hkaiser 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/ | GSoC: https://github.com/STEllAR-GROUP/hpx/wiki/Google-Summer-of-Code-%28GSoC%29-2020
weilewei has quit [Ping timeout: 240 seconds]
Hashmi has quit [Quit: Connection closed for inactivity]
bita has quit [Ping timeout: 260 seconds]
gonidelis has joined #ste||ar
hkaiser has quit [Quit: bye]
diehlpk_work has quit [Quit: Leaving]
gonidelis has quit [Remote host closed the connection]
bita has joined #ste||ar
bita has quit [Quit: Leaving]
<heller1> ms[m]: get well soon!
mdiers_ has quit [Remote host closed the connection]
mdiers_ has joined #ste||ar
kale_ has joined #ste||ar
kale_ has quit [Ping timeout: 256 seconds]
hkaiser has joined #ste||ar
Amy1 has quit [Quit: WeeChat 2.2]
kale_ has joined #ste||ar
kale_ has quit [Remote host closed the connection]
Amy1 has joined #ste||ar
Amy1 has quit [Client Quit]
Amy1 has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
nikunj has quit [Ping timeout: 256 seconds]
weilewei has joined #ste||ar
nikunj has joined #ste||ar
nikunj has quit [Read error: Connection reset by peer]
nikunj has joined #ste||ar
rtohid has joined #ste||ar
kale_ has joined #ste||ar
kale_ has quit [Ping timeout: 264 seconds]
diehlpk_work has joined #ste||ar
kale_ has joined #ste||ar
nikunj97 has joined #ste||ar
kale_ has quit [Ping timeout: 258 seconds]
kale_ has joined #ste||ar
nan11 has joined #ste||ar
rtohid has quit [Ping timeout: 240 seconds]
bita has joined #ste||ar
<diehlpk_work> Is it a bad sign if HPX compiles faster as my lecture slides?
rtohid has joined #ste||ar
kale_ has quit [Ping timeout: 256 seconds]
<hkaiser> diehlpk_work: for HPX?
<diehlpk_work> For my lecture slides?
<hkaiser> heh
<zao> diehlpk_work: I wonder if I have my TikZ-based 3D meshes I generated for my master thesis.
<diehlpk_work> zao, yes tikz is one reason for the long compilation time
<zao> Turns out that even XeLaTeX is not great at rendering 3D meshes, runs out of all sorts of memory after like 1000 triangles.
<zao> Probably not the best thing to perform homogeneous 3D transformations with either.
akheir has joined #ste||ar
gonidelis has joined #ste||ar
gonidelis has quit [Remote host closed the connection]
gonidelis has joined #ste||ar
<hkaiser> zao: so you got your master's for doing 3d simulations with tikz?
<hkaiser> ;-)
<zao> Master was on efficient rendering of crowds of characters in a 3D game, figured I'd have a nice illustration in the thesis and exported an animated mesh straight out of the engine, with the triangles sorted by depth.
<zao> So much effort to not work on the thing I should've been working on :)
<hkaiser> figures
<zao> \begin{figure}
<hkaiser> without \end{figure}, apparently
<hkaiser> diehlpk_work: could you give me the meeting link, pls
<zao> ^^
<gonidelis> may I ask: What is the difference of hpx::futre and std::futre?
<gonidelis> future*
<zao> One is the one specified in the standard (or coming standards), and the other one is the cool improved one in HPX, that's more composable and works in HPX threads and in a distributed manner.
<zao> Regarding your question a few days ago, a compile time of a few seconds for a TU that includes HPX is perfectly normal, there's a lot of headers with complicated code pulled in to pull off the magic.
<gonidelis> TU??
<zao> Translation unit.
<zao> A C++ source file under compilation, with all its includes expanded and all.
<gonidelis> wow !!! Thank you !!!
<zao> One or more TUs are compiled into object files, which are later linked into things.
<zao> The term comes up when talking about visibility and scope.
