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://irc.cct.lsu.edu/
<hkaiser>
Smasher: local_new might not cut it, you'll probably need new_
<mcopik>
hkaiser: there's a slim chance that I'll make it. they made a reservation on my booking for a connecting flight through Montreal and I'll be there at 14:19, so maybe I'll be able to join the last session of the workshop
<mcopik>
assuming that they'll give me a boarding pass in Frankfurt
<mcopik>
because their system does not see my bag anymore
<mcopik>
and the probability of my bag being lost is increasing drastically
<hkaiser>
uhh, well I hope all will go well
<Smasher>
like this, right? hpx::components::new_<hpx_zfm_fingerprint_provider_server>(hpx::find_here());
<hkaiser>
nod
<Smasher>
error: no matching function for call to ‘new_()’
<Smasher>
huh?
<Smasher>
nvm my slip i guesss
<hkaiser>
hpx::new_
<Smasher>
is it in namespace hpx?
<hkaiser>
yah
<Smasher>
i am using hpx::components:new_ everywhere :D
<hkaiser>
works as well, it's imported to hpx
<Smasher>
IT WOOOOOOOOOOOORKS
<Smasher>
HOLY MOLLY
<Smasher>
i just triggered via the network the fingerprint scanner on my raspberry pi
<Smasher>
im excited
<Smasher>
lets see now if the file can be really transfered
<hkaiser>
:D
<Smasher>
0xb1f9c238: ??? + 0xb1f9c238 in /home/pi/xc/local-dbg/lib/libhpxd.so.1
<Smasher>
0xb1f9c3b4: ??? + 0xb1f9c3b4 in /home/pi/xc/local-dbg/lib/libhpxd.so.1
<Smasher>
0xb206d990: hpx::termination_handler(int) + 0x188 in /home/pi/xc/local-dbg/lib/libhpxd.so.1
<Smasher>
im since 20 hours awake and working... gotta take a break i guess and get some sleep
K-ballo has quit [Quit: K-ballo]
<hkaiser>
Smasher: no idea what's wrong, sorry
hkaiser has quit [Quit: bye]
mcopik has quit [Ping timeout: 246 seconds]
vamatya has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 272 seconds]
shoshijak has joined #ste||ar
vamatya has quit [Ping timeout: 268 seconds]
shoshijak has quit [Ping timeout: 240 seconds]
ABresting[m] has joined #ste||ar
shoshijak has joined #ste||ar
david_pfander has joined #ste||ar
bikineev has joined #ste||ar
Matombo has joined #ste||ar
Matombo has quit [Remote host closed the connection]
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
jakemp has quit [Ping timeout: 240 seconds]
bikineev has quit [Remote host closed the connection]
jakemp has joined #ste||ar
bikineev has joined #ste||ar
pree has joined #ste||ar
<pree>
when I try to build hpx-applications with pkg-config in ubuntu 16.04 .I got an error "/usr/bin/ld: cannot find -liostreams". how i can solve this?
<heller__>
pree: which build type is your hpx build?
<pree>
I have built hpx with cmake --DBOOST_ROOT=/root/of/boost/installation \ --DHWLOC_ROOT=/root/of/hwloc/installation [other CMake variable definitions] \ -/path/to/source/tree
<pree>
Cmake
<heller__>
what is "[other CMake variable definitions]"?
<pree>
-DCMAKE_INSTALL_PREFIX=~home/praveen/hpx
<heller__>
should be -lhpx_iostreams
<pree>
-lhpx_iostreams is working but I have noticed another error when I try to execute './hello_world' error while loading shared libraries: libjemalloc.so.2: cannot open shared object file: No such file or directory
<heller__>
it has to be in your LD_LIBRARY_PATH
bikineev has quit [Ping timeout: 268 seconds]
<pree>
Sorry I can't understand it please make it clear.
<heller__>
Smasher: I suspect that just not all actions are registered
<jbjnr>
Smasher: when you get that bstream error, it means that a parcel was sent from one place to another and when it is decoded, something is wrong. in your case, it is highly likely that the type of action or the datatype of one or more argument is differnet on the target node than the sourc node
<jbjnr>
^what heller said. the action type is not registered on one node, so the deserialization layer barfs
<jbjnr>
heller__: rma chunks now supported :)
<jbjnr>
I've added a 3rd chunk type
<hkaiser>
if actions are not registered he would see another error
<K-ballo>
Smasher: given the switch from array<T, N> to vector<T>, I'd suspect vector<T>::size_type to be responsible
<hkaiser>
I think it's data-type size mismatches
<hkaiser>
K-ballo: doh! thanks for catching this
<heller__>
hkaiser: not necessarily. different typeid(...).name()s might lead to that error due to id mismatches
<K-ballo>
shouldn't the archive be dealing with different integer sizes?
<github>
[hpx] hkaiser created fixing_container_serialization (+1 new commit): https://git.io/v9biJ
<jbjnr>
I've just seen severus snake in my pond. I'm delighted. didn't see him at all last summer. he's back!
<hkaiser>
heller__: ok, at some point we were compressing integers tomitigate the bloat, in the new serialization this has been removed
<jbjnr>
it should only be enabled when it's actually need though.
