qrack

qrack copied to clipboard

QEngineShard as single-level binary decision tree

1

Learning from `QBdt`, we see that `QEngineShard` could attach two `QInterface` instances "under" each of its two probability amplitudes, as by multiplication of each `QEngineShard` probability amplitude on their respective...

WrathfulSpatula

Coverage update

2

I haven't checked coverage, lately, which should always be a priority for us. Largely, I know most new optimizations have been refactors of already-covered code, but the numbers have likely...

WrathfulSpatula

Feature: Optional CUDA Support

11

I'm a big fan of the OpenCL standard, and I've said in the past that basically none of the proprietary GPU acceleration standards appeal to me, with the specific possible...

WrathfulSpatula

BDT as QPager "global qubits"

8

A single `QPager` "global qubit" is basically the same thing as a `QBinaryDecisionTree` "node," if the "global qubit" additionally has a normalized pair of scale factors for its two branches....

WrathfulSpatula

"Clifford-permeable domain wall" in QStabilizerHybrid

2

`QBdt` is unstable with `Attach()`, though it is stable without any calls to that method. As I've recently spoken about a "Clifford-permeable domain wall" between stabilizer and (universal) heterogeneous other...

WrathfulSpatula

`QUnit::CINC()` is currently failing a unit test, along with other ALU tests. The change seems to have been precipitated by work with `TrimControls()`, although QUnit ALU methods relied on a...

WrathfulSpatula

Qrack set out trying to avoid _all_ dependency, at its inception, but we have since allowed for a number of very useful _optional_ dependency based features. BLAS/LAPACK would probably be...

WrathfulSpatula

Mixed FP precision for QUnit subsystems

4

Basically, Qrack accepts as standard that fp16 is naturally accurate up to 16qb, fp32 is valid up to 32qb, fp64 is valid up to 64qb... etc.. If we allowed precision...

WrathfulSpatula

Update all plugin support

1

Plugin support has fallen out of date, and it's time to update it, across the board.

WrathfulSpatula

Extend `ZMask()` to `PhaseRootNMask()`

1

There's an obvious extension of `ZMask()` to "masked" `S`, `T`, and general `PhaseRootN` gates.`ZMask()` calculates parity of each masked permutation, and applies a `-1` multiplicative factor if the parity is...

WrathfulSpatula