Skip to content

Releases: vm6502q/amazon-braket-qrack-simulator-python

"Uncompute" (bug/feature)

25 Feb 22:54
Compare
Choose a tag to compare

Bug fix / feature addition: "uncompute" after observable calculation (to allow different bases of observation for the same qubits).

Observables for 0 shots

25 Feb 20:11
Compare
Choose a tag to compare

Features added:

  • Probability and Variance observables for available bases can be calculated exactly for 0 shots.
  • Sample observable can be simulated for 0 shots

This release was tested by running the notebooks in the examples folder.

More observables

25 Feb 12:52
Compare
Choose a tag to compare

Features added:

  • Available observable bases are now (tensor products of) x, y, z, i, and h.
  • Expectation observables for available bases can be calculated exactly for 0 shots.

In a future release, exact calculation of Probability and Variance will require Qrack to expose the ProbBitsAll() method on QInterface. (The method already exists in the C++ code, but there was no demand for this method to be exposed via the shared library API, previously.)

Patch through AWS SDK observables from measurement shots

24 Feb 23:38
Compare
Choose a tag to compare

Features added (for reporting observables estimated from measurement shots):

  • Sample
  • Variance
  • Expectation

This release was tested by running the notebooks in the examples folder.

Debugging from examples

24 Feb 18:13
Compare
Choose a tag to compare

Bugs fixed:

  • Hex measurements are now parsed correctly
  • Extra simulator arguments are parsed correclty

This release was tested by running the notebooks in the examples folder.

First pre-release

23 Feb 23:01
Compare
Choose a tag to compare

This is the first (pre-)release of the Qrack Simulator for Amazon Braket SDK!

The (Amazon Braket SDK) Qrack Simulator is a Python open source library that provides an implementation of a quantum simulator that you can run locally. Qrack is a well-rounded simulator in development for over six years, with GPU acceleration, "ideal" and approximate modes, and novel approaches to canonical algorithms. Both OpenCL and CUDA back ends are available (supporting general GPU and accelerator vendors through the open standard of OpenCL).