Intel® Quantum SDK

Release Notes

• Replaced repeated invocations of compiler optimizations with a single binary that results in:
  • Reduced compilation time for large programs.
  • Additional warnings for potential invalid blocks.
  • Reduced file system access.
  • A new driver that does not require any changes from the user in normal use. 
• Changed behavior for generating intermediate representation (IR) files:
  • The -k flag no longer generates IR files at each stage of compilation. It only outputs the final classical and quantum IR.
  • New flags interrupt or start compilation flow at a specified step:
    • -z (lowercase "z") denotes the starting stage for compilation.
    • -Z (uppercase "Z") denotes the ending stage for compilation.

Release Notes

• Addition of quantum expression (QExpr) and FLEQ:
  • Version 1.1 introduces QExpr, a new data type for expressing quantum instructions, and APIs to express algorithms using a functional programming paradigm.
  • FLEQ provides a powerful, succinct tool for defining quantum algorithms and APIs.
• Addition of more qubit simulation back ends:
  • Tensor network back end: A new target back end uses tensor network methods to simulate quantum circuits with the ability to simulate large numbers of qubits for certain types of circuits.
  • Clifford circuit back end: A new target back end that provides extremely fast results for simulating workloads with only Clifford quantum gates.
  • A user-defined back end: A new interface provides a route to implementing your own qubit simulations as a back end for quantum workloads.
    
• Performance improvements to the existing back ends for Intel Quantum Simulator and the Quantum Dot Simulator.
• A new API provides the capability to add your own noise models for the qubits simulated through the Intel Quantum Simulator.
• When writing a quantum_kernel, developers can declare variables of type qbit in the local scope.
• Enhancement of qubit scheduling and placement, where users can now select between predefined algorithms for:
  • Mapping software qubits (virtual) to hardware qubits (physical)
  • Determining the order of quantum gate and instruction completion
• Added the ability to run user-selected passes and external passes:
  • A new set of compiler features allows you to run optimization passes from LLVM, the Intel® Quantum Compiler, or external libraries. These can be added in between any or all of the unrolling, validation, synthesis, scheduling, or lowering steps of quantum compilation.
  • Users can now run their own passes developed using the open sourced front end for the Intel Quantum Compiler.
    
    

Documentation

Intel Quantum SDK 

Release Notes

• Addition of the Quantum Dot Simulator back end supports a new target back end for the full stack in simulation. This is a physics-based simulation of quantum dot qubits as well as simulation of control electronics.
• Addition of placement and scheduler passes provides automatic mapping and scheduling of quantum algorithms onto qubits. These passes treat the problem of mapping qubit-type variables onto the qubits as a physical resource and take into account the connectivity between qubits.
• Added support for Open Quantum Assembly Language (OpenQASM) 2.0 provides a source-to-source translator module that enables you to directly use programs that you have written to output OpenQASM 2.0.
• Added an interface to Python* to support interacting with the Python parts of the quantum computing ecosystem through a choice of writing and compiling OpenQASM 2.0 or by compiling your own shared object.
• Changes to the API naming convention follows LLVM-style to provide predictability in the future. Several of the existing methods' names in the API of the Intel Quantum SDK were standardized so that they are compliant. This onetime breaking of the user API helps increase future user friendliness.