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/
<Smasher> zao are you taliking about that cryptosystem that im using?
<Smasher> and i think you reference that n.GetPow(2) with exponentiation without builtin modulo ?
bikineev has quit [Remote host closed the connection]
<github> [hpx] K-ballo opened pull request #2643: Polishing feature tests (master...feature-tests) https://git.io/vHJ4f
<Smasher> since that's asymmetric encryption, you still use something likex^y mod n there, for encryption and decryption. but in the case of that cryptosystem it's more like x^y mod n^2, hence that precalculation of n.GetPow(2)
<Smasher> and since that's n^2 the computing is so freakin expensive
jaafar has quit [Ping timeout: 260 seconds]
<zao> Ah.
<Smasher> hkaiser if i want to cout every nth iteration of hpx::parallel::for_loop_n how do i do this correctly? need any thread synchronization?
mcopik has quit [Ping timeout: 268 seconds]
<hkaiser> in a parallel loop?
<Smasher> yes
<Smasher> i wanna say: "Here is 100 iteration...\nWe are at 200th n now\nHello from i=300\n...."
<K-ballo> you'll have to count iterations yourself I guess, introducing contention
<zao> "how much harm could a teeny weeny atomic do?" :D
<Smasher> K-ballo boost atomic something?
<K-ballo> or std:: or whatever works
<Smasher> gimme a hint
<K-ballo> no, let hkaiser answer
<K-ballo> mine was just a guess
<hkaiser> std::atomic<int>
<hkaiser> will hurt performance, though
<Smasher> it's crap anyway :)
<K-ballo> I know we are in unsequenced land, but theoretically the iteration with i = 300 might be the "first" one, right?
<zao> Send notifications of completed tasks to a microservice in the cloud which aggregates and ...
<hkaiser> right, you can only get the nth executed iteration
<zao> (pay no attention to the man in a space ship)
<K-ballo> ...we are not in Kansas anymore
<zao> You can keep a count of how many times your thing is invoked by counting up an atomic, f.ex.
<zao> But that count is pretty much directly invalidated by progress made in parallel.
<zao> I need to figure out if atomic has the add-and-return-the-old-value function.
<K-ballo> yep
<K-ballo> fetch_add vs add_fetch or something like that
<K-ballo> something _subtle_ like that
<zao> ah right.
<K-ballo> ah, no, I'm thinking of the sync builtins I think
<K-ballo> atomic only returns the old value
vamatya_ has quit [Ping timeout: 260 seconds]
EverYoung has quit [Ping timeout: 260 seconds]
eschnett has joined #ste||ar
<Smasher> i havend made a single run with a release build yet...
<Smasher> i try in debug - works fine
<Smasher> recompile everything in release - wtf nothing works
<K-ballo> that's how UB usually works... though in this case it might be just about anything else too
<Smasher> that's if you have to much magic under the hood
<K-ballo> uh? I meant mistakes, not magic
<Smasher> doesnt hpx in release mode prints any exceptions?
<K-ballo> sure, no different than debug in that aspect
<K-ballo> it won't do assertions, not in plain Release mode at least, not sure about RelWithDebInfo
<Smasher> pain Release mode
<Smasher> Segmentation fault
K-ballo has quit [Quit: K-ballo]
<Smasher> #1 0x00000000660e5305 in std::_Sp_counted_ptr<hpx::parcelset::policies::tcp::receiver*, (__gnu_cxx::_Lock_policy)2>::_M_dispose() ()
<Smasher> from C:\msys64\mingw64\bin\libhpx.dll
<Smasher> *SIGH*
<Smasher> hkaiser any ideas? http://sprunge.us/gBMM
<Smasher> oh, a zombie process?
<hkaiser> shrug
<Smasher> nah, thats mine
<Smasher> loacked in debugger
<Smasher> rebuild boost in release mode?
<Smasher> because my application and hpx are both in release now
hkaiser has quit [Quit: bye]
parsa has joined #ste||ar
denis_blank has quit [Quit: denis_blank]
vamatya_ has joined #ste||ar
vamatya_ has quit [Quit: Leaving]
justwagle has joined #ste||ar
justwagle has quit [Quit: Leaving]
parsa has quit [Quit: Zzzzzzzzzzzz]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 240 seconds]
pree has joined #ste||ar
pree has quit [Client Quit]
david_pf_ has joined #ste||ar
bikineev has joined #ste||ar
pree has joined #ste||ar
<pree> Hi anybody connected to rostam with mosh?
