ABresting>
I would really appreciate it if you answer some of the few questions I have mentioned
pree has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
denis_blank has quit [Quit: denis_blank]
<hkaiser>
ABresting: ok
pree has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
shoshijak has quit [Ping timeout: 240 seconds]
bikineev has quit [Remote host closed the connection]
david_pfander has quit [Ping timeout: 272 seconds]
pree has joined #ste||ar
pree has quit [Ping timeout: 245 seconds]
pree has joined #ste||ar
pree has quit [Ping timeout: 240 seconds]
<heller_>
hkaiser: ok, now towards a better parcel_await
<heller_>
hkaiser: give me 30 minutes please
<heller_>
I have almost everything implemented
<hkaiser>
heller_: cool
<heller_>
if those allscale people wouldn't interfere :P
<heller_>
hkaiser: btw, the cmake patches should be especially helpful on cray machines ...
<heller_>
or anything with a network storage...
pree has joined #ste||ar
<heller_>
doesn't feel faster though
<hkaiser>
heller_: half of the tests are still being executed
<heller_>
hkaiser: is there anything blocking #2640 still`
<heller_>
ah, ok
<heller_>
assuming circle-ci completes onw
<hkaiser>
heller_: waiting for circleci
<heller_>
great!
<hkaiser>
did he fix the inspect problems?
<heller_>
I hope so ...
<hkaiser>
k
<heller_>
I need some nice statistics for collaboration etc. for the review :P
<hkaiser>
sure, understood - but he fixed the conflicts only today
<heller_>
I know
<heller_>
I kept pushing him...
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
bikineev has joined #ste||ar
bibek_desktop has quit [Quit: Leaving]
bikineev has quit [Remote host closed the connection]
bibek_desktop has joined #ste||ar
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree has quit [Ping timeout: 260 seconds]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree has joined #ste||ar
pree has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
hkaiser has quit [Quit: bye]
pree has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
pree has quit [Ping timeout: 246 seconds]
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
Matombo has joined #ste||ar
bikineev has joined #ste||ar
<heller_>
jbjnr: yt?
pree has joined #ste||ar
pree has left #ste||ar ["Ex-Chat"]
pree has joined #ste||ar
pree has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
<pree>
While building components -> After creating a shared library. I used this command " export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:$HPX_LOCATION/lib/pkgconfig c++ -o hello_world_client hello_world_client.cpp \ `pkg-config --cflags --libs hpx_application` \ -L${HOME}/my_hpx_libs -lhello_world -lhpx_iostreams" . But it gives me an error as "/usr/bin/ld: cannot find -lhello_world. What does -lhello_world in this command mean whether it
<pree>
help
EverYoung has quit [Remote host closed the connection]
eschnett has quit [Quit: eschnett]
<heller_>
pree: it means it tries to link against a library called libhello_world.so
<heller_>
which is searched in ${HOME}/my_hpx_lib
<pree>
what lhello_world here is ?
<pree>
how can I correct this?
<pree>
whether lhello_world refers to a library name ?
<heller_>
-l<name> tells the linker that you like to link against a library call lib<name> (lower case L)
<heller_>
-L<dir> adds <dir> to the library search path
<pree>
so have to specify the library name without lib right ?
<pree>
Thank you so much
<heller_>
yes
<heller_>
just regular linking of shared libraries
pree has quit [Ping timeout: 260 seconds]
<diehlpk_work>
boost::mutex m;
<diehlpk_work>
~~~~~~~^
<diehlpk_work>
Does anyone know what the problem here
<heller_>
diehlpk_work: why not the one from boost/thread?
<diehlpk_work>
Ok
diehlpk_work has quit [Quit: Leaving]
jgoncal has joined #ste||ar
hkaiser has joined #ste||ar
Matombo has quit [Remote host closed the connection]
<heller_>
hkaiser: not having the zombie state for split_gids_ doesn't fix anything :/
<heller_>
at least not for me
jgoncal has quit [Quit: Page closed]
<hkaiser>
it fixed the zombie state for sure ;)
<heller_>
yeah :P
<heller_>
hkaiser: could you think of any race condition that might occur when calling future_data::set_on_completed?
jgoncal has joined #ste||ar
<hkaiser>
heller_: I'd assume that works, otherwise everything else would break
<heller_>
hmpf
jgoncal has quit [Client Quit]
<heller_>
for some reason, the LF PP throws an operation not permitted error now :/
<heller_>
oh no, the MPI PP does the same...
<heller_>
prior to HPX startup...
<heller_>
my fault...
bikineev_ has joined #ste||ar
bikineev has quit [Ping timeout: 246 seconds]
<github>
[hpx] K-ballo force-pushed drop-vs2013 from 44d8c58 to b12b7ee: https://git.io/vHJsA
<github>
hpx/drop-vs2013 4f35039 Agustin K-ballo Berge: Remove support for VS2013
<github>
hpx/drop-vs2013 006cb46 Agustin K-ballo Berge: Mark C++11 defaulted functions as required
<github>
hpx/drop-vs2013 7549d8d Agustin K-ballo Berge: Mark C++11 noexcept as required
<heller_>
hkaiser: I commited what I have now to lf_multiple_parcels
<github>
[hpx] sithhell force-pushed lf_multiple_parcels from b868003 to 28d62a0: https://git.io/vHnkr
<github>
hpx/lf_multiple_parcels 3540cb1 Thomas Heller: Enabling multiple parcels to be sent at once.
<github>
hpx/lf_multiple_parcels 271bf92 Thomas Heller: Increasing stack size to accomodate for stack overflows with the LF PP
<github>
hpx/lf_multiple_parcels d41cf75 Thomas Heller: Improvements to multiple parcel sends
<hkaiser>
heller_: ok, should I close my PR, then?
<heller_>
hkaiser: your call, if you are happy with the parcel_await changes I made, there probably is no need for it
<hkaiser>
heller_: sure, this PR is currently broken anyways...
<heller_>
the PR doesn't entirely fix the LF hangs ... MPI is going through nicely though
<heller_>
yeah
<github>
[hpx] hkaiser closed pull request #2652: This patch makes sure that all parcels in a batch are properly handled (master...fixing_parcel_await) https://git.io/vHsBD
<heller_>
criticism welcome, of course
<heller_>
I think the parcel_await is as good as it can get at this stage
<heller_>
I am absolutely in favor of moving it to the parcel handler though
wash has quit [Quit: Lost terminal]
<heller_>
that is, as you mentioned, call put parcel after each one is finished, but then either push it to the parcelport or the message handler
bikineev_ has quit [Remote host closed the connection]
<heller_>
but the changes included in this PR should fix the immediate problems we discussed earlier...
<heller_>
hkaiser: ok, I can now see how the parcel awaiting might lead to deadlocks...
<github>
[hpx] hkaiser deleted fixing_parcel_await at 4218539: https://git.io/vHntZ
<heller_>
hmm, maybe not, red herring
<hkaiser>
:D
<heller_>
any parcel that is generated during awaiting is handled seperately...