Skip to content

stop using lowopt in libxc 4.2.3 easyconfigs#7115

Merged
akesandgren merged 2 commits intoeasybuilders:developfrom
boegel:libxc_lowopt
Nov 6, 2018
Merged

stop using lowopt in libxc 4.2.3 easyconfigs#7115
akesandgren merged 2 commits intoeasybuilders:developfrom
boegel:libxc_lowopt

Conversation

@boegel
Copy link
Copy Markdown
Member

@boegel boegel commented Nov 4, 2018

It seems like the enforcing lowopt for recent libxc versions, which was added in #4199, is no longer required with libxc 4.2.3.

We're not We are running make check, which should be sufficient to catch any problems with compiling with -O2 (defaultopt) rather than -O1 (lowopt).

cc @akesandgren, @tovrstra (cfr. #4199), @migueldiascosta (cfr. #5250)

@boegel boegel added the change label Nov 4, 2018
@boegel boegel added this to the 3.8.0 milestone Nov 4, 2018
@boegel boegel requested a review from akesandgren November 4, 2018 15:00
@akesandgren
Copy link
Copy Markdown
Contributor

@boegel "We're not running make check"?? We are running make check.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 4, 2018

@akesandgren Oops, Sunday afternoon brain fart... Fixed in PR description.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 4, 2018

Test report by @boegel
SUCCESS
Build succeeded for 9 out of 9 (5 easyconfigs in this PR)
node3130.skitty.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz, Python 2.7.5
See https://gist.github.com/5d094f9187eb380e3d848b664b3a5847 for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 4, 2018

Test report by @boegel
SUCCESS
Build succeeded for 5 out of 5 (5 easyconfigs in this PR)
node2400.golett.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz, Python 2.7.5
See https://gist.github.com/362722d607a618b271868e9c8da50e1d for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 4, 2018

Test report by @boegel
FAILED
Build succeeded for 4 out of 5 (5 easyconfigs in this PR)
node2073.delcatty.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz, Python 2.7.5
See https://gist.github.com/317ab8e93571e102542aa19fd2a81eae for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 4, 2018

More info for failing make check for libxc-4.2.3-gimkl-2017a.eb on Intel Sandy Bridge (from testsuite/test-suite.log):

* mgga_x_b00                     Li           1        vxc
2 0 4.4333808900e+256 -1.9597684224e+00 1.000000e+00

2 1 6.2976726689e+251 0.0000000000e+00 1.000000e+00

2 2 1.6361078443e+253 0.0000000000e+00 1.000000e+00

2 3 -6.5444313773e+253 -4.5664700550e-02 1.000000e+00

3 0 5.5501858511e+254 -2.4022477938e+00 1.000000e+00

3 1 2.6979913163e+252 0.0000000000e+00 1.000000e+00

3 2 2.3456742475e+253 0.0000000000e+00 1.000000e+00

3 3 -9.3826969900e+253 6.1465768638e-02 1.000000e+00

4 0 -2.7423387624e+255 -4.3927352504e-01 1.000000e+00

4 1 5.7748027454e+254 0.0000000000e+00 1.000000e+00

4 2 6.9474540892e+253 0.0000000000e+00 1.000000e+00

4 3 -2.7789816357e+254 -2.3030456703e-02 1.000000e+00

5 0 2.5412629950e+254 -1.5815394361e-01 1.000000e+00

5 1 2.3795219844e+256 0.0000000000e+00 1.000000e+00

5 2 2.2843416276e+254 0.0000000000e+00 1.000000e+00

5 3 -9.1373665105e+254 -6.7787655709e-01 1.000000e+00

6 0 -2.7191743577e+254 -5.9905856975e-02 1.000000e+00

6 1 6.4427935338e+257 0.0000000000e+00 1.000000e+00

6 2 3.8519473844e+254 0.0000000000e+00 1.000000e+00

6 3 -1.5407789538e+255 -1.8051804720e-01 1.000000e+00

7 0 -9.6783057427e+256 -2.2418533527e-01 1.000000e+00

7 1 2.7434667457e+260 0.0000000000e+00 1.000000e+00

7 2 2.0151213308e+255 0.0000000000e+00 1.000000e+00

7 3 -8.0604853233e+255 -5.2085604397e-12 1.000000e+00

8 0 -1.4421572934e+258 -5.9645065810e-02 1.000000e+00

8 1 2.7474641277e+267 0.0000000000e+00 1.000000e+00

8 2 1.1335985538e+257 0.0000000000e+00 1.000000e+00

8 3 -4.5343942154e+257 -2.7070395604e-29 1.000000e+00
     FAIL

Not sure why that problem only occurs on Intel Sandy Bridge, and not on Haswell or Skylake...

@akesandgren
Copy link
Copy Markdown
Contributor

That's odd. Does it always happen?

@migueldiascosta
Copy link
Copy Markdown
Member

Test report by @migueldiascosta
SUCCESS
Build succeeded for 5 out of 5 (5 easyconfigs in this PR)
grc-cluster1 - Linux centos 6.10, Intel(R) Xeon(R) CPU E5-2640 0 @ 2.50GHz, Python 2.7.14
See https://gist.github.com/d278607c561ab0464dffdc3642b4c9f1 for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 5, 2018

Test report by @boegel
FAILED
Build succeeded for 4 out of 5 (5 easyconfigs in this PR)
node2068.delcatty.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz, Python 2.7.5
See https://gist.github.com/5c0c1e34ba94783ad62fcc37485482d0 for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 5, 2018

@akesandgren Looks consistent... :(

I can roll back the change for libxc-4.2.3-gimkl-2017a.eb, and only drop lowopt for recent toolchai n (2017b or newer)?

@akesandgren
Copy link
Copy Markdown
Contributor

So it's only for gimkl/2017a it happens? Then I'd say revert that one and leave the others.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 5, 2018

Testing again on a full node, previous tests were on a partial node (4 cores, 16GB of RAM), maybe that makes the tests fail?

As @migueldiascosta pointed out, his test was also on Intel Sandy Bridge, and it worked fine there...

If the tests fail again for libxc-4.2.3-gimkl-2017a.eb using a full workernode, I'll indeed roll back the change only for libxc-4.2.3-gimkl-2017a.eb so we can move forward.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 5, 2018

Test report by @boegel
FAILED
Build succeeded for 4 out of 5 (5 easyconfigs in this PR)
node2128.delcatty.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz, Python 2.7.5
See https://gist.github.com/d9d9ab4599d8e79778a42316a06dca37 for a full test report.

@boegel
Copy link
Copy Markdown
Member Author

boegel commented Nov 5, 2018

Test report by @boegel
SUCCESS
Build succeeded for 4 out of 4 (4 easyconfigs in this PR)
node2042.delcatty.os - Linux centos linux 7.5.1804, Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz, Python 2.7.5
See https://gist.github.com/03d83250373c84ad97742ab1184a3211 for a full test report.

Copy link
Copy Markdown
Contributor

@akesandgren akesandgren left a comment

Choose a reason for hiding this comment

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

LGTM

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.

3 participants