[petsc-users] false negatives ontests

[Matthew Knepley]

On Fri, Apr 3, 2020 at 12:34 PM Mark Adams <mfadams at lbl.gov> wrote:

> I run this test and it passes. But I know it's wrong so I* change a 6 to
> a 7 in output/ex11_0.out *and I get this diff AND all the rest.
> Any idea what is going on?
>

Yes. We do not compare any real number, only integers. Once you change an
integer, you also get the diff
of all the real numbers. This is, of course, not perfect, but we did not
have the manpower to overhaul all tests
to account for wibble among machines.

   Matt


> Thanks,
>
> 12:25 mark/feature-xgc-interface-rebase *> ~/Codes/petsc-master$ make -f
> ./gmakefile test globsearch="dm_impls_plex_tutorials-ex11_0" PETSC_DIR=$PWD
> Using MAKEFLAGS: PETSC_DIR=/Users/markadams/Codes/petsc-master
> globsearch=dm_impls_plex_tutorials-ex11_0
>         TEST
> arch-macosx-gnu-g/tests/counts/dm_impls_plex_tutorials-ex11_0.counts
>  ok dm_impls_plex_tutorials-ex11_0
> not ok diff-dm_impls_plex_tutorials-ex11_0 # Error code: 1
> #       55c55
> #       <               ***** FormRHSSource: have new_imp_rate=  4.975e+01
> dt=0.125 stepi=1 time=  1.000e-01
> #       ---
> #       >               ***** FormRHSSource: have new_imp_rate=  4.975e-04
> dt=0.125 stepi=1 time=  1.000e-01
> #       64,76c64,76
> #       <     0 SNES Function norm 3.285336127013e-04
> #       <     1 SNES Function norm 9.604648941118e-07
> #       <     2 SNES Function norm 1.964241915657e-08
> #       <     3 SNES Function norm 8.824980789615e-10
> #       <     4 SNES Function norm 4.556683618010e-11
> #       <     5 SNES Function norm 2.470586731358e-12
> #       <     6 SNES Function norm 1.385600544214e-13
>
> *#       <   Nonlinear solve converged due to CONVERGED_SNORM_RELATIVE
> iterations 7*#       <               ***** FormRHSSource: have
> new_imp_rate=  1.981e+03 dt=0.125 stepi=1 time=  1.625e-01
> #       < [0] TSAdaptChoose_Basic(): Estimated scaled local truncation
> error 0.466012, accepting step of size 0.125
> #       <   2) species-0: charge density= -1.6018377279742e+01 z-momentum=
> -2.8181973830519e-19 energy=  1.2008542897309e+05
> #       <   2) species-1: charge density=  1.6024885661178e+01 z-momentum=
> -9.0961762196061e-18 energy=  1.2013832642627e+05
> #       <         2) Total: charge density=  6.5083814358857e-03,
> momentum= -9.3779959579113e-18, energy=  2.4022375539935e+05 (m_i[0]/m_e =
> 3670.94, 80 cells)
> #       ---
> #       >     0 SNES Function norm 3.285336109721e-04
> #       >     1 SNES Function norm 9.602918060455e-07
> #       >     2 SNES Function norm 1.961138067634e-08
> #       >     3 SNES Function norm 8.793960138447e-10
> #       >     4 SNES Function norm 4.534301360955e-11
> #       >     5 SNES Function norm 2.455937690832e-12
> #       >     6 SNES Function norm 1.404599098408e-13
>
> *#       >   Nonlinear solve converged due to CONVERGED_SNORM_RELATIVE
> iterations 6*#       >               ***** FormRHSSource: have
> new_imp_rate=  5.192e-04 dt=0.125 stepi=1 time=  1.625e-01
> #       > [0] TSAdaptChoose_Basic(): Estimated scaled local truncation
> error 0.674842, accepting step of size 0.125
> #       >   2) species-0: charge density= -1.6018371940813e+01 z-momentum=
> -6.0282835882130e-19 energy=  1.2008542888021e+05
> #       >   2) species-1: charge density=  1.6024885680151e+01 z-momentum=
> -9.8338166614696e-18 energy=  1.2013832641421e+05
> #       >         2) Total: charge density=  6.5137393379366e-03,
> momentum= -1.0436645020291e-17, energy=  2.4022375529442e+05 (m_i[0]/m_e =
> 3670.94, 80 cells)
> #       78c78
> #       < 2 TS dt 0.15625 time 0.225
> #       ---
> #       > 2 TS dt 0.136947 time 0.225
>
> # -------------
> #   Summary
> # -------------
> # FAILED diff-dm_impls_plex_tutorials-ex11_0
> # success 1/2 tests (50.0%)
> # failed 1/2 tests (50.0%)
> # todo 0/2 tests (0.0%)
> # skip 0/2 tests (0.0%)
> #
> # Wall clock time for tests: 48 sec
> # Approximate CPU time (not incl. build time): 47.74 sec
> #
> # To rerun failed tests:
> #     /usr/bin/make -f gmakefile test test-fail=1
> #
> # Timing summary (actual test time / total CPU time):
>


-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener

https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20200403/fbaa9d0c/attachment-0001.html>