K-ballo 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/
nanmiao11 has quit [Remote host closed the connection]
nanmiao11 has joined #ste||ar
dcbdnl has joined #ste||ar
<dcbdnl>
Hi. Does AGAS or HPX in general allow global objects to reside on disk? So if a computation is being done and there is no more ram, would HPX write some global objects to disk and then bring them back when necessary?
<hkaiser>
ms[m]: no
<hkaiser>
dcbdnl: in theory yes, but we have not fully implemented this
<hkaiser>
dcbdnl: migration can be seen not only as an operation of moving objects in between localities, but also as making them persistent or bringing them back
<khuck>
ms[m]: I eventually got HPX and Phylanx to build with sanitizer flogs... and I got: `201: ==1381527==ERROR: AddressSanitizer: stack-buffer-underflow on address 0x7f5fcbb47a30 at pc 0x0000004800d7 bp 0x7f5fcbb47750 sp 0x7f5fcbb47748 `
<khuck>
which is I think coming from HPX in `src/runtime/parcelset/parcel.cpp:190`
<khuck>
(well, address sanitizer says it is)
<khuck>
although I don't think that's the fatal error I am looking for, so never mind for now
dcbdnl has quit [Quit: dcbdnl]
<diehlpk_work>
khuck, We need to shorten the IDPDS paper. Would you mind to give it a try as a native speaker and shorten sentences?
<khuck>
how much shorter?
<diehlpk_work>
1.5 columns
<khuck>
3/4 a page?
<diehlpk_work>
I removed half a page by compact the authors already
<diehlpk_work>
Yes
<khuck>
how soon?
<diehlpk_work>
<diehlpk_work>
October 12, 2020
<diehlpk_work>
I submitted the abstract today
<khuck>
Yeah, I think so
<diehlpk_work>
Ok, the current version is on Github
<khuck>
the usual LSU paper repo?
<khuck>
github.com:STEllAR-GROUP/documents
<diehlpk_work>
Yes
<khuck>
which directory?
<khuck>
the SC20_octotiger?
<diehlpk_work>
khuck, see pm
dcbdnl has joined #ste||ar
dcbdnl has quit [Client Quit]
hkaiser has joined #ste||ar
<K-ballo>
what did I change that caused warnings from compatibility headers to suddenly start poping up?
<K-ballo>
I haven't changed any includes, so.. I must have messed up the diagnostic disabling mechanism?
<K-ballo>
but still, why do many of them come from building hpx itself?
<hkaiser>
K-ballo: ms[m] actually enabled the warnings just the other day ;-)
<K-ballo>
aah ok
<hkaiser>
K-ballo: hmm, I though that change was already merged, but it wasn't (#4968)
<K-ballo>
I would expect hpx itself to not trigger deprecated header warnings.. we are not boost
<ms[m]>
Yeah, 4968 is also for different deprecation warnings
weilewei has joined #ste||ar
<ms[m]>
K-ballo: my guess is mistakes in the compatibility headers :/ at least at some point I've fixed incorrect ifdefs in them
<ms[m]>
So even more reason to generate them
<K-ballo>
I should probably wait for deprecated includes to be fixed on master
<K-ballo>
oh, the check is inverse
<K-ballo>
#if have_warnings rather than #if warnings_disabled