Skip to content

Latest commit

 

History

History
757 lines (541 loc) · 24.2 KB

advanced_usage.rst

File metadata and controls

757 lines (541 loc) · 24.2 KB
Raw

Advanced usage

Note

If you're running locally, remember set up the QVM and quilc in server mode before trying to use them: server.

pyQuil configuration

:py~pyquil.api.QCSClient instructs pyQuil on how to connect with the components needed to compile and run programs (quilc, qvm, and QCS). Any APIs that take a configuration object as input (e.g. :py~pyquil.api.get_qc) typically do so optionally, so that a default configuration can be loaded for you if one is not provided. You can override this default configuration by either instantiating your own :py~pyquil.api.QCSClient object and providing it as input to the function in question, or by setting the QCS_SETTINGS_FILE_PATH and/or QCS_SECRETS_FILE_PATH environment variables to have pyQuil load its settings and secrets from specific locations. By default, configuration will be loaded from $HOME/.qcs/settings.toml and $HOME/.qcs/secrets.toml.

Additionally, you can override whichever QVM and quilc URLs are loaded from settings.toml (profiles.<profile>.applications.pyquil.qvm_url and profiles.<profile>.applications.pyquil.quilc_url fields) by setting the QCS_SETTINGS_APPLICATIONS_QVM_URL and/or QCS_SETTINGS_APPLICATIONS_QUILC_URL environment variables. If these URLs are missing from settings.toml and are not set by environment variables, the following defaults will be used (as they correspond to the default behavior of the QVM and quilc when running locally):

  • QVM URL: http://127.0.0.1:5000
  • quilc URL: tcp://127.0.0.1:5555

Multithreading

:py~pyquil.api.QuantumComputer objects are safe to share between threads, enabling you to execute and retrieve results for multiple programs or parameter values at once. Note that :py~pyquil.Program and :py~pyquil.api.EncryptedProgram are not thread-safe, and should be copied (with copy()) before use in a concurrent context.

Note

The QVM processes incoming requests in parallel, while a QPU may process them sequentially or in parallel (depending on the qubits used). If you encounter timeouts while trying to run large numbers of programs against a QPU, try increasing the execution_timeout parameter on calls to :py~pyquil.get_qc (specified in seconds).

Note

We suggest running jobs with a minimum of 2x parallelism, so that the QVM or QPU is fully occupied while your program runs and no time is wasted in between jobs.

Note

Because pyQuil does not currently have an asyncio API it is recommended to use ThreadPools.

Below is an example that demonstrates how to use pyQuil in a multithreading scenario:

from multiprocessing.pool import ThreadPool

from pyquil import get_qc, Program
from pyquil.api import QCSClient

qc = get_qc("Aspen-M-3")


def run(program: Program):
    return qc.run(qc.compile(program)).get_register_map().get("ro")


programs = [
    Program(
        "DECLARE ro BIT",
        "RX(pi) 0",
        "MEASURE 0 ro",
    ).wrap_in_numshots_loop(10),
] * 20

with ThreadPool(5) as pool:
    results = pool.map(run, programs)

for i, result in enumerate(results):
    print(f"Results for program {i}:\n{result}\n")

Alternative QPU endpoints

Rigetti QCS supports alternative endpoints for access to a QPU architecture, useful for very particular cases. Generally, this is useful to call "mock" or test endpoints, which simulate the results of execution for the purposes of integration testing without the need for an active reservation or contention with other users. See the QCS API Docs for more information on QPU Endpoints.

To be able to call these endpoints using pyQuil, enter the endpoint_id of your desired endpoint in one of the sites where quantum_processor_id is used:

# Option 1
qc = get_qc("Aspen-M-3", endpoint_id="my_endpoint")

# Option 2
qam = QPU(quantum_processor_id="Aspen-M-3", endpoint_id="my_endpoint")

After doing so, for all intents and purposes - compilation, optimization, etc - your program will behave the same as when using "default" endpoint for a given quantum processor, except that it will be executed by an alternate QCS service, and the results of execution should not be treated as correct or meaningful.

