#ste||ar on 2020-02-24 — irc logs at irclog.cct.lsu.edu

2019-12-03 02:04 hkaiser changed the topic of #ste||ar to: The topic is '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/

03:12 hkaiser has quit [Quit: bye]

07:05 nikunj has quit [Remote host closed the connection]

07:10 nikunj has joined #ste||ar

08:01 nikunj has quit [Read error: Connection reset by peer]

08:01 nikunj has joined #ste||ar

08:54 Guest49223 has joined #ste||ar

08:56 nikunj has quit [Ping timeout: 272 seconds]

08:57 nikunj has joined #ste||ar

09:00 nikunj has quit [Read error: Connection reset by peer]

09:02 nikunj has joined #ste||ar

09:02 nikunj97 has joined #ste||ar

09:24 nikunj has quit [Read error: Connection reset by peer]

09:24 nikunj has joined #ste||ar

09:28 nikunj97 has quit [Ping timeout: 240 seconds]

09:28 nikunj has quit [Ping timeout: 252 seconds]

09:29 nikunj97 has joined #ste||ar

09:32 Nikunj__ has joined #ste||ar

09:35 nikunj97 has quit [Ping timeout: 272 seconds]

09:35 nikunj97 has joined #ste||ar

09:38 Nikunj__ has quit [Ping timeout: 260 seconds]

10:21 nikunj97 has quit [Ping timeout: 240 seconds]

11:52 nikunj has joined #ste||ar

11:54 nikunj has quit [Read error: Connection reset by peer]

11:55 nikunj has joined #ste||ar

12:02 nikunj has quit [Ping timeout: 258 seconds]

12:02 nikunj97 has joined #ste||ar

12:03 nikunj has joined #ste||ar

12:03 Nikunj__ has joined #ste||ar

12:03 nikunj has quit [Read error: Connection reset by peer]

12:04 nikunj has joined #ste||ar

12:07 nikunj97 has quit [Ping timeout: 258 seconds]

12:59 Guest49223 has quit [Ping timeout: 268 seconds]

13:07 hkaiser has joined #ste||ar

13:08 V|r has joined #ste||ar

13:23 <mdiers_> hello, unfortunately i still have a performance problem. i modified the hello_world example for this. and put it here: https://gist.github.com/m-diers/10e0bed1ac978c7f27713c7767f45acf i have recreated the problem in work_load. since it is an iterative algorithm, the structure of work_load cannot be changed very well. can someone take a look at it and give me some recommendations?

13:30 ronniegandhi has joined #ste||ar

13:38 <hkaiser> mdiers_: did you measure the idle-rates?

13:53 <mdiers_> hkaiser: the idle rate of the systems? the idle-rate of the system is near zero.

13:55 <hkaiser> mdiers_: the hpx idle-rate performance counter

14:04 K-ballo has joined #ste||ar

14:12 K-ballo has quit [Ping timeout: 260 seconds]

14:12 <mdiers_> hkaiser: /threads/idle-loop-count/instantaneous that is the only idle counter?

14:12 <hkaiser> no

14:12 <hkaiser> not that one

14:13 <hkaiser> sec

14:13 <hkaiser> here: /threads/idle-rate

14:13 <hkaiser> that requires rebuilding hpx with -DHPX_WITH_THREAD_IDLE_RATES=On however

14:14 <mdiers_> hkaiser: /threads{locality#*/total}/idle-rate that is the right one, but it is away ;-( I will do a rebuild

14:14 <hkaiser> mdiers_: the idle-rate counter gives you a nice aggregated measure of how well your code is parallelized

14:22 <mdiers_> hkaiser: its ot: how is the flag to build shared libs?

14:41 hkaiser has quit [Quit: bye]

14:59 K-ballo has joined #ste||ar

15:04 hkaiser has joined #ste||ar

15:22 diehlpk_work has joined #ste||ar

15:33 Nikunj__ has quit [Ping timeout: 260 seconds]

15:38 nikunj has quit [Ping timeout: 258 seconds]

15:42 nikunj has joined #ste||ar

15:45 nikunj has quit [Read error: Connection reset by peer]

15:46 nikunj has joined #ste||ar

15:58 nikunj has quit [Read error: Connection reset by peer]

15:58 nikunj has joined #ste||ar

16:02 mdiers_ has quit [Quit: mdiers_]

16:08 mdiers_ has joined #ste||ar

16:08 mdiers_ has quit [Client Quit]

