CTests fail for current HEAD on chrysalis #565

Activity

1. philipwjones

   added

   bugSomething isn't workingSomething isn't working

   on Sep 18, 2026
2. philipwjones

   added the

   Bug

   issue type on Sep 18, 2026
3. xylar commented on Sep 21, 2026

   

   xylar

   on Sep 21, 2026

   More actions

   @philipwjones, I can't reproduce this. Develop at 5dfd2ae passes all 49 CTests on Chrysalis with Intel (and gnu) in Release, built and run with the Polaris Omega CTest utility, which links OmegaSphereMesh.nc to cosine_bell_icos480.omega_vars.260807.nc from https://web.lcrc.anl.gov/public/e3sm/polaris/ocean/omega_ctest/. Log:

   /lcrc/group/e3sm/ac.xylar/polaris_1.1/chrysalis/test_20260921/omega_ctest_565/develop_5dfd2ae/omega_ctest_chrysalis_intel.o1292001
   

   The four failing tests are exactly the ones that use OmegaSphereMesh.nc, so I suspect an outdated sphere mesh file. Which file is yours linked to? The 260807 file is the one the QuickStart points to; the expected norms in the sphere tests are tuned to it. Note that #547 will change the expected mesh again (to 260911).

   ---

   Posted by Claude Code on @xylar's behalf. The testing, analysis and wording above are AI-authored; please check them accordingly.
4. philipwjones commented on Sep 21, 2026

   

   philipwjones

   on Sep 21, 2026

   Author

   More actions

   Ah, that was it, sorry. When I updated the other files recently, I grabbed the 260911 version assuming it was the latest....
5. philipwjones

   closed this as completedon Sep 21, 2026