<jbjnr>
^needed
<heller__>
hkaiser: yes, this has been removed as the compression showed up pretty high when profiling
<hkaiser>
jbjnr: what do you mean
<Smasher>
hkaiser oh another update , cool :)
<jbjnr>
-Ihpx.serialization.arch.independent=1 or something
<jbjnr>
so that the 99.9% of the time when we run on homogeneous systems we don't pay extra or anything
<jbjnr>
mind you. size_t is 64 nits everywhere anyway
<jbjnr>
^bits. <grrrr>
<hkaiser>
jbjnr: it isn't
<hkaiser>
on 32bit platforms its usially uint32_t
<jbjnr>
I mean everywhere we run the other 99.9 % of the time
<hkaiser>
well, sure
<jbjnr>
so promoting to 64 isn't costing anything
<jbjnr>
so you can ignore me
<jbjnr>
(if you weren't already)
<hkaiser>
as long as yu're talking about size_t, sure, but uint16_t would still get promoted
bikineev has quit [Ping timeout: 268 seconds]
<jbjnr>
help! - suppose I serialize some stuff, one object of which is an rma::vector - it gets placed in the archive as an rma_chunk. fine. if this is sent over the wire, it will have the rma data transferred via the parcelport and it will be copied into some buffer on the other side. If however I write a unit test that simply serializes and rma vector into an archive, then deserializes it into...
<jbjnr>
...another rma vector - no data transfer takes place.
<jbjnr>
Is there a way of knowing if the deserialization is local, or not?
<jbjnr>
cos we can add a memcopy if local, but I'm not sure when to trigger it.
<hkaiser>
if serialization occurs it has to be remote (in real world use cases)
<jbjnr>
in my unit test example - what to do? make a better test?
<hkaiser>
so in your test you have to manually add the memcpy before deserializing
<jbjnr>
hmmmm
<jbjnr>
I'll ponder some more
bikineev has joined #ste||ar
<jbjnr>
bikineev: if I want to run the cpp serializers tests, do I just clone and go, or are there tricky other things that need to be done?
<github>
[hpx] hkaiser deleted algorithms_signatures at 053db1b: https://git.io/v9bxL
<Smasher>
hkaiser maybe you know why gdb still doesnt load symbols for my shared library? i have the following picture http://sprunge.us/bjEN to load the symbols i do: add-symbol-file /cygdrive/d/git/ba-phmf/cmake-build-debug-arm-noide/phmf/hd_bozorth3/libhpx_bz3_client.so 0x21c498 it outputs Reading symbols from /cygdrive/d/git/ba-phmf/cmake-build-debug-arm-noide/phmf/hd_bozorth3/libhpx_bz3_clie
<Smasher>
nt.so...done.
<Smasher>
but the picture doesn't change
<Smasher>
only if u have a minute :)
akheir has joined #ste||ar
<hkaiser>
Smasher: no idea what's wrong
<Smasher>
aha!
<wash>
hkaiser: btw, they said you'd be calling in via skype for the steering committee meeting. Just wanted to confirm?
<Smasher>
set solib-search-path /cygdrive/d/git/ba-phmf/cmake-build-debug-arm-noide/phmf/hd_bozorth3/
<hkaiser>
wash: I will, thanks
<hkaiser>
wash: I guess no need for you to show up, then
<wash>
gotcha
<Smasher>
hkaiser dont you use gdb a lot for your work?
<hkaiser>
Smasher: not at all
<hkaiser>
I use VS
<Smasher>
ah i see
<Smasher>
i already knew that actually
<Smasher>
hkaiser have you already did crosscompiling with vs?
<Smasher>
done*
<hkaiser>
none
<hkaiser>
I use gdb if I have to directly on linux
<jbjnr>
did anyone else try building that cpp-serializers stuff?
thundergroudon has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
pree has joined #ste||ar
pree has quit [Ping timeout: 260 seconds]
<Smasher>
i currently have an issue that i cant call the same action on a component twice
<Smasher>
i get 0xb2902f70: hpx::threads::coroutines::detail::coroutine_impl::operator()() + 0x110 in /home/pi/xc/local-dbg/lib/libhpxd.so.1: HPX(invalid_status)
<github>
hpx/master da51f54 Hartmut Kaiser: Fixing problem with previous commit...
<heller__>
:D
<hkaiser>
heller__: sanfu here on my machine..., sorry for that
<heller__>
no problem
<heller__>
hkaiser: "wo gehobelt wird fallen späne..."
taeguk has quit [Ping timeout: 260 seconds]
hkaiser has quit [Quit: bye]
<Smasher>
heller__ nanu, was für Weißheiten :)
<Smasher>
-> From chipping come chips.
<Smasher>
Or i actually like this translation more: You have to break an egg to make an omelette.
<K-ballo>
now I'm hungry
thundergroudon has quit [Ping timeout: 246 seconds]
aserio has quit [Quit: aserio]
bikineev has joined #ste||ar
david_pf_ has joined #ste||ar
<Smasher>
K-ballo le baguette?
<K-ballo>
la
<Smasher>
une
vamatya_ has joined #ste||ar
vamatya has quit [Ping timeout: 268 seconds]
hkaiser has joined #ste||ar
<Smasher>
hmmm
<Smasher>
something new here
<Smasher>
{what}: new allocator failed to allocate memory: HPX(out_of_memory)
<Smasher>
39% free here
<hkaiser>
stack overflow?
<Smasher>
very suspicious
<Smasher>
i dont do anything except int hpx_main() { hpx::new_<bz_server>(hpx::find_here()); hpx::agas::register_name("phmf_server", hpx::find_here()); }
<Smasher>
that is okay...
<Smasher>
then i start the client in debug mode, with gdb attached
<Smasher>
set the brakepoint to the first line of hpx_main
<Smasher>
run it and the server crashes before gdb halts
<hkaiser>
run the server with gdb and see where it crashes
akheir has quit [Remote host closed the connection]