Ctests on Perlmutter-GPU with nvidiagpu running much more slowly than on Perlmutter-CPU with nvidia

[altheaden]

@xylar and I noticed that ctests on pm-cpu with nvidia ran at about 90 seconds, but on pm-gpu with nvidiagpu they ran at 1538.76 seconds.

[xylar]

This seems qualitatively similar to the slow Intel performance we've seen: #119 But this case seems even worse. 25 minutes instead of 1.5 is really quite bad!

[xylar]

It could be that Polaris or Omega itself isn't setting some environment variables that are needed or something like that.

[xylar]

I ran the tests myself and they didn't take that long. They're still quite slow but only took 577.79 seconds (about 10 minutes)