Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

- Added yml files for MacOS and Linux CI

- Linked the status to README

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Add one-qubit IRTransformation for ion trap backend

Handle Honeywell Accelerator backend status

Currently, we terminate if the backend is not online.
This can be inconvenient for batch mode (vector<Composite>) since the backend can go into 'in maintenance' status.

Hence, changed the logic to only terminate if the status is 'offline' (according to the calendar).
Otherwise, try polling.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Limit QLM core count in power access mode (qlmaas)

The scheduler on the QLM will not allow more than 75 cores to be used for any single jobs.

Depending on the problem size (number of qubits), the QLM may pick an optimal core count exceeding this limit, hence needs to specifically limit the core count.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Updated the RO error decorator

Changed the order of retrieving ro probabilities: if a file path is explicitly specified, we should always use it rather than the data from the accelerator.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

IonQ Accelerator bug fix 

Fixing bug: closing brace before code

Users reported incorrect ro mitigation decorator.

This is due to the ro error probabilities are incorrect. The code to populate the meas1_prep0 data was accidentally placed outside the if scope, hence contain bogus data.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

- Batch mode: override base RemoteAccelerator to run one-by-one.

- Fixed U3 decompose for IonQProgramVisitor: looks like the decomposition is not correct (incorrect results). Also, IonQ supports native sqrt-X gate, hence just use the same U3 -> (sx-rz) as that of IBM.

Tested by: running on IonQ remote simulator.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Signed-off-by: Austin Adams <aja@gatech.edu>

Create plugin for GTRI testbed backend

Fixed https://github.com/eclipse/xacc/issues/491

Fixed a bug in single-qubit matrix decomposition

The sub-routine must return once it found the decomp.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Added tests

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Fixed https://github.com/eclipse/xacc/issues/463

More reliable finding of preamble.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>
Signed-off-by: Thien Nguyen <thien.md.nguyen@gmail.com>

Added some missing headers for GCC 11 compatibility

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Adding <exception> include to be compatible w/ gcc 11.1

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Signed-off-by: Alex McCaskey <mccaskeyaj@ornl.gov>

Signed-off-by: Alex McCaskey <mccaskeyaj@ornl.gov>

Signed-off-by: Alex McCaskey <mccaskeyaj@ornl.gov>

Signed-off-by: Alex McCaskey <mccaskeyaj@ornl.gov>

Fixed https://github.com/eclipse/xacc/issues/466

Cherrypicking the eigen fix to XACC:

https://gitlab.com/libeigen/eigen/-/commit/66d073c38e3cd5dad974deea7b3d1d45247ea55b



Updating the whole eigen to the latest version brings other unneccesary compatibility issue.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Adding layer metadata to gate node in the DAG representation

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

The string output now contains the layer info.

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Signed-off-by: Thien Nguyen <nguyentm@ornl.gov>

Add some instructions for common ion trap native gates

AcceleratorBuffer: Use exp-val-z extra info for getExpectationValueZ() if present