bobakk3r has quit [Quit: No Ping reply in 180 seconds.]
bobakk3r has joined #ste||ar
hkaiser has quit [Ping timeout: 256 seconds]
araml has joined #ste||ar
V|r has quit [Ping timeout: 255 seconds]
V|r has joined #ste||ar
V|r has joined #ste||ar
V|r has quit [Changing host]
araml has quit [Read error: Connection reset by peer]
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Ping timeout: 256 seconds]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
spaceVStab has joined #ste||ar
spaceVStab has quit [Remote host closed the connection]
hkaiser has joined #ste||ar
<mdiers_>
simberg: yt?
<mdiers_>
simbergm: yt?
nikunj has joined #ste||ar
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar
hkaiser has quit [Ping timeout: 256 seconds]
bita has quit [Read error: Connection reset by peer]
nikunj97 has joined #ste||ar
nikunj has quit [Ping timeout: 268 seconds]
Nikunj__ has joined #ste||ar
nikunj97 has quit [Ping timeout: 256 seconds]
nikunj has joined #ste||ar
Nikunj__ has quit [Ping timeout: 258 seconds]
nikunj has quit [Remote host closed the connection]
nikunj has joined #ste||ar
nikunj has quit [Remote host closed the connection]
hkaiser has joined #ste||ar
giannis has joined #ste||ar
giannis is now known as Guest81362
Guest81362 is now known as gonidelis
<gonidelis>
hello! i am trying to install HPX on my ubuntu16 machine and i have encountered a problem with hwloc on my installation
<zao>
How fun! What happens?
<gonidelis>
i am installing hwloc using `sudo apt install hwloc` but after trying `cmake -DHPX_WITH_MALLOC=system` my cmake pops "Hwloc could not be found..."
<zao>
Do you have libhwloc-dev?
<zao>
Packages on Ubuntu are often split into a executable/runtime-library package and a matching development package with headers/etc.
<gonidelis>
Oh no! Let me try it. Thanks a lot for your response. Yeah I get it...
<gonidelis>
Well thanks a lot! It worked. WOW! Thank you...
<zao>
Yay!
<gonidelis>
well then I try `make all` and I get "CMakeFiles/Makefile2:46816: recipe for target 'libs/threadmanager/CMakeFiles/hpx_threadmanager.dir/all' failed". Do you know what may be the cause?
<zao>
The cause may be further up in your output.
<zao>
Not off-hand, but it was long since I last built HPX.
<gonidelis>
you are right. it actually must be : * error: call of overloaded ‘get()’ is ambiguous*
bita has joined #ste||ar
<gonidelis>
are `tcmalloc` or/and `jemalloc` mandatory in order for `make install` to succeed?
<zao>
I'd expect `system` allocator to be sufficient.
<simbergm>
gonidelis: what compiler are you using? I suspect you need a newer gcc...
<gonidelis>
`gcc --version` outputs: *5.4.0*
<zao>
simbergm: Documentation still documents 4.9 or something.
<simbergm>
grr, we need to update the documentation in that case...
<simbergm>
gonidelis: are you building 1.4.0 or master?
<simbergm>
for master you probably need 7 or higher
<simbergm>
and if you're really on ubuntu 16.04 I recommend you upgrade...
<zao>
simbergm: Hey, my cluster is still 16.04.
<simbergm>
1.4.0/1 should still work with gcc 5
<simbergm>
heh, whatever works for you
<gonidelis>
yeah I am building master. Should I upgrade my gcc I 'll try again and I will inform you on my progress. Thanks a lot.
<simbergm>
yep, in that case gcc 7 or higher, or clang 4 or higher (if I remember correctly)
<gonidelis>
Well I recently reinstalled Ubuntu16. The reason I picked this distro is because there are plenty fixes on the internet as it is older than 18. Do you think I should upgrade in order to contribute to HPX?
hkaiser has quit [Quit: bye]
<zao>
gonidelis: On cluster systems an older OS tends to be "fine" as long as one can provide decent compilers and libraries in a module system on top, as long as drivers and kernel bugs aren't too intrusive.
<zao>
On a workstation, 18.04 is a fine version to be on, somewhat newer versions in the OS itself, desktop environments tend to work better.
<zao>
gonidelis: I'm guessing you intend to apply for GSoC, maybe?
<gonidelis>
well, yes I do...
mdiers_ has quit [*.net *.split]
Yorlik has quit [*.net *.split]
mdiers_ has joined #ste||ar
Yorlik has joined #ste||ar
<zao>
(I'm not, if you're scared of competition :D)
<zao>
You can probably develop on anything, but something that's comfortably above minspec is probably helpful.
<gonidelis>
Well I am installing Ubuntu18 right now as installing gcc-7 didn't help (as expected)
<zao>
Did you specify that it should use the compiler (with `-DCMAKE_C_COMPILER=gcc-7 -DCMAKE_CXX_COMPILER=g++-7` or suchlike)?
<gonidelis>
wasn't aware of this particular directive. my bad.
<zao>
As the distro and all libs are built with one particular compiler, additional compiler packages tend to have alternate names, so that when something uses "gcc" or "g++", they get the normal one and you get to specify explicitly if you need to use a different one.
<gonidelis>
yeah I absolutely get it. Thank you. Actually I am waiting for `make` to complete right now. I wanted to ask something though: you previously said that your cluster runs Ubuntu16 and you mentioned 'workstation' as an alternative. As I am a newcomer and I certainly hope to stay (I love c++ and parallelization/distributes systems is my passion) could
<gonidelis>
you tell me whether any particular hardware is needed in order to contribute to HPX (not just for GSoC)?
<zao>
Most of it is just regular PC development.
<gonidelis>
I have been developing all my code/projects so far on a ThinkPad T430 with an i5-3320 quad core and 8gb ram. Oh I see... So it must be performance tests that you run on clusters right? I mean in order to achieve computational improvement...
<zao>
The TCP parcelport might need some networking, the infiniband parcelport would need infiniband connectivity, there's some side projects that use GPUs.
<zao>
The schedulers might have fun corner cases if you have a more complicated machine layout.
<zao>
Depends a bit on what project you intend to propose.
<zao>
If you're going to build tests/examples on that 8GB machine, you can probably not build with high -j parallelism, as linking them takes a lot of RAM.
<zao>
"my" cluster would be a bit of a misnomer by the way, as it's the one I use at work.
<zao>
The HPX group has access to resources too, but it tends to help if you can do some local development.
<gonidelis>
ok I 'll definately take my hardware into account as far as project selection is concerned. What about contribution in general? Could I become an HPX developer with that kind of hardware?
<gonidelis>
Yeah. I imagined that you will have access as a team on certain resources (that was my next question actually ;p ). that might answer my first one too...
<zao>
I forget how it has gone the previous years, but I'd expect that accepted students would have access to the smaller resources by need (which still has a few nodes in a batch system).
<zao>
A full build of HPX with tests and examples tends to take a lot of time, but actively developing and building parts of it tends to be fine on most setups.
<zao>
I've built HPX a whole bunch of times on my old work macbook, which was along those specs.
<gonidelis>
perfect! I ll try to get familiarized with HPX then, trying to implement some tests / test-code as soon as possible and I will soon provide you with a proposal. Thanks for your help/info!
gonidelis4 has joined #ste||ar
gonidelis4 has quit [Remote host closed the connection]
gonidelis has quit [Remote host closed the connection]
gonidelis has joined #ste||ar
gonidelis has quit [Remote host closed the connection]