Does it make sense that Hohlraum flux query regression test yields different results on Windows?

[biagas]

For hf_3_0.txt as an example, the regression results from pascal are:

The hohlraum flux query over 1 energy groups was written to hf0.ult.
Sum is 0.538024.
Energy group values are:  0.538024

There isn't a mode-specific baseline, so serial and parallel runs yield same results. Not so on Windows

From serial run on Windows:

The hohlraum flux query over 1 energy groups was written to hf0.ult.
Sum is 0.495349.
Energy group values are:  0.495349

From parallel run on Windows:

The hohlraum flux query over 1 energy groups was written to hf0.ult.
Sum is 0.497103.
Energy group values are:  0.497103

[brugger1]

I believe the hohlraum flux query uses a random number generator to define the rays for the query.