16:08 mdiers_ has joined #ste||ar

16:12 mdiers_ has quit [Client Quit]

16:12 mdiers_ has joined #ste||ar

16:13 mdiers_ has quit [Client Quit]

16:13 mdiers_ has joined #ste||ar

16:13 mdiers_ has quit [Client Quit]

16:14 mdiers_ has joined #ste||ar

16:14 mdiers_ has quit [Client Quit]

16:15 mdiers_ has joined #ste||ar

16:16 <mdiers_> hkaiser updated the gist with idle-rates (0.01%)

16:18 nikunj has quit [Read error: Connection reset by peer]

16:18 nikunj has joined #ste||ar

16:23 nikunj has quit [Ping timeout: 272 seconds]

16:24 nikunj has joined #ste||ar

16:28 nikunj has quit [Remote host closed the connection]

16:30 nikunj has joined #ste||ar

16:44 nikunj has quit [Ping timeout: 265 seconds]

16:46 nikunj has joined #ste||ar

16:47 nikunj97 has joined #ste||ar

16:51 <hkaiser> mdiers_: can you re-link the gist here?

16:53 weilewei has joined #ste||ar

16:54 <weilewei> simbergm can you please update spack tag for hpx 1.4.1? Or let me know where can I obtain sha256 hash for the commit?

16:55 <weilewei> https://github.com/spack/spack/blob/94acd8f235d8131939a2485dae7fc98177b87901/var/spack/repos/builtin/packages/hpx/package.py#L18

16:56 <hkaiser> weilewei: Rod is responsible for Spack, he maintains the package for HPX

16:57 <hkaiser> please get in contact with him to make it happen.

16:57 <weilewei> hkaiser thanks! I will contact Rod instead. simbergm please miss the message

17:28 ronniegandhi has quit [Ping timeout: 255 seconds]

17:55 weilewei has quit [Remote host closed the connection]

18:03 parsa has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]

18:03 parsa has joined #ste||ar

18:14 nikunj97 has quit [Ping timeout: 255 seconds]

18:14 Nikunj__ has joined #ste||ar

18:15 nikunj97 has joined #ste||ar

18:19 Nikunj__ has quit [Ping timeout: 240 seconds]

18:19 Nikunj__ has joined #ste||ar

18:21 nikunj97 has quit [Ping timeout: 255 seconds]

18:35 nikunj97 has joined #ste||ar

18:37 Nikunj__ has quit [Ping timeout: 240 seconds]

18:59 <nikunj97> has anyone tried building boost with armclang?

18:59 <nikunj97> It says unsupported toolset when I'm trying to build it

19:00 <nikunj97> K-ballo ^^

19:01 <zao> What platform is that on?

19:01 <nikunj97> I'm on hisilicon

19:01 <zao> If you're cross-compiling, are you using something like the cmake/toolchains/ we have?

19:02 <nikunj97> I'm not cross-compiling, I'm compiling on aarch64

19:02 <nikunj97> with armclang

19:02 <nikunj97> eventually I want to build hpx

19:02 <nikunj97> but I can't get past boost as of now

19:02 <zao> ahahaha

19:03 <zao> Sorry :D

19:03 <zao> Boost is amusing with its ideas of knowing things about the target.

19:03 <zao> What library is it that is bailing out?

19:03 <nikunj97> it just won't take in armclang as a toolset

19:03 <nikunj97> it's failing at the bootstrap level

19:04 <zao> Is it supposed to?

19:04 <nikunj97> when I try gcc on bootstrap and force armclang on b2 then I get error about no .jam file found

19:04 <nikunj97> zao, I don't suppose it should

19:04 <nikunj97> Unknown toolset: armclang

19:04 <zao> I'd expect to use the "clang" toolset and explicitly specify the name of the compiler binaries in *-config.jam

19:04 <nikunj97> this is in boostrap.log

19:05 <nikunj97> armclang is supposed to be arm-hpc-compiler something

19:05 <nikunj97> it was a separate module on the cluster

19:05 <zao> Did you see this wiki page? https://gitlab.com/arm-hpc/packages/-/wikis/packages/boost

19:05 <nikunj97> it has llvm backend and a different frontend

19:05 <zao> Seems to suggest to bootstrap with GCC and use "clang" with an override in user-config.jam

19:05 <nikunj97> so I thought armclang made sense

19:06 <nikunj97> ohh thanks for the link!

19:06 <nikunj97> I should've seen that

19:07 <nikunj97> I'm using 1.65 though

19:07 <nikunj97> let's see if it works. Thanks a lot for your expertise ;)

19:11 <nikunj97> zao, finally some compilation happening! Thanks again :D

19:12 <zao> \o/

19:31 hkaiser has quit [Quit: bye]

19:34 <nikunj97> Performing Test HPX_WITH_BUILTIN_INTEGER_PACK - Failed. What does the test do?

19:45 nikunj97 has quit [Quit: Leaving]

19:50 nikunj97 has joined #ste||ar

19:55 <diehlpk_work> nikunj, You could look into the Fedora spec file for boost

19:56 <diehlpk_work> They build for arm64 and arm32 and their package works for me

19:56 <nikunj97> diehlpk_work, I keep that in mind next time

20:00 <nikunj97> diehlpk_work, the spec file seems to be working with gcc as the toolset

20:01 <nikunj97> diehlpk_work, see https://src.fedoraproject.org/rpms/boost/blob/master/f/boost.spec#_679

20:03 <nikunj97> I wanted to build with armclang and not gcc. There's an extra step with armclang as zao pointed out.

20:03 <nikunj97> zao, it's nice to see HPX getting build and run examples without returning any errors ;)

20:04 <nikunj97> I can now say that HPX builds on hisilicon1616 as well!

20:09 Nikunj__ has joined #ste||ar

20:12 nikunj97 has quit [Ping timeout: 272 seconds]

20:19 nikunj97 has joined #ste||ar

20:21 hkaiser has joined #ste||ar

20:23 Nikunj__ has quit [Ping timeout: 260 seconds]

20:27 nikunj97 has quit [Quit: Leaving]

20:42 diehlpk has joined #ste||ar

20:43 <diehlpk> hkaiser, Blaze 3.7 is available in the upcoming Fedora release

20:43 <diehlpk> hkaiser, Can you please add Google Summer of Code to the irc welcome message?

20:44 <diehlpk> | GSoC: https://github.com/STEllAR-GROUP/hpx/wiki/Google-Summer-of-Code-%28GSoC%29-2020

20:46 <diehlpk> https://github.com/STEllAR-GROUP/hpx/wiki/Google-Summer-of-Code-%28GSoC%29-2020#task-bench-a-parameterized-benchmark-for-evaluating-parallel-runtime-performance

20:46 <diehlpk> We do not have any mentor and additonal information there.

20:46 <diehlpk> Who is the mentor for this one?

20:46 <diehlpk> simbergm, jbjnr ?

20:46 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

20:48 <hkaiser> diehlpk: it was John's idea

20:48 <diehlpk> Ok, so I will add his name as a mentor

20:59 diehlpk has quit [Ping timeout: 260 seconds]

21:38 Abhishek09 has joined #ste||ar

21:40 <Abhishek09> diehlpk_work: When there is good time to talk with @rtohid

21:41 <diehlpk_work> Abhishek09, You might want to send him an email and ask him.

21:41 <diehlpk_work> I do not know his schedule

21:41 <Abhishek09> eMail?

21:42 <Abhishek09> diehlpk_work: Which project you mentoring this year?

21:42 <diehlpk_work> Depends on what submissions we have

21:44 <Abhishek09> But you are not interested in pip package anymore?

21:45 <Abhishek09> Ok please provide email id of rtohid

21:45 <diehlpk_work> I think since Rod is one of the main developers of phylanx, it might be better if he will mentor this project

21:46 <diehlpk_work> he has much more knowledge about the software and what to do

21:47 <diehlpk_work> Abhishek09, https://www.cct.lsu.edu/about/staff

21:47 <diehlpk_work> Please find his contact information there

21:49 <Abhishek09> He is not listed there

21:49 <Abhishek09> His full mame?

21:49 <Abhishek09> name

21:55 <Abhishek09> hkaiser: Do you have?

22:04 <hkaiser> Abhishek09: what do I have?

22:04 <Abhishek09> Email id of Rod

22:05 <hkaiser> see the staff page diehlpk_work posted above, his name is Tohid, Mohammad Rastegar

22:12 <Abhishek09> Thanks hkaiser

22:15 Abhishek09 has quit [Remote host closed the connection]

22:16 <diehlpk_work> parsa, Can you please add the AGAS results to the paper?

22:17 <diehlpk_work> You should have the last missing run by the end of this week

22:17 <diehlpk_work> Cori is down until Wed