Using libquil for Quilc and QVM

Note

This feature is experimental and may not work for all platforms.

libquil provides the functionality of Quilc and QVM in a library that can be used without having to run Quilc and QVM as servers, which can make developing with pyQuil easier.

To use libquil, first follow its installation instructions. Once libquil and its dependencies are installed, you will need to run the following command to install a compatible version of qcs-sdk-python:

poetry run pip install --config-settings=build-args='--features libquil' qcs-sdk-python --force-reinstall --no-binary qcs-sdk-python

You can then check that libquil is available to pyQuil by executing the following Python code

from pyquil.diagnostics import get_report
print(get_report())

Towards the end of the output, you will see a libquil section like below

libquil:
    available: true
    quilc version: 1.27.0
    qvm version: 1.17.2 (077ba23)

If you do not see available: true then re-try installation. If you continue to have issues, please report them on github.

If installation was successful, you can now use libquil in pyQuil: the get_qc function provides two keyword parameters quilc_client and qvm_client which can be set to use libquil:

from pyquil import get_qc
from qcs_sdk.compiler.quilc import QuilcClient
from qcs_sdk.qvm import QVMClient

qc = get_qc("8q-qvm", quilc_client=QuilcClient.new_libquil(), qvm_client=QVMClient.new_libquil())

Please report issues on github.

Using qubit placeholders

Note

The functionality provided inline by QubitPlaceholders is similar to writing a function which returns a Program, with qubit indices taken as arguments to the function.

In pyQuil, we typically use integers to identify qubits

placeholders

from pyquil import Program from pyquil.gates import CNOT, H print(Program(H(0), CNOT(0, 1)))

placeholders

H 0 CNOT 0 1

However, when running on real, near-term QPUs we care about what particular physical qubits our program will run on. In fact, we may want to run the same program on an assortment of different qubits. This is where using QubitPlaceholders comes in.

placeholders

from pyquil import Program from pyquil.gates import H, CNOT

placeholders

from pyquil.quilatom import QubitPlaceholder q0 = QubitPlaceholder() q1 = QubitPlaceholder() p = Program(H(q0), CNOT(q0, q1)) print(p)

placeholders

H Placeholder(QubitPlaceholder(0x...)) CNOT Placeholder(QubitPlaceholder(0x...)) Placeholder(QubitPlaceholder(0x...))

H Placeholder(QubitPlaceholder(0x600002DEB5B0)) CNOT Placeholder(QubitPlaceholder(0x600002DEB5B0)) Placeholder(QubitPlaceholder(0x600002DEABB0))

Addressing qubits

If your program uses QubitPlaceholders, the placeholders must be resolved before your program can be run. If you try to run a program with unresolved placeholders, you will get an error:

print(p.out())

RuntimeError: Qubit q4402789176 has not been assigned an index

Instead, you must explicitly map the placeholders to physical qubits. By default, the function :py~pyquil.quil.address_qubits will address qubits from 0 to N, skipping indices that are already used in the program.

placeholders

from pyquil.quil import address_qubits print(address_qubits(p))

placeholders

H 0 CNOT 0 1

The real power comes into play when you provide an explicit mapping:

placeholders

print(address_qubits(p, qubit_mapping={

q0: 14, q1: 19,

}))

placeholders

H 14 CNOT 14 19

As an alternative to a mapping, you can consider using :py~pyquil.quil.Program.resolve_placeholders_with_custom_resolvers. This method accepts any function that takes a placeholder as an argument, and returns a fixed value for that placeholder (or None, if you want it to remain unresolved).

placeholders

from typing import Optional from pyquil import Program, get_qc from pyquil.gates import H, CNOT from pyquil.quilatom import QubitPlaceholder

placeholders

q0 = QubitPlaceholder() q1 = QubitPlaceholder() p = Program(H(q0), CNOT(q0, q1)) qc = get_qc("2q-qvm")

def qubit_resolver(placeholder: QubitPlaceholder) -> Optional[int]:
if placeholder == q0:

return 0

if placeholder == q1:

return None

p.resolve_placeholders_with_custom_resolvers(qubit_resolver=qubit_resolver) print(p)

placeholders

H 0 CNOT 0 Placeholder(...)

