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/
<github>
hpx/P0443R2 744d8dd Hartmut Kaiser: Finally rename apply_execute() to post()
<github>
hpx/P0443R2 86b39fc Hartmut Kaiser: Adding has_<op>_member<T> to executor tests
Matombo has joined #ste||ar
hkaiser_ has quit [Quit: bye]
bikineev has quit [Remote host closed the connection]
Matombo has quit [Remote host closed the connection]
diehlpk has quit [Remote host closed the connection]
diehlpk has joined #ste||ar
mcopik has quit [Ping timeout: 255 seconds]
K-ballo has quit [Quit: K-ballo]
diehlpk has quit [Ping timeout: 240 seconds]
diehlpk has joined #ste||ar
denis_blank has quit [Quit: denis_blank]
diehlpk has quit [Remote host closed the connection]
jaafar has joined #ste||ar
jaafar has quit [Client Quit]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
<pree>
parsa[w} : yt ?
<pree>
I try to some tests in rostam. when I execute this command after compilation "./ch" just works fine. But When I use srun -p leo -N 2 ./ch it tells srun: error: leo00: task 0: Aborted (core dumped)
<pree>
If there is error in program how the "./ch" works ?
<pree>
For even some simple programs (just printing my name) .. it tells error: trillian00: task 0: Exited with exit code 160
<pree>
What this means ? and how to correct it ?
<pree>
Thank you
<jbjnr>
heller_: SC papers notification is June 15th, but GB finalists won't be announced until sometime after that date
taeguk has quit [Quit: Page closed]
pree has quit [Ping timeout: 268 seconds]
Remko has joined #ste||ar
david_pfander1 has joined #ste||ar
Remko has quit [Remote host closed the connection]
david_pfander1 has quit [Ping timeout: 246 seconds]
Matombo has joined #ste||ar
<jbjnr>
wash cppcon submission deadline was yesterday - I didn't look. I'm too late!
<jbjnr>
heller_: I'm putting you down as a 'collaborator' on my internal project sheet.
bikineev has joined #ste||ar
<jbjnr>
bikineev: yt?
<bikineev>
jbjnr: hey, yeah
<jbjnr>
just wanted to say thanks for sending me those numbers
<jbjnr>
paper submitted. all ok
<jbjnr>
we are pretty awesome overall I think.
<bikineev>
jbjnr: no problem. Thank you for doing this
<bikineev>
(and putting me in the paper)
<jbjnr>
you worked hard on that stuff -even if it was a while ago
Matombo has quit [Remote host closed the connection]
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
pree has joined #ste||ar
shoshijak has joined #ste||ar
mcopik has joined #ste||ar
ajax67 has joined #ste||ar
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
Matombo has joined #ste||ar
<jbjnr>
very very quiet round here at the moment
<hkaiser>
g'morning
<jbjnr>
hi
Matombo has quit [Ping timeout: 240 seconds]
<pree>
hi
eschnett has quit [Quit: eschnett]
<zao>
*ribbit*
* K-ballo
yawns
aserio has joined #ste||ar
<hkaiser>
heller_: yt?
<pree>
parsa[w]: yt?
eschnett has joined #ste||ar
Matombo has joined #ste||ar
<github>
[hpx] hkaiser created lf_multiple_parcels at 5ac0d87 (+0 new commits): https://git.io/vHnkr
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
zbyerly_ has quit [Ping timeout: 255 seconds]
shoshijak has quit [Ping timeout: 240 seconds]
EverYoung has quit [Remote host closed the connection]
shoshijak has joined #ste||ar
mcopik has quit [Ping timeout: 240 seconds]
<pree>
Getting "srun: error: marvin00: task 0: Exited with exit code 160" when trying to execute files.
<pree>
What's this mean? In rostam
EverYoung has joined #ste||ar
diehlpk_work has joined #ste||ar
<diehlpk_work>
wash[m], yt?
<zao>
pree: Sounds like the program exited with a non-zero exit code. The meaning of it is up to whatever exited.
<zao>
(srun should just passthrough the code)
<zao>
Can you run simpler programs, plain C++, or even something from the system like 'hostname'?
<diehlpk_work>
Can I combine openmp with hpx?
<pree>
This is some sort of error ?? @zao
<zao>
pree: If you have a core dump, you can do a post-mortem debugging by running gdb executable_name corefile_name
<zao>
bt may reveal the call stack of the failure.
<zao>
Nothing else in your output?
<pree>
I ran simple program just printing my name. works fine for ./a.out
shoshijak has quit [Ping timeout: 240 seconds]
<pree>
okay I will try gdb
<pree>
corefile_name ??
<zao>
When a program crashes and coredumps are enabled, it will result in a "core file", a file containing the state of the program and its memory.
<zao>
With it you can interact with the debugger as if it was attached at the final moment of the process life.
<pree>
zao : Thank you
<zao>
Where they end up varies depending on how the system is configured and if the ulimit for their maximum size is large enough.
<zao>
Typically in the directory you ran from.
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
<david_pfander>
Hi, how do I set the CXX and CC compiler for HPX, hpx seems to ignore CMAKE_C_COMPILER (and its CXX version) as a parameter to cmake?
<david_pfander>
(verified that it works in general by compiling another cmake project with that compiler (Vc))
shoshijak has joined #ste||ar
<aserio>
david_pfander: where are you building?
<david_pfander>
aserio: complicated, on a HLRS machine (supercomputing center here in Stuttgart)
<david_pfander>
as tave is down
<david_pfander>
BTW cmake version is 3.0.2
<aserio>
and it is ignoring the compiler that you pass to it?
<david_pfander>
aserio: I think I found the solution
<aserio>
:D its amazing what I can do for people
<david_pfander>
aserio: toolchain file overwrites it, total forgot about that :(
<aserio>
:p
<zao>
Well done, aserio!
<david_pfander>
aserio: thanks, Adrian! :)
<aserio>
all in a days work!
shoshijak has quit [Ping timeout: 240 seconds]
ArashA has joined #ste||ar
bikineev has joined #ste||ar
shoshijak has joined #ste||ar
ajaivgeorge has joined #ste||ar
ajaivgeorge has quit [Ping timeout: 240 seconds]
shoshijak has quit [Ping timeout: 240 seconds]
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 240 seconds]
K-ballo1 is now known as K-ballo
ajaivgeorge has joined #ste||ar
Matombo has quit [Ping timeout: 240 seconds]
ajaivgeorge_ has joined #ste||ar
ArashA has quit [Quit: This computer has gone to sleep]
<pree>
Is anyone know where is my mentor parsa amini ?
ajaivgeorge has quit [Ping timeout: 240 seconds]
ArashA has joined #ste||ar
shoshijak has joined #ste||ar
<jbjnr>
pree: well he's at LSU normally, but if he's not on IRC then he might be away from the office or something. When harmut (hkaiser) reapears, ask him, their offices are only a few doors aart
<github>
[hpx] hkaiser created detect_cpp_mode (+1 new commit): https://git.io/vH963
<github>
hpx/detect_cpp_mode f218738 Hartmut Kaiser: Refining C++ language mode detection for MSVC...
<github>
[hpx] hkaiser opened pull request #2694: Refining C++ language mode detection for MSVC (master...detect_cpp_mode) https://git.io/vH96W
aserio has joined #ste||ar
pree has quit [Ping timeout: 260 seconds]
pree has joined #ste||ar
bikineev has quit [Remote host closed the connection]
eschnett has quit [Quit: eschnett]
shoshijak has quit [Quit: Ex-Chat]
shoshijak has joined #ste||ar
shoshijak has quit [Client Quit]
shoshijak has joined #ste||ar
denis_blank has joined #ste||ar
eschnett has joined #ste||ar
shoshijak has quit [Ping timeout: 240 seconds]
K-ballo1 has joined #ste||ar
Matombo has joined #ste||ar
K-ballo has quit [Ping timeout: 255 seconds]
K-ballo1 is now known as K-ballo
hkaiser has quit [Quit: bye]
mcopik_ has quit [Ping timeout: 246 seconds]
jaafar has quit [Ping timeout: 240 seconds]
bikineev has joined #ste||ar
bikineev has quit [Ping timeout: 255 seconds]
hkaiser has joined #ste||ar
<heller_>
hkaiser: hey
<hkaiser>
hey heller_
<hkaiser>
welcome back ;)
<heller_>
thanks ;)
<heller_>
lucky you
<hkaiser>
am I?
