update all main files with the same runtime_error exception handling

[luhuhis]

resolves #24

[clarkedavida]

Everything passes with CUDA 11.6 in this branch except the following:

• DslashMulti
• HBOR_multiple
• hisqForce
• hisqForceImagMu
• rhmcTest
• rhmcTest_4pf The first three are fixed in a0e0a5ff7f38d3907e7db2abb96aa7dd93568fe0, d5758b447bcb7b5c21f173d7b55f45bbb3ec365c, and d278c36edfc9ec5c429ef4fd162577aeb0c0597a, respectively. The last three presumably suffer from the same issue that hisqForce did with the introduction of CUDA 11.6?

[dbollweg]

The issues with cuda 11.6 are unrelated to the changes introduced in this pr. I think we should be able to merge if everything passes on 11.0?

[clarkedavida]

I tend to agree with you, Dennis. Could either you or Lukas do the 11.0 tests and then merge it in if everything is fine?

[lukas-mazur]

I agree, I will run the tests

[lukas-mazur]

All tests (except the above mentioned tests) pass on single GPU. However, since try-catch adds some overhead, it would be good if timing tests could be done, before we merge this branch. Just to be sure.

[luhuhis]

I think someone already tested this some time ago (maybe @dbollweg ?), but I agree. Although I would be surprised if it there is was any difference.