Noisy simulations on GPU

Noisy simulations running in "stochastic" mode can be offloaded to run on a GPU. As in the case of ideal simulations, to use the GPU based simulator, it is sufficient to set the fields "use_GPU" (to 'True') and "precision" (to either 1 or 2) while initializing the 'NoisyQProc'. In this notebook we show some use cases, where we obtain a speedup in running the simulations on a GPU.

import numpy as np

from qat.core import Observable, Term
from qat.lang.AQASM import Program
from qat.lang.AQASM.qftarith import QFT
from qat.hardware import DefaultHardwareModel
from qat.quops.quantum_channels import ParametricAmplitudeDamping, ParametricPureDephasing
from qlmaas.noisy.noisy_simulation import compute_fidelity
from qlmaas.qpus import NoisyQProc

As an example, we consider the Quantum Fourier Transform (QFT) circuit and we suppose a simple noise model with idle qubits subject to parametric amplitude damping.

nqbits = 14
prog = Program()
reg = prog.qalloc(nqbits)
prog.apply(QFT(nqbits), reg)
circ = prog.to_circ()

hardware_model = DefaultHardwareModel(gate_times = {"H": 0.2, "C-PH": lambda angle:0.65},
                                      idle_noise = [ParametricAmplitudeDamping(T_1 = 75)])    

Here we choose to run the simulation with 1000 samples and initialize the NoisyQProc with different arguments to compare the output and their respective runtimes.

n_samples = 1000

noisy_qpu_gpu_single = NoisyQProc(hardware_model=hardware_model, sim_method="stochastic", 
                                  n_samples=n_samples, use_GPU=True, precision=1)
noisy_qpu_gpu_double = NoisyQProc(hardware_model=hardware_model, sim_method="stochastic", 
                                  n_samples=n_samples, use_GPU=True, precision=2)
noisy_qpu_cpu = NoisyQProc(hardware_model=hardware_model, sim_method="stochastic", n_samples=n_samples)

I Fidelity of noisy QFT

%%time
fid_cpu, err_cpu = compute_fidelity(circ, noisy_qpu_cpu) # Simulation running on 1 node (cpu)
print(fid_cpu, err_cpu)

Submitted a new batch: Job96

0.9039919448391515 0.010890140359847728
CPU times: user 6.8 ms, sys: 3.56 ms, total: 10.4 ms
Wall time: 6.43 s

%%time
fid_gpu_single, err_gpu_single = compute_fidelity(circ, noisy_qpu_gpu_single) # Simulation running in single precision on a GPU
print(fid_gpu_single, err_gpu_single)

Submitted a new batch: Job105

0.9177095565677693 0.010437858305897487
CPU times: user 7.1 ms, sys: 3.81 ms, total: 10.9 ms
Wall time: 4.32 s

%%time
fid_gpu_double, err_gpu_double = compute_fidelity(circ, noisy_qpu_gpu_double) # Simulation running in double precision on a GPU
print(fid_gpu_double, err_gpu_double)

Submitted a new batch: Job107

0.89512193880791 0.010808849469555416
CPU times: user 5.44 ms, sys: 2.53 ms, total: 7.97 ms
Wall time: 4.27 s

Here we compare the stochastic results with a deterministic evaluation and check the time it takes to get an exact value.

noisy_qpu_det = NoisyQProc(hardware_model=hardware_model, sim_method="deterministic-vectorized")

%%time
fid_cpu_det, _ = compute_fidelity(circ, noisy_qpu_det) # Deterministic simulation running on 1 node (cpu)
print(fid_cpu_det)

Submitted a new batch: Job109

0.9076533148415741
CPU times: user 62.5 ms, sys: 16.8 ms, total: 79.3 ms
Wall time: 1min 28s

II Sampling a noisy QFT

job = circ.to_job(nbshots=100)

%%time
res_cpu = noisy_qpu_cpu.submit(job)

Submitted a new batch: Job176
CPU times: user 3.59 ms, sys: 2.52 ms, total: 6.11 ms
Wall time: 143 ms

%%time
res_gpu_single = noisy_qpu_gpu_single.submit(job)

Submitted a new batch: Job177
CPU times: user 2.13 ms, sys: 73 μs, total: 2.21 ms
Wall time: 144 ms

%%time
res_gpu_double = noisy_qpu_gpu_double.submit(job)

Submitted a new batch: Job178
CPU times: user 2.53 ms, sys: 2.18 ms, total: 4.71 ms
Wall time: 145 ms

III Observable evaluation

Here we generate a random observable with 40 terms and evaluate it

n_terms = 40
terms = ["X", "Y", "Z"]
pauli_terms = []
for _ in range(n_terms):
    term = ""
    for _ in range(np.random.choice([1, 2], 1)[0]):
        term += np.random.choice(terms, 1)[0]
    pauli_terms.append(Term(1.0, term, list(np.random.choice(nqbits, len(term), replace=False))))
    
obs = Observable(nqbits, pauli_terms=pauli_terms)

job_obs = circ.to_job("OBS", observable=obs)

%%time
res_cpu_obs = noisy_qpu_cpu.submit(job_obs)
print(res_cpu_obs.value, res_cpu_obs.error)

Submitted a new batch: Job179

9.572164953530006 0.07283062670590353
CPU times: user 8.45 ms, sys: 0 ns, total: 8.45 ms
Wall time: 4.34 s

%%time
res_gpu_single_obs = noisy_qpu_gpu_single.submit(job_obs)
print(res_gpu_single_obs.value, res_gpu_single_obs.error)

Submitted a new batch: Job188

9.668764715785713 0.07317348698271015
CPU times: user 7.93 ms, sys: 3.57 ms, total: 11.5 ms
Wall time: 4.3 s

%%time
res_gpu_double_obs = noisy_qpu_gpu_double.submit(job_obs)
print(res_gpu_double_obs.value, res_gpu_double_obs.error)

Submitted a new batch: Job196

9.472209908673358 0.07372277468206316
CPU times: user 8.19 ms, sys: 1.57 ms, total: 9.76 ms
Wall time: 4.31 s