|
1 | | - |
2 | | -Density Matrix Simulators |
| 1 | +Noisy Simulators |
3 | 2 | ================================== |
4 | 3 |
|
| 4 | +Trajectory Noisy Simulation |
| 5 | +++++++++++++++++++++++++++++++++++ |
| 6 | + |
| 7 | +The :code:`nvidia` target supports noisy quantum circuit simulations using |
| 8 | +quantum trajectory method across all configurations: single GPU, multi-node |
| 9 | +multi-GPU, and with host memory. When simulating many trajectories with small |
| 10 | +state vectors, the simulation is batched for optimal performance. |
| 11 | + |
| 12 | +When a :code:`noise_model` is provided to CUDA-Q, the :code:`nvidia` target |
| 13 | +will incorporate quantum noise into the quantum circuit simulation according |
| 14 | +to the noise model specified. |
| 15 | + |
| 16 | + |
| 17 | +.. tab:: Python |
| 18 | + |
| 19 | + .. literalinclude:: ../../../snippets/python/using/backends/trajectory.py |
| 20 | + :language: python |
| 21 | + :start-after: [Begin Docs] |
| 22 | + |
| 23 | + .. code:: bash |
| 24 | + |
| 25 | + python3 program.py |
| 26 | + { 00:15 01:92 10:81 11:812 } |
| 27 | +
|
| 28 | +.. tab:: C++ |
| 29 | +
|
| 30 | + .. literalinclude:: ../../../snippets/cpp/using/backends/trajectory.cpp |
| 31 | + :language: cpp |
| 32 | + :start-after: [Begin Documentation] |
| 33 | +
|
| 34 | + .. code:: bash |
| 35 | +
|
| 36 | + nvq++ --target nvidia program.cpp [...] -o program.x |
| 37 | + ./program.x |
| 38 | + { 00:15 01:92 10:81 11:812 } |
| 39 | +
|
| 40 | +
|
| 41 | +In the case of bit-string measurement sampling as in the above example, each measurement 'shot' is executed as a trajectory, whereby Kraus operators specified in the noise model are sampled. |
| 42 | +
|
| 43 | +For observable expectation value estimation, the statistical error scales asymptotically as :math:`1/\sqrt{N_{trajectories}}`, where :math:`N_{trajectories}` is the number of trajectories. |
| 44 | +Hence, depending on the required level of accuracy, the number of trajectories can be specified accordingly. |
| 45 | +
|
| 46 | +.. tab:: Python |
| 47 | +
|
| 48 | + .. literalinclude:: ../../../snippets/python/using/backends/trajectory_observe.py |
| 49 | + :language: python |
| 50 | + :start-after: [Begin Docs] |
| 51 | +
|
| 52 | + .. code:: bash |
| 53 | + |
| 54 | + python3 program.py |
| 55 | + Noisy <Z> with 1024 trajectories = -0.810546875 |
| 56 | + Noisy <Z> with 8192 trajectories = -0.800048828125 |
| 57 | +
|
| 58 | +.. tab:: C++ |
| 59 | +
|
| 60 | + .. literalinclude:: ../../../snippets/cpp/using/backends/trajectory_observe.cpp |
| 61 | + :language: cpp |
| 62 | + :start-after: [Begin Documentation] |
| 63 | +
|
| 64 | + .. code:: bash |
| 65 | +
|
| 66 | + nvq++ --target nvidia program.cpp [...] -o program.x |
| 67 | + ./program.x |
| 68 | + Noisy <Z> with 1024 trajectories = -0.810547 |
| 69 | + Noisy <Z> with 8192 trajectories = -0.800049 |
| 70 | +
|
| 71 | +
|
| 72 | +The following environment variable options are applicable to the :code:`nvidia` target for trajectory noisy simulation. Any environment variables must be set |
| 73 | +prior to setting the target. |
| 74 | +
|
| 75 | +.. list-table:: **Additional environment variable options for trajectory simulation** |
| 76 | + :widths: 20 30 50 |
| 77 | +
|
| 78 | + * - Option |
| 79 | + - Value |
| 80 | + - Description |
| 81 | + * - ``CUDAQ_OBSERVE_NUM_TRAJECTORIES`` |
| 82 | + - positive integer |
| 83 | + - The default number of trajectories for observe simulation if none was provided in the `observe` call. The default value is 1000. |
| 84 | + * - ``CUDAQ_BATCH_SIZE`` |
| 85 | + - positive integer or `NONE` |
| 86 | + - The number of state vectors in the batched mode. If `NONE`, the batch size will be calculated based on the available device memory. Default is `NONE`. |
| 87 | + * - ``CUDAQ_BATCHED_SIM_MAX_BRANCHES`` |
| 88 | + - positive integer |
| 89 | + - The number of trajectory branches to be tracked simultaneously in the gate fusion. Default is 16. |
| 90 | + * - ``CUDAQ_BATCHED_SIM_MAX_QUBITS`` |
| 91 | + - positive integer |
| 92 | + - The max number of qubits for batching. If the qubit count in the circuit is more than this value, batched trajectory simulation will be disabled. The default value is 20. |
| 93 | + * - ``CUDAQ_BATCHED_SIM_MIN_BATCH_SIZE`` |
| 94 | + - positive integer |
| 95 | + - The minimum number of trajectories for batching. If the number of trajectories is less than this value, batched trajectory simulation will be disabled. Default value is 4. |
| 96 | +
|
| 97 | +.. note:: |
| 98 | + |
| 99 | + Batched trajectory simulation is only available on the single-GPU execution mode of the :code:`nvidia` target. |
| 100 | + |
| 101 | + If batched trajectory simulation is not activated, e.g., due to problem size, number of trajectories, or the nature of the circuit (dynamic circuits with mid-circuit measurements and conditional branching), the required number of trajectories will be executed sequentially. |
| 102 | +
|
| 103 | +
|
5 | 104 |
|
6 | 105 | Density Matrix |
7 | 106 | ++++++++++++++++ |
|
0 commit comments