-
Notifications
You must be signed in to change notification settings - Fork 358
Race condition with a threaded test with matrix on? #3360
Copy link
Copy link
Open
Labels
bugsomething is working incorrectlysomething is working incorrectlyclosed: wontfixWe won't fix this issue, because it would be too difficult and/or isn't important enough to fixWe won't fix this issue, because it would be too difficult and/or isn't important enough to fixpriority: lowBackground task that doesn't need to be done right away.Background task that doesn't need to be done right away.
Milestone
Metadata
Metadata
Assignees
Labels
bugsomething is working incorrectlysomething is working incorrectlyclosed: wontfixWe won't fix this issue, because it would be too difficult and/or isn't important enough to fixWe won't fix this issue, because it would be too difficult and/or isn't important enough to fixpriority: lowBackground task that doesn't need to be done right away.Background task that doesn't need to be done right away.
Type
Fields
Give feedbackNo fields configured for Bug.
Brief summary of bug
In ctsm5.3.065 I had the test
ERP_P64x2_D_Ld5.f10_f10_mg37.I1850Clm50Bgc.derecho_intel.clm-ciso--clm-matrixcnOn_ignore_warnings
fail at runtime. But, after resubmitting it worked fine.
This is just to bring awareness that this is a potential problem and a simple resubmit solves it.
General bug information
CTSM version you are using: ctsm5.3.064-72-g4037e5a8d
Does this bug cause significantly incorrect results in the model's science? no
Configurations affected: threading with soil matrix on
Details of bug
It dies in a SHR_ASSERT statement that's ensuring the matrix size of two sparse matrices multiplied together are the same size.
Important output or errors that show the problem
cesm.log: