Using v2.5.1 at NERSC. State file attached: quickview-state_aerosols_linear_log_symlog.json
Below are 6 screenshots showing 2 variables in 2 rows. Both variables are positive-definite, and each variable is presented 3 times using linear, log, and symlog scales (left to right). I had expected the log and symlog scales to give very similar images. This was indeed the case in the 1st row of screenshots but not the case in the 2nd row.
Note that the 1st variable has very large values: ~1e8 to ~1e10; the 2nd variable has very small values: ~1e-13 to ~2e-10. I'm guessing the linear threshold used for symlog is not accommodating the small values of the 2nd variable?
Using v2.5.1 at NERSC. State file attached: quickview-state_aerosols_linear_log_symlog.json
Below are 6 screenshots showing 2 variables in 2 rows. Both variables are positive-definite, and each variable is presented 3 times using linear, log, and symlog scales (left to right). I had expected the log and symlog scales to give very similar images. This was indeed the case in the 1st row of screenshots but not the case in the 2nd row.
Note that the 1st variable has very large values: ~1e8 to ~1e10; the 2nd variable has very small values: ~1e-13 to ~2e-10. I'm guessing the linear threshold used for symlog is not accommodating the small values of the 2nd variable?