Performance fix

Improved QML blocks execution performance.

QML block support and fixes

• Supported qureg state expectations and fixed terminology for probability measures.
• Increased qSim capabilities supporting QML blocks execution.
• Improved socket transfer performance (10x w.r.t. previous setup - now ~100 msg/s).

Updated repo with Python access client

Code fixes for few bugs.
Python access client module added.

First repo release

• First complete release for GPU CUDA devices
• Supported 1-qubit and 2-qubits core function sets and handled multi-controlled long-short range cases.
• Finalised QASM message structure for handling multi-control / short-long range fuction params.
• Supported function blocks for macro transformations handling concerning:
  • n-qubits swap and c-swap
• qSim client access documented in "qSim_How_To_User_Guide".
• Supported a "cpu-mode" for run on non-GPU devices, to be activated as make macro (QSIM_CPU).