qgear

run_qiskit_aer

 run_qiskit_aer (qcL, shots)

Run Qiskit Aer simulation

source

run_cudaqft

 run_cudaqft (shots, num_gpus, num_qubit, nc=1, target='nvidia', verb=1)

Run CUDA-Q QFT circuits

source

input_shard

 input_shard (bigD, myRank, numRank, verb=1)

Shard dataset across MPI ranks

source

run_gate_job

 run_gate_job (exp:str, backend:str='nvidia', numshots:int=1024,
               basePath:str=None, qft:bool=False,
               target_option:str='fp32', verbosity:int=1)

*Run a gate list experiment.

Args: exp: Experiment name (without .gate_list.h5) backend: Backend to use (‘nvidia’, ‘qiskit-cpu’, ‘tensornet’, ‘qpp-cpu’) numshots: Shots per circuit basePath: Base directory for input/output (or ‘env’ to use $Cudaq_dataVault) qft: If True, run QFT kernel instead of gate list target_option: Target options (default ‘fp32’) verbosity: Verbosity level (0-3)*

source

run_qcrank

 run_qcrank (circ_name:str, inp_path:str='out', out_path:str='out',
             backend:str='nvidia', num_shot_per_addr:int=400,
             exp_name:str=None, verb:int=1)

*Run a QCrank simulation with CUDA-Q.

Args: circ_name: Circuit name without extension. inp_path: Path to input .qcrank_inp.h5 file. out_path: Directory for outputs. backend: CUDA-Q backend target. num_shot_per_addr: Shots per address. exp_name: Optional experiment name to override auto-generated job ID. verb: Verbosity level.*

source

run_cudaq

 run_cudaq (gateD, shots, verb=1, backend='qpp-cpu')

Run CUDA-Q simulation for all circuits in gateD.

source

harvest_cudaq_backRun_submitMeta

 harvest_cudaq_backRun_submitMeta (md, backend:str, exp_name:str=None)

Fill metadata with backend run info.

source

make_qcrank

 make_qcrank (md, barrier=True)

Create a parameterized QCrank circuit object.

source

canned_qcrank_inp

 canned_qcrank_inp (inp_path:str, circ_name:str, num_shot_per_addr:int)

Load prepacked QCrank HDF5 input and update metadata with shot count.

source

rank_print

 rank_print (*args, **kwargs)

Simplified rank_print (no MPI), always prints.