<heller_>
that I am back :P
<hkaiser>
ahh - absolutely!
<heller_>
we are back too early ...
<heller_>
... kid got sick so we had to abort
<heller_>
jbjnr: copy that.
<hkaiser>
sorry...
<heller_>
no problem ;)
<heller_>
we probably went to the less east german place in the world ;)
<hkaiser>
heh
<hkaiser>
Bonn?
<heller_>
na, it was called "Kultuinsel Einsiedel"
<heller_>
supposedly a theme park without commercialism
<hkaiser>
'supposedly'?
<heller_>
well, they selled their own brand, of course ;)
<hkaiser>
ahh it's in the 'deep east' - how did you survive?
<heller_>
but yeah, no fast food or somesuch, all self made, some place near görlitz, it was very nice indeed
<heller_>
it was very very nice
<hkaiser>
good
<heller_>
slogan: "wo grenzen fliessend sind, man kann sogar drin schwimmen" ;)
pree has quit [Ping timeout: 246 seconds]
bikineev has joined #ste||ar
<heller_>
anyway ... next event coming as well ... have to go to marburg tomorrow until wednesday night ...
<heller_>
so, all fun and no play for me
pree has joined #ste||ar
<jbjnr>
hpx::init: std::exception caught: Requested AGAS host (148.187.45.207) not found in node list
<jbjnr>
heller_: copy what?
<heller_>
jbjnr: that you put me down as collaborator
<jbjnr>
hkaiser: the error above - what might have gone wrong ?
<jbjnr>
heller_: yes. We are having a group reassignment and I have to list my collaborators etc. I ended up just putiting "stellar group" and listing us all
<hkaiser>
hpx::init looks through the list of nodes participating in the application
<jbjnr>
hope you had a nice break
<jbjnr>
been quiet here withut you
<hkaiser>
that's either picked up from the nodefile or the environment or the command line
<hkaiser>
and the AGAS server is not in that list
<heller_>
jbjnr: break not fully over yet
<jbjnr>
hkaiser: it's a simple srun -n 1 hello_world
<hkaiser>
misconfigured system?
<jbjnr>
I guess it measnd we beoke something dundamental
<jbjnr>
^broke
<hkaiser>
different network interfaces?
<heller_>
jbjnr: but my fingers can't sit still ... this reference counting bug is haunting me at night. hard to to anything without electricity though
<jbjnr>
it's our resource partitioner branch after merging to master
<jbjnr>
with master ...
<hkaiser>
heller_: that PR I mentioned reverts the crutial changes
<jbjnr>
heller_: did you see the paper - getting so close to MPI now ...
<heller_>
jbjnr: had a quick look, yes, the graphs look just amazing
<hkaiser>
we've been living with the zombie-map for a while, so it will work for another couple of days
<jbjnr>
I might try implementing the one endpoint per thread thingy to see how much more I can squeeze out
<hkaiser>
jbjnr: could we consolidate things first, please?
<heller_>
hkaiser: the zombie map is probably just a cosmetic change
<jbjnr>
hkaiser: sure
<hkaiser>
heller_: I'm sure that's what was causing the ref-counting issues
<jbjnr>
I've got real work to do, so I won't really do much more on it for a bit
<hkaiser>
jbjnr: sure, no worries
<pree>
hkaiser : Sorry to interept. I'm not able to contact parsa yet. What to do ?? Thank you
<heller_>
hkaiser: ok, you reverted more than that though
<heller_>
hkaiser: did you run the tests?
