Improve test suite speed: pytest-xdist; qcut and qchem markers; update reconstruct tests

[antalszava]

Context

We aim to lower the runtime of our test suites.

To achieve that, there are further approaches that can be followed up:

• Slow tests can be revisited: understanding why the test takes a long time is key;
• Distributing the entirety of the test suite into multiple GitHub runners helps;
• As pytest has capabilities for running tests, those can also be utilized.

Changes

• Creates two new markers in the test suite:
  • qcut
  • qchem
• Separates qcut and qchem into their own runners;
• Uses pytest-xdist to distribute the test runs across 2 cores (-n auto; GitHub seems to provision 2-core machines);
• One test case in tests/fourier/test_reconstruct.py frequently runs for almost 1 minute in itself; revisited this file and the test case in specific.

Posibble Drawbacks

• When using pytest-xdist, the separate files are not shown in the log, just dots for the entire suite:
  

Note, that upon a failing test, the file name will be showing nonetheless.

[github-actions]

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md with:

• A one-to-two sentence description of the change. You may include a small working example for new features.
• A link back to this PR.
• Your name (or GitHub username) in the contributors section.

[codecov]

Codecov Report

Merging #2504 (441e2fe) into master (207074e) will increase coverage by 2.24%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #2504      +/-   ##
==========================================
+ Coverage   97.23%   99.47%   +2.24%     
==========================================
  Files         244      244              
  Lines       19433    19433              
==========================================
+ Hits        18895    19331     +436     
+ Misses        538      102     -436     

Impacted Files	Coverage Δ
pennylane/_qubit_device.py	98.73% <0.00%> (+0.31%)	⬆️
pennylane/interfaces/tensorflow_autograph.py	100.00% <0.00%> (+1.17%)	⬆️
...ane/transforms/optimization/single_qubit_fusion.py	100.00% <0.00%> (+1.92%)	⬆️
pennylane/interfaces/tensorflow.py	100.00% <0.00%> (+2.04%)	⬆️
pennylane/devices/default_qubit_tf.py	91.93% <0.00%> (+3.22%)	⬆️
pennylane/interfaces/torch.py	100.00% <0.00%> (+3.27%)	⬆️
pennylane/collections/qnode_collection.py	100.00% <0.00%> (+3.50%)	⬆️
pennylane/devices/default_qubit_torch.py	92.07% <0.00%> (+5.94%)	⬆️
pennylane/interfaces/execution.py	100.00% <0.00%> (+6.29%)	⬆️
pennylane/math/single_dispatch.py	98.84% <0.00%> (+7.33%)	⬆️
... and 11 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 207074e...441e2fe. Read the comment docs.

[rmoyard]

Great improvements 💯

[antalszava]

[sc-17363]