hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar-group.org | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | This channel is logged: irclog.cct.lsu.edu
K-ballo has quit [Quit: K-ballo]
hkaiser has quit [Quit: Bye!]
Yorlik has joined #ste||ar
mdiers[m] has joined #ste||ar
<mdiers[m]> Could someone maybe still try to reproduce my problem https://github.com/STEllAR-GROUP/hpx/issues/5802?
<john98zakaria[m]> <mdiers[m]> "Could someone maybe still try to..." <- I could try running it on my very humble i5 4460 with 4 cores on gcc
<john98zakaria[m]> Or do you need a fancy monster machine. (I am a noob BTW)
<mdiers[m]> It is important that it is started as multiple processes (srun/mpirun), in your case with 4 processes on one node! (With one process per physical core.)
<john98zakaria[m]> <mdiers[m]> "It is important that it is..." <- Shooting stars everywhere
<john98zakaria[m]> mpirun -n 4 ./lost_ready_future --hpx:ini=hpx.parcel.mpi.enable!=1 --hpx:ini=hpx.parcel.tcp.enable!=0
<john98zakaria[m]> Which means I couldn't reproduce the issue
<mdiers[m]> <john98zakaria[m]> "Shooting stars everywhere..." <- With this, I was able to reproduce it at least somewhat. I also have to run it several times in some cases. On older generations of cpus I don't get it reproduced that well either.
<mdiers[m]> <john98zakaria[m]> " mpirun -n 4 ./lost_ready_future..." <- It is also important to use an optimized version. (release)
<john98zakaria[m]> I am using RelWithDebug so it should be equivalent
hkaiser has joined #ste||ar
K-ballo has joined #ste||ar
hkaiser has quit [Quit: Bye!]
hkaiser has joined #ste||ar
<K-ballo> gonidelis[m]: thoughts on moussaka?
hkaiser has quit [Quit: Bye!]
<john98zakaria[m]> I planning to contribute by creating example code
<john98zakaria[m]> What criteria does example code need to fulfill in order to be accepted into the repository?
<john98zakaria[m]> s/creating/writing/
<gonidelis[m]> K-ballo: what the.... 😅
<gonidelis[m]> who taught you this bad word?
<gonidelis[m]> john98zakaria: what examples code?
<gonidelis[m]> clang-format -i -style=file <all your edited files>
<gonidelis[m]> ?
<john98zakaria[m]> gonidelis[m]: There is an example folder in the repository
<gonidelis[m]> oh you mean examples. ok what kind of example do you plan on writing?
<gonidelis[m]> very nice idea btw ;)
<john98zakaria[m]> My bachelor thesis is on molecular simulation on distributed systems
<john98zakaria[m]> Mainly trying to swap out mpi with hpx
<john98zakaria[m]> So I wanted to contribute some parts of the thesis into the repository as well
<john98zakaria[m]> Mainly focusing on
<john98zakaria[m]> Molecular Dynamics simulation of a Lennard-Jones fluid
<gonidelis[m]> 0.0
<gonidelis[m]> that's amazing
<gonidelis[m]> do you think you can have a standalone, single-file example for that? john98zakaria
<gonidelis[m]> except formattin i dont think we have any other strict standards tbh?
<gonidelis[m]> (no question mark ^^)
<john98zakaria[m]> I don't think so, it would more likely become folder
<john98zakaria[m]> Or I could extract the communication part out
<john98zakaria[m]> And remove the physics from it
<gonidelis[m]> hm...... we have to ask hkaiser for that
<gonidelis[m]> but why don't you publish the project as a standalone github repo and then we can have it on our docs webpage or sth
<john98zakaria[m]> gonidelis[m]: I think that's actually better
<gonidelis[m]> looking forward to it ;)
<john98zakaria[m]> My supervisor had already used hpx on node level parallelism
<john98zakaria[m]> Here is the paper
hkaiser has joined #ste||ar
<Yorlik> Do we have a physicist amongst the HPX people? Maybe someone who knows the perils of simulation? I have to make a very serious technology decision about physics simulation in our HPX empowered gameserver.
<diehlpk_work> Yorlik, We have some astrophysicist, we collaborate with
<Yorlik> diehlpk_work, Do you think such an un-serious matter like a gameserver could be discussed with her/him?
<diehlpk_work> You can always ask Dominic
<Yorlik> Allright - I'll watch out for him.
<diehlpk_work> He is not on irc
<Yorlik> How could I contact him? Do you speak to him / could ask for me?
<hkaiser> Yorlik: Dominic Marcello <dmarce504@gmail.com>
<Yorlik> I'm already writing an email right now.
ruchipakhle has joined #ste||ar
ruchipakhle has quit [Quit: Client closed]
Yorlik has quit [Ping timeout: 252 seconds]