Continuous Integration via Travis

[L30nardoSV]

This PR aims to add a preliminary support of the Continuous Integration (CI) feature via Travis into AutoDock-GPU.

1. Every commit triggers a set of automatic test on different systems featuring multi-core CPUs. This is so far limited to Ubuntu, but can be later extended to other Linux distributions, or even to macOS (depending on available support).

2. Current configuration runs on the following Ubuntu distributions: Focal (20.04), Bionic (18.04), Xenial (16.04).

• On Focal, four test cases are performed, namely NUMWI = { 1, 4, 8, 16 }.
• On Bionic and Xenial, this is currently limited to only NUMWI = 1 for each distribution.

3. Proposed tests are available within this run_test.sh script. Their length is limited to lower bounds (in terms of #molecules, #LGA runs, etc) in order finish testing in a reasonable amount of time (approx. 10 min for each case).

4. All builds are listed here, while details of lastest build can be found here. Notice that running clinfo shows that only two CPU cores are available per system.

5. Minor code clean-up on Makefiles.OpenCL and Makefile.Cuda.

Looking forward to your feedback.

[L30nardoSV]

I updated the PR so that CI is based on GitHub actions instead of Travis.

@atillack @diogomart @jeeberhardt, please have a look and provide feedback :)

[atillack]

@L30nardoSV That's great - thank you. I think the only thing missing now is to automatically compare the test run results against reference values (+/- 0.3 kcal/mol maybe) and use it :-)