Skip to content

{math}[GCC/7.3.0-2.30,iccifort-2018.3.222-GCC-7.3.0-2.30] Arb v2.16.0, FLINT v2.5.2#8137

Merged
vanzod merged 5 commits intoeasybuilders:developfrom
SethosII:20190425133110_new_pr_Arb2160
Jul 5, 2019
Merged

{math}[GCC/7.3.0-2.30,iccifort-2018.3.222-GCC-7.3.0-2.30] Arb v2.16.0, FLINT v2.5.2#8137
vanzod merged 5 commits intoeasybuilders:developfrom
SethosII:20190425133110_new_pr_Arb2160

Conversation

@SethosII
Copy link
Copy Markdown
Contributor

(created using eb --new-pr)

@SethosII
Copy link
Copy Markdown
Contributor Author

Test report by @SethosII
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in this PR)
ribnode015 - Linux ubuntu 18.04, Intel(R) Xeon(R) CPU E5645 @ 2.40GHz, Python 2.7.15rc1
See https://gist.github.com/58cd9f61e0a0295bcfd2c3331c9467ab for a full test report.

@vanzod
Copy link
Copy Markdown
Member

vanzod commented May 2, 2019

@SethosII Given that both libraries include math functions and the documentation puts a lot of emphasis on their performance, I would argue that it may be worth building separately with both GCC and icciifort compilers. Thoughts?

@vanzod vanzod added the new label May 2, 2019
@vanzod vanzod added this to the 3.x milestone May 2, 2019
@SethosII
Copy link
Copy Markdown
Contributor Author

SethosII commented May 3, 2019

@vanzod So I should change the toolchain to foss?

@vanzod
Copy link
Copy Markdown
Member

vanzod commented May 3, 2019

@SethosII No, since there should not be any requirement for either MPI, OpenBLAS or FFTW.
GCC-7.3.0-2.30 makes more sense in the context of using the lowest possible toolchain.

@akesandgren
Copy link
Copy Markdown
Contributor

@SethosII ping, I agree with @vanzod please use GCC-7.3.0-2.30 as toolchain instead.
And it wouldn't hurt to add easyconfigs for icc/2018.3.222-GCC-7.3.0-2.30 too.

@SethosII
Copy link
Copy Markdown
Contributor Author

@akesandgren @vanzod I changed the toolchain. I don't have the intel compiler so I can't test this.

@akesandgren
Copy link
Copy Markdown
Contributor

You don't have to test the versions for the icc/2018.3.222-GCC-7.3.0-2.30 toolchain. We can do that.

But please do add Arb and FLINT easyconfigs with icc/2018.3.222-GCC-7.3.0-2.30 toolchain to this PR.
We'll run the test-reports for it.

@SethosII
Copy link
Copy Markdown
Contributor Author

@akesandgren Maybe a dumb question but is icc a real toolchain? I couldn't find any existing easyconfig with this one, is iccifort the real toolchain?

@akesandgren
Copy link
Copy Markdown
Contributor

Hmm, you're probably right, @boegel you know the correct answer here...

@akesandgren
Copy link
Copy Markdown
Contributor

Quite so. After actually looking at the situation it's iccifort-2018.3.222-GCC-7.3.0-2.30.eb you should use. That corresponds to intel/2018b

@akesandgren akesandgren changed the title {math}[GCCcore/7.3.0] Arb v2.16.0, FLINT v2.5.2 {math}[GCC/7.3.0-2.30] Arb v2.16.0, FLINT v2.5.2 May 20, 2019
@akesandgren
Copy link
Copy Markdown
Contributor

Test report by @akesandgren
FAILED
Build succeeded for 3 out of 4 (4 easyconfigs in this PR)
b-an03.hpc2n.umu.se - Linux ubuntu 16.04, Intel(R) Xeon(R) CPU E5-2690 v4 @ 2.60GHz, Python 2.7.12
See https://gist.github.com/94378ab98ba6774a9e045c93e72fe3a0 for a full test report.

portable, and extensively tested."""

toolchain = {'name': 'iccifort', 'version': '2018.3.222-GCC-7.3.0-2.30'}

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can't find the real reason for the failing test (not enough time), but add a
toolchainopts = {'noopt': True}
to Arb for the iccifort version and it will pass the tests.

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@SethosII Do you have any chance to test this build on a more recent CPU architecture and see if there are any other solutions aside from what Åke suggested?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@vanzod I don't have the intel compiler, so I can't test the failing easyconfig myself.

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@SethosII Then please go ahead and apply the change that @akesandgren proposed.

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@vanzod @akesandgren I added the toolchain option.

Comment thread easybuild/easyconfigs/f/FLINT/FLINT-2.5.2-GCC-7.3.0-2.30.eb Outdated
@vanzod
Copy link
Copy Markdown
Member

vanzod commented Jun 17, 2019

Test report by @vanzod
FAILED
Build succeeded for 3 out of 4 (4 easyconfigs in this PR)
cermis - Linux debian 9.9, Intel(R) Xeon(R) CPU X5670 @ 2.93GHz, Python 2.7.13
See https://gist.github.com/e51698b7859d2b6f4b1d1e9e5d5cb0d5 for a full test report.

@verdurin
Copy link
Copy Markdown
Member

Test report by @verdurin
SUCCESS
Build succeeded for 4 out of 4 (4 easyconfigs in this PR)
rescomp2 - Linux centos linux 7.3.1611, Intel(R) Xeon(R) Gold 6126 CPU @ 2.60GHz, Python 2.7.5
See https://gist.github.com/067ed09b25457ca1db8df6afb460b83a for a full test report.

@vanzod
Copy link
Copy Markdown
Member

vanzod commented Jul 5, 2019

Test report by @vanzod
SUCCESS
Build succeeded for 4 out of 4 (4 easyconfigs in this PR)
cermis - Linux debian 9.9, Intel(R) Xeon(R) CPU X5670 @ 2.93GHz, Python 2.7.13
See https://gist.github.com/7e5b06808d5c8eca9e98f2a58573a0c0 for a full test report.

@vanzod vanzod dismissed akesandgren’s stale review July 5, 2019 17:20

Requested changes have been applied

@vanzod vanzod modified the milestones: 3.x, next release (3.9.3) Jul 5, 2019
@vanzod
Copy link
Copy Markdown
Member

vanzod commented Jul 5, 2019

Going in, thanks @SethosII!

@vanzod vanzod merged commit 3deb394 into easybuilders:develop Jul 5, 2019
portable, and extensively tested."""

toolchain = {'name': 'iccifort', 'version': '2018.3.222-GCC-7.3.0-2.30'}
toolchainopts = {'noopt': True}
Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@akesandgren This was added by @SethosII based on your suggestion.

But, this corresponds to using -O0, which makes the resulting build painfully slow?

@boegel boegel changed the title {math}[GCC/7.3.0-2.30] Arb v2.16.0, FLINT v2.5.2 {math}[GCC/7.3.0-2.30,iccifort-2018.3.222-GCC-7.3.0-2.30] Arb v2.16.0, FLINT v2.5.2 Jul 6, 2019
@SethosII SethosII deleted the 20190425133110_new_pr_Arb2160 branch July 8, 2019 09:57
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants