DALTON 2018 on x86_64

Problems with Dalton installation? Find answers or ask for help here
Post Reply
RikaK
Posts: 11
Joined: 09 Jun 2014, 03:40
First name(s): Rika
Last name(s): Kobayashi
Affiliation: Australian National University
Country: Australia

DALTON 2018 on x86_64

Post by RikaK » 16 Aug 2019, 06:16

Back to my favourite topic.
I just noticed that Dalton 2018 is out so have checked out Dalton2018.2 and taken it through a build.
With intel-fc/2018.3.222 intel-cc/2018.3.222 intel-mkl/2018.3.222 the test suite is giving me
95% tests passed, 24 tests failed out of 496
(I have still to chase down the failures).
But as our system guys are now going zero-tolerance on test-suite failures is there a recommended compiler version for the test-suite to pass?
Dalton used to provide such a recommendation in the past which I found helpful but I couldn't immediately see one (apart from "Don't use Intel 2015") for Dalton 2018.

The following tests FAILED:
340 - choles_ccsdpt_energy (Failed)
422 - cc2_r12_aux_ne (Failed)
435 - cc2_r12_aux_ch4_cbs1 (Failed)
437 - cc2_r12_aux_ch4_cbs3 (Failed)
440 - ccsdr12_bh_direct (Failed)
441 - ccsdr12_n2_direct (Failed)
442 - ccsdr12_bh (Failed)
443 - ccsdr12_n2 (Failed)
446 - ccsdr12_n2_sym_exc (Failed)
447 - ccsdr12_n2_sym_exc_lhtr (Failed)
449 - ccsdr12_n2_gamma (Failed)
450 - ccsdr12_n2_gamma2 (Failed)
451 - ccsdr12_h2o_gamma (Failed)
452 - ccsdr12_h2o_gamma2 (Failed)
453 - ccsdr12_be (Failed)
454 - cc3r12_n2_sym_exc (Failed)
455 - cc3r12_n2_sym_exc_lhtr (Failed)
456 - cc3r12_n2_sym_exc_tight (Failed)
457 - cc3r12_n2_sym_exc_lhtr_tight (Failed)
487 - benchmark_eri_adz (Timeout)
489 - benchmark_eri_atzs (Failed)
491 - benchmark_eri_r12xl (Timeout)
492 - benchmark_her_adz (Timeout)
494 - benchmark_her_atzs (Timeout)
Errors while running CTest

User avatar
magnus
Posts: 471
Joined: 27 Jun 2013, 16:32
First name(s): Jógvan Magnus
Middle name(s): Haugaard
Last name(s): Olsen
Affiliation: Hylleraas Centre, UiT The Arctic University of Norway
Country: Norway

Re: DALTON 2018 on x86_64

Post by magnus » 16 Aug 2019, 09:43

I can reproduce these errors and had to go all the way back to compiler version 15 (specifically tested 15.0.3) which corresponds to 2015 update 3.

User avatar
magnus
Posts: 471
Joined: 27 Jun 2013, 16:32
First name(s): Jógvan Magnus
Middle name(s): Haugaard
Last name(s): Olsen
Affiliation: Hylleraas Centre, UiT The Arctic University of Norway
Country: Norway

Re: DALTON 2018 on x86_64

Post by magnus » 16 Aug 2019, 09:50

I've created an issue for this: https://gitlab.com/dalton/dalton/issues/101

RikaK
Posts: 11
Joined: 09 Jun 2014, 03:40
First name(s): Rika
Last name(s): Kobayashi
Affiliation: Australian National University
Country: Australia

Re: DALTON 2018 on x86_64

Post by RikaK » 17 Aug 2019, 01:37

Hi Magnus,
Thanks for the confirmation and follow-up. Unfortunately/fortunately, we're upgrading our supercomputer but we are not carrying over any of the older compilers or libraries so 15.0.3 (we have intel-fc/15.0.3.187) won't be an option. Can't help being amused by the irony of the "Many tests fail using Intel 2015. Therefore we currently do not recommend to compile Dalton 2018.1 using Intel 2015." advice though.
Cheers,
Rika
PS And I was going to say that I have just tried intel2019 but I see from gitlab issues that you have too.

User avatar
magnus
Posts: 471
Joined: 27 Jun 2013, 16:32
First name(s): Jógvan Magnus
Middle name(s): Haugaard
Last name(s): Olsen
Affiliation: Hylleraas Centre, UiT The Arctic University of Norway
Country: Norway

Re: DALTON 2018 on x86_64

Post by magnus » 17 Aug 2019, 18:26

Ah yes, that is an oversight. We had some issues with Intel 2015 but that was a long time ago. The documentation is not up to date.

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: DALTON 2018 on x86_64

Post by taylor » 18 Aug 2019, 08:34

Update3 for the 2015 compiler is fine, and most of the time so is 2016.U0 (to my surprise...). No comment on more recent Intel compilers because we have issues with licences that involve a four-way dialog between Intel, Dell, people who are no longer here, and ourselves. Oh, and university purchasing.

Meanwhile, I posted a question about the latest Portland Group compilers on the developer's forum three months ago: no replies, which suggests no-one is using the pgroup compilers??

Best regards
Pete

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest