Fix single-point matrixCN fails

[slevis-lmwg]

Description of changes

Change SparseMatrixMultiplyMod.F90 line 1246:
SHR_ASSERT_FL((size(filter_actunit_C) > num_actunit_C), sourcefile, __LINE__)
to
SHR_ASSERT_FL((size(filter_actunit_C) >= num_actunit_C), sourcefile, __LINE__)
to help some failing matrixCN tests pass.

Specific notes

Contributors other than yourself, if any:
@ekluzek

CTSM Issues Fixed (include github issue #):
Fix #2780

Are answers expected to change (and if so in what way)?
No.

Any User Interface Changes (namelist or namelist defaults changes)?
No.

Does this create a need to change or add documentation? Did you do so?
No.

Testing performed, if any:
Failing tests now pass (see issue #2780).

[slevis-lmwg]

Rebase to cesm3_0_beta04_changes

[ekluzek]

Thanks for this. Just needs to be done elsewhere as well≥

[ekluzek]

We realized this should be done elsewhere, so I'll look at it again when this is done to approve again.

[slevis-lmwg]

Testing on derecho and izumi in prep for merge:
OK ./run_sys_tests -s matrixcn -c ctsm5.3.009 --skip-generate

[slevis-lmwg]

@ekluzek this is ready for me to merge with your approval.

[ekluzek]

Awesome this is great.

Erik, you approved this a few minutes ago, but this review is still blocking the merge. I'm dismissing your review in case it unblocks the merge.

[slevis-lmwg]

@ekluzek for now this is still blocked from merge...

[slevis-lmwg]

For the record:
I'm temporarily changing the beta04 branch settings to bypass the need for another review and approval.

[slevis-lmwg]

I have changed back the settings to require a review and approval.