address Debian experimental issues - Redmine #3205
The results of the beta1 show a couple of failures:
https://buildd.debian.org/status/package.php?p=gromacs&suite=experimental
Some of these have been resolved in beta2, other have (likely) not, in particular:
-
the arm64 build fails: 13: [ RUN ] HardwareTopologyTest.NumaCacheSelfconsistency 13: /<>/src/gromacs/hardware/tests/hardwaretopology.cpp:205: Failure 13: Expected: (c.size) > (0), actual: 0 vs 0 13: /<>/src/gromacs/hardware/tests/hardwaretopology.cpp:205: Failure 13: Expected: (c.size) > (0), actual: 0 vs 0 13: [ FAILED ] HardwareTopologyTest.NumaCacheSelfconsistency (9 ms) 13: -------- 4 tests from HardwareTopologyTest (35 ms total)
-
i386: ListedForcesTest, MathUnitTest fail
-
ppc64el:SimulatorsAreEquivalentDefaultModular/SimulatorComparisonTest.WithinTolerances/3
On most the other hardware there are failures in MdrunTests, FileIOTests, MdrunNonIntegratorTests
(from redmine: issue id 3205, created on 2019-11-15 by pszilard, closed on 2020-01-10)
- Changesets:
- Revision c5a01a73 by Mark Abraham on 2019-11-16T21:47:17Z:
Stop testing that hwloc cache-size detection works properly
This was never a unit test of whether GROMACS HardwareTopology can
detect cache line size, because it requires both that hwloc works
correctly on the execution hardware, and that GROMACS code uses its
results correctly. So we should stop running such tests.
Refs #3205
Change-Id: I4d13e39620335b235f9969e5e09299fb9b41b1a0
- Revision d76ad197 by Mark Abraham on 2019-11-16T22:42:26Z:
Relaxed listed forces test tolerances
Different compilers on different hardware generate different FP
instructions, which we need to tolerate.
Refs #3205
Change-Id: I965cecd82593dd7a0dd9e34d69937d06cd60735c
- Revision bc628c63 by Mark Abraham on 2019-11-26T14:47:50Z:
Relax tolerance for bd tests
The implementation is somehow not very reproducible.
Refs #3205
Change-Id: Ibf9de00d6e67b8e2237cc8726f300b2bf8ad2327
- Revision 245b4433 by Test User on 2019-12-02T13:52:27Z:
Relax test tolerance for cross correlation measure for densities
Some compilers do not perform the optimisations that are necessary to
reach default real tolerance when calculating the cross correlation of
larger three dimensional densities (100x100x100 voxels)
This patch loosens the requirement on the precision of cross correlation
evaluations.
refs #3205
Change-Id: Iad1bb1b17bb6935b2ce2537bcbb6c3902172a6bc
- Revision a83b3bdd by Test User on 2019-12-05T08:00:59Z:
Relax cross-correlation test precision
Relase the floating point precision in cross-correlation evaluations on
large data arrays further. i386 fails here with debian experimental build.
Refs #3205
Change-Id: I86c9ad1804da5be98671ac14a5bc37ea5636aad0
- Revision 93b5f239 by Mark Abraham on 2019-12-06T09:44:46Z:
Provide more feedback when listed forces tests fail
Refs #3205
Change-Id: Ic012aeaf8b35af88b78e109acabb6048d593cab3
- Revision c13f4653 by Mark Abraham on 2019-12-06T13:26:17Z:
Avoid 386 O3 codegen bug
Refs #3205
Change-Id: I2aa5018bde089b11497892dd68a7f9715e69e20e
- Revision af698c8f by Mark Abraham on 2019-12-31T13:56:42Z:
Avoid 386 O3 codegen bug even more
Refs #3205
Change-Id: I78f3cc03da36e129a842507f9cdd3d170beb8c08