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
<Smasher> good morning ;)
<pree> good evening :)
<Smasher> im having again some strange starting problems
<Smasher> im getting either http://sprunge.us/gBMM or just nothing happens
<Smasher> so frustrating sometimes
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
K-ballo has joined #ste||ar
bikineev has quit [Ping timeout: 246 seconds]
<Smasher> D:/git/hpx/src/hpx_main.cpp:22: undefined reference to `__imp__ZN3hpx16get_config_entryERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEES7_'
<Smasher> what have i failed here again? :)
<K-ballo> not again :/
<K-ballo> that's the only one?
<Smasher> yes
<Smasher> its a small helper executable
<Smasher> hmmhm
<K-ballo> which libstdc++ abi did you end up using?
<Smasher> oh, it could be a slip...
<Smasher> copypaste mistake %)
<Smasher> i copied hpx_setup_target(..) but forgot to modift target
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
david_pf_ has quit [Ping timeout: 260 seconds]
hkaiser has joined #ste||ar
<hkaiser> pree: I'll answer your email later today
parsa has joined #ste||ar
<pree> thanks@ hkaiser
david_pf_ has joined #ste||ar
david_pf_ has quit [Remote host closed the connection]
<github> [hpx] hkaiser closed pull request #2637: Fix typos and erroneous things in examples and tutorials. (master...taeguk-patch-typo-1) https://git.io/v9hnp
ajaivgeorge has joined #ste||ar
pree_ has joined #ste||ar
ajaivgeorge_ has joined #ste||ar
ajaivgeorge has quit [Read error: Connection reset by peer]
pree has quit [Ping timeout: 260 seconds]
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar
eschnett has quit [Quit: eschnett]
ajaivgeorge_ has quit [Quit: ajaivgeorge_]
ajaivgeorge_ has joined #ste||ar
ajaivgeorge_ has quit [Client Quit]
eschnett has joined #ste||ar
mcopik has joined #ste||ar
pree_ has quit [Quit: Page closed]
denis_blank has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
ajaivgeorge has joined #ste||ar
ajaivgeorge has quit [Read error: Connection reset by peer]
ajaivgeorge has joined #ste||ar
pree has joined #ste||ar
ajaivgeorge has quit [Quit: ajaivgeorge]
EverYoung has quit [Ping timeout: 245 seconds]
<K-ballo> LOL boost.context
<zao> Sup?
david_pf_ has joined #ste||ar
<K-ballo> >> Great, that made HPX work again when using Boost.Context on Windows (x64).
<K-ballo> > I suggest you do an extensive testing - Windows is not my target development platform ...
<K-ballo> (I didn't add the dots, those are in the original response)
<zao> So what's the process of removing a library from Boost?
<K-ballo> once a library it's in it's in, and the author may do whatever it wishes to with it
<zao> Boost needs garbage collection.
<hkaiser> K-ballo: nod, funny...
<hkaiser> I'm waiting for a point in time where Boost.Context works for all platform we care about at the same time, at which point I will fork it ;)
<hkaiser> thundergroudon[m: did you have a chance to try the 32bit fixes?
<K-ballo> might take a while.. one step forward, two steps back
<hkaiser> indeed
<hkaiser> next boost version looks promising so far ;)
diehlpk has joined #ste||ar
<heller__> Smasher: always make sure none of your other processes are running
<heller__> hkaiser: lol
<Smasher> heller__ i do
<Smasher> still not a single run in release mode ^_^
<Smasher> i will finish my db suff and will try release mode again
<Smasher> i need it to measure performance
pree has quit [Ping timeout: 260 seconds]
pree has joined #ste||ar
pree has quit [Ping timeout: 260 seconds]
pree has joined #ste||ar
<thundergroudon[m> hkaiser: Just finished trying now, the fixes seem to work
<thundergroudon[m> Thanks hkaiser
david_pf_ has quit [Remote host closed the connection]
<hkaiser> thundergroudon[m: ok, thanks
<github> [hpx] hkaiser deleted fixing_32bit_msvc at d5058c5: https://git.io/vHU3s
bikineev has joined #ste||ar
EverYoung has joined #ste||ar
diehlpk has quit [Ping timeout: 260 seconds]
bikineev has quit [Remote host closed the connection]
denis_blank has quit [Quit: denis_blank]
shoshijak has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
<Smasher> what would be the best method to change from debug to release?
