Timothy Smith
Timothy Smith
@trilinos/tpetra @bartlettroscoe ## Motivation Informational output is of limited utility when tests are passing, and can randomly break automated PR testing when it overlaps with and breaks up the "test...
@trilinos/tpetra ## Motivation Test refactoring and consolidation. Replaces #10728 due to CDash issues. ## Testing Test refactoring -- tests should continue to pass
## Enhancement @trilinos/tpetra Currently Tpetra does not have low-level, targeted performance tests. If the performance dashboard review process becomes less labor-intensive, adding these tests could be very helpful for isolating...
## Enhancement @trilinos/tpetra The "4-arg" (varying number of packets per LID) variant of Distributor's doPosts operation does not have a unit test covering the "slow path" case (when the data...
Some tests are failing in CPU builds with segfaults that appear to be from OpenFAST. Presumably these are all the same bug, although we can split up this issue if...
The nightly dashboards report several clang warnings about overridden functions in the master element code. @overfelt these warnings began to appear after #1022 merged, although I'm a little confused about...
Issues that need to be resolved to make the dashboard green: * #976 * #977 * #979 * #973 * #982 * #983 * #984 * #985 * #994 *...
The SNL dashboard has configuration errors on all Cuda lines, likely due to a Trilinos change. Associated Trilinos issue: https://github.com/trilinos/Trilinos/issues/10954
On all nightly GPU test lines that don't include OpenFast, AMSKernelHex8Mesh.NGP_ams_forcing is failing with "terminate called without an active exception". (The OpenFast lines fail differently, due to #979)
Is there an actual issue, or do we just need to raise the tolerances? Is there a problem with using the same set of gold files for 6 distinct dashboard...