A fast quantum stabilizer circuit simulator.
Project description
Stim
Stim is a fast simulator for quantum stabilizer circuits.
API references are available on the stim github wiki: https://github.com/quantumlib/stim/wiki
Stim can be installed into a python 3 environment using pip:
pip install stim
Once stim is installed, you can import stim and use it.
There are three supported use cases:
- Interactive simulation with
stim.TableauSimulator. - High speed sampling with samplers compiled from
stim.Circuit. - Independent exploration using
stim.Tableauandstim.PauliString.
Interactive Simulation
Use stim.TableauSimulator to simulate operations one by one while inspecting the results:
import stim
s = stim.TableauSimulator()
# Create a GHZ state.
s.h(0)
s.cnot(0, 1)
s.cnot(0, 2)
# Look at the simulator state re-inverted to be forwards:
t = s.current_inverse_tableau()
print(t**-1)
# prints:
# +-xz-xz-xz-
# | ++ ++ ++
# | ZX _Z _Z
# | _X XZ __
# | _X __ XZ
# Measure the GHZ state.
print(s.measure_many(0, 1, 2))
# prints one of:
# [True, True, True]
# or:
# [False, False, False]
High Speed Sampling
By creating a stim.Circuit and compiling it into a sampler, samples can be generated very quickly:
import stim
# Create a circuit that measures a large GHZ state.
c = stim.Circuit()
c.append_operation("H", [0])
for k in range(1, 30):
c.append_operation("CNOT", [0, k])
c.append_operation("M", range(30))
# Compile the circuit into a high performance sampler.
sampler = c.compile_sampler()
# Collect a batch of samples.
# Note: the ideal batch size, in terms of speed per sample, is roughly 1024.
# Smaller batches are slower because they are not sufficiently vectorized.
# Bigger batches are slower because they use more memory.
batch = sampler.sample(1024)
print(type(batch)) # numpy.ndarray
print(batch.dtype) # numpy.uint8
print(batch.shape) # (1024, 30)
print(batch)
# Prints something like:
# [[1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]
# [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
# [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]
# ...
# [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]
# [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]
# [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
# [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]]
This also works on circuits that include noise:
import stim
import numpy as np
c = stim.Circuit("""
X_ERROR(0.1) 0
Y_ERROR(0.2) 1
Z_ERROR(0.3) 2
DEPOLARIZE1(0.4) 3
DEPOLARIZE2(0.5) 4 5
M 0 1 2 3 4 5
""")
batch = c.compile_sampler().sample(2**20)
print(np.mean(batch, axis=0).round(3))
# Prints something like:
# [0.1 0.2 0. 0.267 0.267 0.266]
You can also sample annotated detection events using stim.Circuit.compile_detector_sampler.
For a list of gates that can appear in a stim.Circuit, see the latest readme on github.
Independent Exploration
Stim provides data types stim.PauliString and stim.Tableau, which support a variety of fast operations.
import stim
xx = stim.PauliString("XX")
yy = stim.PauliString("YY")
assert xx * yy == -stim.PauliString("ZZ")
s = stim.Tableau.from_named_gate("S")
print(repr(s))
# prints:
# stim.Tableau.from_conjugated_generators(
# xs=[
# stim.PauliString("+Y"),
# ],
# zs=[
# stim.PauliString("+Z"),
# ],
# )
s_dag = stim.Tableau.from_named_gate("S_DAG")
assert s**-1 == s_dag
assert s**1000000003 == s_dag
cnot = stim.Tableau.from_named_gate("CNOT")
cz = stim.Tableau.from_named_gate("CZ")
h = stim.Tableau.from_named_gate("H")
t = stim.Tableau(5)
t.append(cnot, [1, 4])
t.append(h, [4])
t.append(cz, [1, 4])
t.prepend(h, [4])
assert t == stim.Tableau(5)
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file stim-1.4.0.tar.gz.
File metadata
- Download URL: stim-1.4.0.tar.gz
- Upload date:
- Size: 135.6 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.4.1 importlib_metadata/4.6.0 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.61.1 CPython/3.8.8
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | eb6911290d29110abc1c0aee39d02f64ae4e9d3042fe1dac42d8a2aedab3e4fe |
|
| MD5 | c60fad1b2fd33618045411d61d35cc79 |
|
| BLAKE2b-256 | 910bf976af8fa92bbe04569e3c243b6f5a28395b66ec5281e7cbb31a4d0e47dd |
