Micro-optimize basis state conversion

[josh146]

Context:

• In fix: convert measurements to indices without allocating a high-dimens… amazon-braket/amazon-braket-sdk-python#203, it was noted that the current implementation for converting between binary samples and the equivalent base-10 representation does not work for devices with more than 32 qubits.

Description of the Change:

• Modifies QubitDevice, replacing the current conversion logic with one that works for > 32 qubits.

Benefits:

• Samples can now be processed if the devices has more than 32 qubits

• The new logic is easier to read and understand than np.ravel_multi_index

• The new logic is ever so slightly faster:

  

Possible Drawbacks:

There are two other places where np.ravel_multi_index is used:

• default.qubit, when applying a statevector to a subset of wires
• default.mixed, same as above.

It might be worth exploring if a better implementation can be used here as well.

Related GitHub Issues: n/a

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.73%. Comparing base (a4060d6) to head (3d46779).
Report is 2634 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1088   +/-   ##
=======================================
  Coverage   97.73%   97.73%           
=======================================
  Files         154      154           
  Lines       11679    11681    +2     
=======================================
+ Hits        11414    11416    +2     
  Misses        265      265           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[chaserileyroberts]

Let me grab a before/after benchmark from asv

[chaserileyroberts]

This is very clever!