Requesting a register of qubit placeholders

Usually, your algorithm will use an assortment of qubits. You can use the convenience function :py~pyquil.quilatom.QubitPlaceholder.register to request a register of qubits to build your program.

register

from pyquil import Program from pyquil.gates import H from pyquil.quilatom import QubitPlaceholder from pyquil.quil import address_qubits

register

qbyte = QubitPlaceholder.register(8) p_evens = Program(H(q) for q in qbyte) print(address_qubits(p_evens, {q: i*2 for i, q in enumerate(qbyte)}))

register

H 0 H 2 H 4 H 6 H 8 H 10 H 12 H 14

Classical control flow

Here are a couple quick examples that show how much richer a Quil program can be with classical control flow.

Warning

Dynamic control flow can have unexpected effects on readout data. See accessing_raw_execution_data for more information.

While loops

In this first example, we create a while loop by following these steps:

  1. Declare a register called flag_register to use as a boolean test for looping.
  2. Initialize this register to 1, so our while loop will execute. This is often called the loop preamble or loop initialization.
  3. Write the body of the loop in its own :py~pyquil.quil.Program. This will be a program that applies an X gate followed by an H gate on our qubit.
  4. Use the :py~pyquil.quil.Program.while_do method to add control flow.
  5. Call :py~pyquil.quil.Program.resolve_label_placeholders to resolve the label placeholders inserted by while_do.

control-flow

from pyquil import Program from pyquil.gates import *

# Initialize the Program and declare a 1 bit memory space for our boolean flag outer_loop = Program() flag_register = outer_loop.declare('flag_register', 'BIT')

# Set the initial flag value to 1 outer_loop += MOVE(flag_register, 1)

# Define the body of the loop with a new Program inner_loop = Program() inner_loop += Program(X(0), H(0)) inner_loop += MEASURE(0, flag_register)

# Run inner_loop in a loop until flag_register is 0 outer_loop.while_do(flag_register, inner_loop) outer_loop.resolve_label_placeholders()

print(outer_loop)

control-flow

DECLARE flag_register BIT[1] MOVE flag_register[0] 1 LABEL @START_0 JUMP-UNLESS @END_0 flag_register[0] X 0 H 0 MEASURE 0 flag_register[0] JUMP @START_0 LABEL @END_0

Notice that the outer_loop program applied a Quil instruction directly to a classical register. There are several classical commands that can be used in this fashion:

  • NOT which flips a classical bit
  • AND which operates on two classical bits
  • IOR which operates on two classical bits
  • MOVE which moves the value of a classical bit at one classical address into another
  • EXCHANGE which swaps the value of two classical bits

Note

The approach documented here can be used to construct a "numshots" loop in pure Quil. See the :py~pyquil.quil.Program.with_loop method and build_a_fixed_count_loop for more information.

If, then

In this next example, we show how to do conditional branching in the form of the traditional if construct as in many programming languages. Much like the last example, we construct programs for each branch of the if, and put it all together by using the :py~pyquil.quil.Program.if_then method.

control-flow

# Declare our memory spaces branching_prog = Program() ro = branching_prog.declare('ro', 'BIT') test_register = branching_prog.declare('test_register', 'BIT')

# Construct each branch of our if-statement. We can have empty branches # simply by having empty programs. then_branch = Program(X(0)) else_branch = Program()

# Construct our program so that the result in test_register is equally likely to be a 0 or 1 branching_prog += H(1) branching_prog += MEASURE(1, test_register)

# Add the conditional branching branching_prog.if_then(test_register, then_branch, else_branch)

# Measure qubit 0 into our readout register branching_prog += MEASURE(0, ro) branching_prog.resolve_label_placeholders()

print(branching_prog)

control-flow

DECLARE ro BIT[1] DECLARE test_register BIT[1] H 1 MEASURE 1 test_register[0] JUMP-WHEN @THEN_0 test_register[0] JUMP @END_0 LABEL @THEN_0 X 0 LABEL @END_0 MEASURE 0 ro[0]

We can run this program a few times to see what we get in the readout register ro.

control-flow

from pyquil import get_qc

qc = get_qc("2q-qvm") branching_prog.wrap_in_numshots_loop(10) result = qc.run(branching_prog) print(result.get_register_map()['test_register'])

control-flow

[[...]

