Possible regression in the hydro test suite

Hi all,

I've been running some hydro tests (for my M&M scheme) that started looking weird after I rebased. It seems there's something odd going on with the hydro v.s. around 2 months ago, possibly to do with time-stepping.

To see what I mean, compare the following three plots:

Original

GCC8.1 MacOS

ICC COSMA-7

Note the significant difference in the velocity profile (the new ones are more wiggley). This difference is significantly amplified when looking at the Pressure-Energy scheme,

Original

GCC8.1 MacOS

ICC COSMA-7

Any ideas? Disabling optimization and vectorization does not change things.

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information