tp2.1 tests occasional differences in netcdf file

[JessicaMeixner-NOAA]

Describe the bug
Occasionally the netcdf gridded files in tp2.1 tests will have different number of outputs, despite no other differences in the regression tests.

To Reproduce
Get lucky to happen to see this? Is hard to replicate issue

Expected behavior
Same code should produce same output.

Screenshots
For example difference will look like in output/ww3_tp2.1/work_PR2_UNO/ww3.196806.nc_diff.txt

6c6
<       time = UNLIMITED ; // (6 currently)
---
>       time = UNLIMITED ; // (17 currently)
80c80
<               :stop_date = "1968-06-06 05:00:00" ;
---
>               :stop_date = "1968-06-06 04:48:00" ;
95,96c95,97
<  time = -7879, -7878.95833333333, -7878.91666666667, -7878.875, 
<     -7878.83333333333, -7878.79166666667 ;
---
>  time = -7879, -7878.9875, -7878.975, -7878.9625, -7878.95, -7878.9375, 
>     -7878.925, -7878.9125, -7878.9, -7878.8875, -7878.875, -7878.8625, 
>     -7878.85, -7878.8375, -7878.825, -7878.8125, -7878.8 ;

Additional context
There are no errors or any thing else pointing to issues that I've been able to find. This is basically running the same branch twice. I belive @sbanihash might have seen this before too.