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/
jaafar has quit [Ping timeout: 255 seconds]
vamatya_ has joined #ste||ar
vamatya_ has quit [Read error: Connection reset by peer]
K-ballo has quit [Quit: K-ballo]
hkaiser has quit [Quit: bye]
<heller> jbjnr: what patch?
pree has joined #ste||ar
pree has quit [Read error: Connection reset by peer]
pree has joined #ste||ar
jaafar has joined #ste||ar
<pree> hey people !
<pree> where is the primary implementation of SYCL in HPx.compute
<pree> It was mentioned it is there
<pree> whether it is there ?
jaafar has quit [Ping timeout: 255 seconds]
pree has quit [Ping timeout: 260 seconds]
pree has joined #ste||ar
<zao> Can't really find anything in hpxcl, so heaven knows :)
<pree> zao: to me ?
<pree> I think SYCL should be in compute not in HPXCL
<pree> : )
<zao> Isn't HPXCL HPX.Compute?
<zao> Or did they merge stuff into HPX actual?
<zao> (I'm out of touch :D)
<pree> HPXCL is independent from HPX .
<pree> HPXCL uses HPX
<pree> but HPX.compute is actually a part of HPX runtime
<pree> for supporting work dispatch to heterogenous systems
K-ballo has joined #ste||ar
hkaiser has joined #ste||ar
pree has quit [Ping timeout: 260 seconds]
pree has joined #ste||ar
<K-ballo> what's the status of #2829 (pack traversal)?
<hkaiser> K-ballo: I was waiting for the dust around the original pack patch to settle
<hkaiser> wanted to avoid even more problems
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 255 seconds]
EverYoung has joined #ste||ar
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoung has quit [Ping timeout: 246 seconds]
EverYoun_ has joined #ste||ar
EverYoun_ has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
EverYoung has joined #ste||ar
EverYoung has quit [Remote host closed the connection]
<zao> Running the test suite repeatedly - "180 - tests.unit.lcos.distributed.tcp.async_continue_cb_colocated (Failed)" some of the runs.
<heller> do you have an error or something?
<zao> Not quite...
<zao> Seems like something got stolen or something?
<zao> (I don't know what it's supposed to test)
<zao> I'll keep on running, just wanted to mention that it seems like a spurious test.
<zao> This is on current master.
<heller> yeah
<heller> the callbacks are too called too late
<heller> looks like there is some synchronization missing
<heller> the callback has to be called after the future of the call is ready. this looks like a genuine bug
<heller> the failure would make sense if there would have been no synchronization at all before the callback flag gets checked
<zao> Single failure in tests.unit.lcos.distributed.tcp.async_cb_remote_client too, https://gist.github.com/zao/23d5a657e07899870eafca9da2ca5631
<heller> probably related
jaafar has joined #ste||ar
<heller> at least, it looks like i have the current failure under control again, my only problem still is the throttle test
<heller> to get that diagnosed, i'll need to update scimitar and adapt it to the RP changes
<zao> Filed an issue for it, for permanence - https://github.com/STEllAR-GROUP/hpx/issues/2935
<pree> zao : how can i reproduce that error, running it in my pc doesn't shows up any error as you mentioned ? Thanks
<pree> * local machine
<zao> pree: I'm just running `ninja tests` multiple times, each of them happened once for five full test suite runs.
<zao> I'm sure you can run individual tests somehow, but I don't know how :)
<hkaiser> heller: just get rid of the throttle stuff in hpx
<hkaiser> problem solved ;)
<pree> zao : whether you are telling running tests mulitple times leads to this probelm such as spurious thread wake up's ? I don't know what is ninja tests though
<pree> thanks
<zao> pree: `ninja` is the build system I use instead of make.
<zao> Building the 'tests' target compiles all the tests if needed, and runs the test suite.
<zao> pree: It's never running multiple things at the same time, just running test suites back-to-back to catch tests that only fail sometimes.
<pree> zao : well understood : ) thanks
jaafar has quit [Ping timeout: 246 seconds]
<heller> hkaiser: can of worms ;)
pree has quit [Quit: AaBbCc]
<heller> hkaiser: having a working scimitar doesn't hurt either ;)
<heller> would be really cool to have pretty printers auto generated with a clang tool
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
<hkaiser> heller: sounds like a gsoc project
<heller> yeah
<hkaiser> take a note before we forget it ;)
<heller> this would also solve the maintenance for the natvis stuff
<zao> Test suite seems rather stable otherwise - https://gist.github.com/zao/dc94b3ec7fcccc224aa1ed3f1c818384
<zao> (over 14 runs)
<heller> zao: thanks! that is very useful
<heller> if we had a bot that's able to trigger those builds (and capture the output) over ssh, would you integrate some of your boxes in them?
<heller> with*
<zao> If this machine remains stable and I can keep the OS, there's definitely a possibility.
<zao> It's my home NAS built on the Ryzen machine, which I've recently moved to Linux to see if it stops hanging.
<zao> No hangs thus far.
<zao> As for integrating with a bot, would need some work to build a script to run. Right now it's just an existing build and a bash script spinning on `ninja tests` into sequential result files.
<zao> (got some additional memory ordered up for it, so I can build better)
<zao> It's going to yawn as a NAS, so it's a good use of the CPUs.
<zao> 8c/16t
diehlpk has joined #ste||ar
EverYoung has joined #ste||ar
EverYoung has quit [Ping timeout: 258 seconds]
diehlpk has quit [Ping timeout: 248 seconds]