<Smasher> would it be sufficent to make sure my binaries are getting linked to release versions of hpx binaries?
EverYoung has quit [Ping timeout: 246 seconds]
pree has quit [Ping timeout: 260 seconds]
<zao> Smasher: If you want to use a particular flavor of HPX, you should use the flags exposed by that flavor of HPX, and link the corresponding libraries in that tree.
<zao> Typically you'd use the CMake or PkgConfig things to get the right flags and libs in.
<Smasher> ?
<Smasher> use the flags exposed by that flavor of HPX?
<Smasher> all?
<Smasher> i used to to just to build hpx with -DCMAKE_BUILD_TYPE=Debug
<zao> Most things are encoded in the defines.hpp and friends.
<Smasher> and the same for my project
<zao> Continue.
<Smasher> hmm, headers
<Smasher> im not sure if i can just just invoke make install then
<Smasher> or should i clean up my system first
<zao> Installed in system locations, did we? :)
<zao> Excellent way to accidentally pick up the wrong headers/libs or create dangerous mixes.
pree has joined #ste||ar
justwagle has joined #ste||ar
justwagle has quit [Client Quit]
david_pf_ has joined #ste||ar
<zao> Smasher: I'd strongly recommend -DCMAKE_INSTALL_PREFIX
<zao> And for finding, -DCMAKE_PREFIX_PATH or whatnot.
<Smasher> Installed in system locations, did we? :)
parsa has joined #ste||ar
<Smasher> :DDDDD YES
<Smasher> not i am regreting this
jaafar has joined #ste||ar
<Smasher> not = now
<zao> Reasonably easy to clean.
<zao> Clear out include/hpx, lib/libhpx*, lib/hpx/, and I guess look for pkgconfig/cmake files.
<zao> Might be good to make a install somewhere and see what you end up with in the target tree.
<Smasher> and i also have bin/libhpx*.dll
<Smasher> i will clean this mess in a moment
<Smasher> i have a strange issue with a whared pointer
<Smasher> i have a function GetEncryptionModulus() which returns BigInteger&, i.e. a reference, right?
<Smasher> return *this->encryptionModulus;
<Smasher> but the comment says: @return a read-only reference to the encryption modulus
<Smasher> how is it read only if thats a reference?
shoshijak has quit [Ping timeout: 246 seconds]
<Smasher> zao ^
<zao> Either documentation sucks, APIs prevent use of const due to shitty C reasons, or it's all wrong.
<zao> "read-only" as in if you mutate it, it's your fault.
<Smasher> im doing this
<Smasher> shared_ptr<BigInteger> encryptionModulus = make_shared<BigInteger>(paillier->GetEncryptionModulus());
<Smasher> that should work, right?
<zao> Yup.
<zao> Unless BigInteger's ctor does something naughty with references.
<zao> Don't remember if make_shared forwards properly or if it copies.
<Smasher> i use it like this afterwards
<Smasher> PaillierCiphertext(BigInteger(d_kj[l], 16), encryptionModulus)
<zao> Intent should be to make a copy via construction into a heap-allocated object held by shared_ptr.
<Smasher> and that ciphertext constructor initializes a field with that param
<Smasher> i.e. encryptionModulus(encryptionModulus)
<Smasher> when i debug and look at that object it's pointer is 0x00
<zao> If you're in release, you can't trust what the debugger says for many things.
<zao> Can't really say much here without source, a bit too paraphrased.
<Smasher> im not i release yet
shoshijak has joined #ste||ar
parsa has quit [Read error: Connection reset by peer]
jaafar has quit [Ping timeout: 246 seconds]
pree has quit [Ping timeout: 260 seconds]
david_pf_ has quit [Remote host closed the connection]
shoshijak has quit [Ping timeout: 246 seconds]
bikineev has joined #ste||ar
bikineev has quit [Remote host closed the connection]
bikineev has joined #ste||ar
parsa has joined #ste||ar
patg has quit [Read error: Connection reset by peer]
denis_blank has joined #ste||ar
parsa has quit [Quit: Zzzzzzzzzzzz]
parsa has joined #ste||ar