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/
<K-ballo>
it will do nested-bind evaluation and placeholder propagation
<hkaiser>
yah
<K-ballo>
util::protect isn't enough, that only covers nested bind evaluation
<hkaiser>
what can we do instead?
<K-ballo>
we can't use bind with user provided arguments, not directly
<K-ballo>
I suppose we could extend util::protect to work with placeholders (needs modifying bind as well), and then protect every single user provided argument given to bind
<K-ballo>
-- HPX_WITH_MAX_CPU_COUNT=64
<K-ballo>
-- HPX_WITH_MORE_THAN_64_THREADS=(
<github>
[hpx] K-ballo created more-than-64-threads-default (+1 new commit): https://git.io/vdNut
<github>
hpx/more-than-64-threads-default 5ee529f Agustin K-ballo Berge: Fix HPX_WITH_MORE_THAN_64_THREADS default value
<github>
[hpx] K-ballo opened pull request #2969: Fix HPX_WITH_MORE_THAN_64_THREADS default value (master...more-than-64-threads-default) https://git.io/vdNun
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
parsa has joined #ste||ar
hkaiser has quit [Quit: bye]
parsa has quit [Client Quit]
parsa has joined #ste||ar
parsa has quit [Client Quit]
K-ballo has quit [Quit: K-ballo]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 264 seconds]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 252 seconds]
gedaj has quit [Read error: Connection reset by peer]
gedaj_ has joined #ste||ar
parsa has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 240 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
gedaj_ has quit [Read error: Connection reset by peer]
gedaj__ has joined #ste||ar
gedaj__ has quit [Remote host closed the connection]
gedaj_ has quit [Remote host closed the connection]
gedaj__ has joined #ste||ar
parsa has quit [Ping timeout: 240 seconds]
parsa has joined #ste||ar
gedaj__ has quit [Quit: Leaving]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 252 seconds]
gedaj has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
thundergroudon[m has quit [Ping timeout: 248 seconds]
taeguk[m] has quit [Ping timeout: 255 seconds]
thundergroudon[m has joined #ste||ar
taeguk[m] has joined #ste||ar
<jbjnr>
hkaiser: "how is mine different" - well I've tried to explain many times already - the async execution compiles fine, there is no problem. The problem is the args that are passed in to async_execute - try this https://gist.github.com/biddisco/aeaaa9c58069a41ba2a60d47131c0c13 and you will see that the arguments to async_execute are totally different from the simle void future
<jbjnr>
we now have two args instead of 1 and that's why my numa function test fails. it expects one argument - the future (or unwrapped one) from the continuation, but it gets two argumant instead
<K-ballo>
uh.. entry point must be defined.. I'm doing something wrong
<jbjnr>
when async_execute is called from a normal async call, we do not see these multiple args. then the numa_hint_function compiles and runs as expected.
<K-ballo>
why wouldn't hpx_main.hpp define main?
<jbjnr>
it should
<K-ballo>
works when #including hpx/hpx.hpp before hpx_main.hpp
<K-ballo>
scratch that, I was testing a different code path
<K-ballo>
may need some #define that I don't have or something..
<K-ballo>
I only see one definition of main, in a file that's not being included, and no other declaration
<jbjnr>
the async produces quite sensible args. the continuation is not useful for me directly, unless I somehow specialize my numa hint for all that extra guff that I'm getting for free.
<jbjnr>
anyway hkaiser is not around so I might as well not try to explain.
<jbjnr>
K-ballo: hashtag - worksforme
<K-ballo>
pity, I wanted to finally be able to run your branch
<K-ballo>
the example doesn't work, fails to link some guided executor internals
<K-ballo>
I mean the original example
<jbjnr>
try the minimal example in the git s above
<jbjnr>
gist^
<K-ballo>
doesn't work, missing main
<jbjnr>
it's essentially the same issue
<K-ballo>
is what I've been trying to do
<jbjnr>
lo
<jbjnr>
lol
<K-ballo>
maybe I should try master instead....
<jbjnr>
what OS are you using?
<K-ballo>
well master doesn't link either, so at least that's not on you :P
<K-ballo>
windows
<jbjnr>
I'm assuming hk uses windows, and I'm usin linux
<jbjnr>
^aha ok.
<K-ballo>
I recently had to rebuild hpx from scratch, cmake config and all, maybe I broke something
<K-ballo>
uh... my linker has gone crazy
<hkaiser>
jbjnr: ok, cool - nice and small test
<hkaiser>
will look
<K-ballo>
recreated project from scratch, main behaves now
<K-ballo>
jbjnr: yey! now we are in business
<jbjnr>
so if you can explain to me how to get simple arguments in the continuation async like we get in the normal async, then all is well.
<K-ballo>
why would we even have a HPX_HAVE_CXX14_RETURN_TYPE_DEDUCTION ?
<K-ballo>
jbjnr: the arguments you expect are a single function object, correct?
<jbjnr>
yes, a single function, and simple args
<jbjnr>
if you see the async output, it is a simple function, plus undecorated args. the continuation output has a pair type function and the args are wrapped in stuff, and there are twon instead of 1, (or 3 instead of 2 etc etc if you use bigger examples)
<jbjnr>
the problem is that I forward the Ts... into my numa hint function and they don't work for the continuation case.
gedaj has quit [Remote host closed the connection]
<K-ballo>
jbjnr: I believe I misunderstood what your issue was
diehlpk has joined #ste||ar
gedaj has joined #ste||ar
gedaj has quit [Remote host closed the connection]