<hkaiser>
pree: I told him to talk to you today
<hkaiser>
aserio: ^^
jaafar has joined #ste||ar
<hkaiser>
yes
<pree>
Okay. Thank you
<hkaiser>
couldn't reproduce the issue, though - not before not after the change
<heller_>
this looks like it is not going out of scope in time though
<heller_>
can't reproduce it here either
<hkaiser>
heller_: I believe that the new logic you implemented in parcel_await does things more than once for certain parcels
<hkaiser>
id-splitting done twice leads to this effect
<heller_>
*nod*
<heller_>
odd that I can't reproduce it though
<hkaiser>
happens only if more than one parcel is sent concurrently
<heller_>
ok
<heller_>
could reproduce it now
<heller_>
good
<heller_>
the more threads per locality, the higher the chances
<heller_>
jbjnr: where did you get the information about GB notification from?
<hkaiser>
what info?
<heller_>
that the notifications will be sent out some time around the 15th
<heller_>
my info is second week in june
<hkaiser>
well, that sounds sensible as paper notifications are due that day as well
<heller_>
june 15th is the date for the regular SC papers
<heller_>
and he said GB ones will be after
<hkaiser>
heller_: well, some GB papers can go to the main program if rejected, so this is connected somehow
aserio has quit [Quit: aserio]
<heller_>
yeah
<heller_>
we didn't submit to the main track to close that loophole :P
<hkaiser>
we didn't because the deadline for this was 2 weeks earlier
<wash[m]>
jbjnr: hey, what's up?
<hkaiser>
wash[m]: how's the GSoC student doing?
<wash[m]>
Hey, pretty good. I will send you an update by mail this evening
<denis_blank>
hkaiser: I'm thinking about adding a HPX_HAVE_EXPRESSION_SFINAE define. However it's difficult to test a particular compiler for full expression sfinae support and I would prefer to go the way of hardcoding the compiler version. However, (for reason) this isn't the way the other feature tests are going, do you have an idea about this?
<hkaiser>
denis_blank: otoh, the solution using a workaround should work for conforming compilers as well, shouldn't it?
<heller_>
hkaiser: #2692 still has the problem
<hkaiser>
denis_blank: even more as there are compilers like nvcc which are barely c++11
<heller_>
hkaiser: re SC paper, yes, I was just trying to find other excuses ;)
<hkaiser>
heller_: ok, so it is something else, apparently
<heller_>
yeah
Matombo has quit [Remote host closed the connection]
<hkaiser>
heller_: well, at least you can use this PR for your fixes :-P
<heller_>
yes
<heller_>
i guess the best thing for now would be to revert the last PR
<heller_>
completely. and work on a proper fix for the problem step by step
<hkaiser>
do we know that #2619 actually causes this problem?
<heller_>
good question ;)
<heller_>
let me check
<hkaiser>
the rest of #2619 was related to th eLF pp, no?
<denis_blank>
hkaiser: Yes, the solution using no heavy expression sfinae should build on all compilers, however GCC and Clang both support the expression heavy version which is better for performance reasons (as explained today). Probably I need to test icc and nvcc as well.
<hkaiser>
it could have been #2658 as well
<heller_>
more or less, yes
<hkaiser>
denis_blank: don't worry about optimization at this point, really
<heller_>
hkaiser: I'll have a look
<denis_blank>
hkaiser: Yes, however I recognized that the optimization I applied works on GCC and Clang without any issues already, so I would prefer the way of just removing it from MSVC builds, if that's ok.
<hkaiser>
denis_blank: gcc 4.9?
<denis_blank>
hkaiser: I was using gcc 5.4 for development, I'll test gcc 4.9 shortly
<K-ballo>
don't we have one?
<heller_>
hkaiser: bisecting now...
<K-ballo>
a feature-test, based on the forms of SFINAE that we know are ok to use with our targets
<hkaiser>
K-ballo: for nvcc we don't use feature tests
<hkaiser>
we feature test th eunderlying compiler instead
<K-ballo>
I don't understand, I'll get back to it later
hkaiser has quit [Read error: Connection reset by peer]
jaafar has joined #ste||ar
hkaiser has joined #ste||ar
jaafar has quit [Ping timeout: 240 seconds]
jaafar has joined #ste||ar
denis_blank has quit [Quit: denis_blank]
EverYoun_ has joined #ste||ar
EverYoung has quit [Ping timeout: 246 seconds]
EverYoun_ has quit [Ping timeout: 240 seconds]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]