<gonidelis> yeah I know the theory but I lack terminology-knowledge somehow. Thank you ;p
<gonidelis> Also thank you for responding to a 3-days-ago quesiton. Really apriciate it. As I have not digged into building large stuff yet I was of the impression that the compilation time would take like some millie-seconds.
<zao> A lot of it is due to that as fancy C++ code is template-based, there's a fair bit of complexity for the compiler to work through every time a header is included.
<zao> While in good old C and boring C++, headers tend to contain less as implementations tend to be separately compiled.
K-ballo has quit [Remote host closed the connection]
akheir has quit [Read error: Connection reset by peer]
K-ballo has joined #ste||ar
akheir has joined #ste||ar
rtohid has left #ste||ar [#ste||ar]
akheir has quit [Quit: Leaving]
rtohid has joined #ste||ar
<parsa> hkaiser: ping
<hkaiser> hey
<parsa> see pm
<hkaiser> jbjnr: I have rebased the APEX branch, there is one compilation error now: https://app.circleci.com/pipelines/github/STEllAR-GROUP/hpx/2839/workflows/fb0b2137-af7e-4a37-bd1f-5b5904aa9d56/jobs/176746
<hkaiser> jbjnr: would you mind having a look, please?
<hkaiser> there are two tests failing as well, but those should be fixed on a separate PR, already
gonidelis has quit [Remote host closed the connection]
Nikunj__ has joined #ste||ar
nikunj97 has quit [Ping timeout: 260 seconds]
<bita> hkaiser, I have changed perftests, but still failing. Do you have any advice? https://github.com/STEllAR-GROUP/phylanx/pull/1143/commits/547f3a7be1fcc6d66f7b64112f82cb3687cce089
<bita> In ^ cmake seems to work in making 3 files but both num_localities() and hpx::get_num_localities() see only 1 locality
<hkaiser> bita: sec
<hkaiser> bita: the cmake file looks ok
<Yorlik> Why in hell is this allowed? I wish it would throw: https://godbolt.org/z/nxPXeW
<Yorlik> (std::variant)
<bita> hkaiser, it does not get num_localities as the one that is set in the cmake. Would it be resolved if we run it on Rostam?
<K-ballo> Yorlik: why would it throw?
<Yorlik> It is initialized to nothing. it pick a default type and a default value. I consider this dangerous.
<Yorlik> Sure - you could see it as a feature in certain use cases ...
<zao> Yorlik: The wonderful debates of never-empty variants or not.
<Yorlik> Yup. I just found my position in that debate - lol.
<Yorlik> Had some wonderful bigs due to it. Stuff just worked so nicely until runtime.
<Yorlik> bugs.
<zao> It's even documented behaviour :P
<Yorlik> NP - I'll just implement a .validate something. But youz can really not assume anything ever - :D
<zao> You can probably make it behave like you want by having a non-default-constructible "empty" 0th type.
<zao> On the nullary ctor - "This overload only participates in overload resolution if std::is_default_constructible_v<T_0> is true."
<Yorlik> Makes sense.
<K-ballo> if your variant has meaning you'd want to wrap it anyhow
<Yorlik> It is wrapped. It's internal code onside the class using it
<Yorlik> I use it as a variant of different shared_ptr types to different node types in a quadtree.
<Yorlik> leaves vs. branching nodes
<Yorlik> Ya - nice one
<Yorlik> But I need bad default constructed variants. I'll just use a void* and check for 0
<Yorlik> I kinda ssumed i'd get the -1 index
<Yorlik> Wrong thought, obviously.
<hkaiser> bita yah, that might help
<hkaiser> could be that the testing infrastructure passes the number of localities only if run via slurm
<bita> Okay, thanks. I am writing a simpler test for num_localities()
<hkaiser> bita: nice!
<bita> apparently I have missed one ${param}! It is working now, I will check and push it
<hkaiser> bita: thanks
<bita> thank you
rtohid has left #ste||ar [#ste||ar]
bita has quit [Ping timeout: 260 seconds]
Nikunj__ has quit [Ping timeout: 260 seconds]
bita has joined #ste||ar
bita has quit [Ping timeout: 260 seconds]