[...] [...] [...] [...] [...] [...] [...] [...] [...]]

[[1]

[1] [1] [0] [1] [0] [0] [1] [1] [0]]

Sentinel based loop

Now that we understand how to create loops and conditionals, we can put them together to create a sentinel controlled loop. That is, we'll repeat the body of a program until a certain condition is met. In this example, we'll use the classic bell state program to demonstrate the concept. However, this technique can be applied to any program with a probabilistic outcome that we want to repeat until we get a desired result.

To start, let's import everything we'll need:

sentinel-based-loop

# Import some types we'll use from typing import Optional, Tuple

# We'll use numpy to help us validate our results import numpy as np

# We'll need to create a program and define an executor from pyquil import Program, get_qc # We'll use these gates in our program from pyquil.gates import CNOT, H, X # We'll also need the help of a few control flow instructions from pyquil.quilbase import Halt, Qubit, MemoryReference, JumpTarget, Jump from pyquil.quilatom import Label

Building our program

Adding control flow to a program introduces complexity, especially as we add more branches to the program. To manage this complexity we'll use some of the methods we learned about in the previous sections as well as by breaking down the program into its constituent parts.

The program body

First, let's define the body of our program. This is the part of the program that we'll repeat until we get the result we desire. In this case, we'll create a bell state between two qubits and measure them:

sentinel-based-loop

def body(qubits: Tuple[Qubit, Qubit], measures: MemoryReference) -> Program:

"""Constructs a bell state between the given qubit and measures them into the given memory reference.""" program = Program(H(qubits[0]), CNOT(qubits)) program.measure_all(zip(qubits, measures)) return program

Resetting state

For this program, we'll say our desired result is that both qubits measure to 0. After an unsuccessful attempt where they measure to 1, we'll want to reset the state of the qubits before trying again. To do this, we'll create a program that applies an X gate to the qubits if either of them measured to 1:

sentinel-based-loop

def reset_bell_state(qubits: Tuple[Qubit, Qubit], measures: MemoryReference) -> Program:

"""Resets the state of the qubits if either of them measured to 1.""" program = Program() program.if_then(measures[0], Program(X(qubits[0]))) program.if_then(measures[1], Program(X(qubits[1]))) return program

Enforcing a sentinel condition

Next, we'll construct the part of our program that enforces the sentinel condition. In this case, we'll end the program if the given memory reference is 0, otherwise we'll want to reset the state of our qubits and jump back to the beginning of the program. We'll construct the branch that ends the program using Quil's Halt instruction, and we'll accept the alternative branch as an argument to our function and pass it in the next step:

sentinel-based-loop

def enforce_sentinel(mem_ref: MemoryReference, else_program: Program) -> Program:

"""Ends the program if mem_ref is 0, otherwise executes else_program.""" program = Program() # We use the if_then method here to help us construct our branch. As described above, # if_then takes a memory reference and two programs. It constructs a branch that # runs the first program if mem_ref is 1, otherwise it runs the second program. # Since we want to end the program if mem_ref is 0, we pass in our HALTing # program as the second program and the alternative branch as the first. program.if_then(mem_ref, else_program, Program(Halt())) return program

Putting it all together

With each component of our program ready, we just need to compose all the pieces:

sentinel-based-loop

def sentinel_program(qubits: Tuple[Qubit, Qubit]) -> Program:

# Create a label to reference the start of the program start_label = Label("start-loop")

# Use the label to create a jump target at the beginning of the program program = Program(JumpTarget(start_label))

# Declare a register to measure the qubits into measures = program.declare("measures", "BIT", 2)

# Add the loop body to our program program += body(qubits, measures)

# If the sentinel condition isn't met, then we: reset = Program( # Reset the state of our qubits reset_bell_state(qubits, measures), # Jump back to the start of the program Jump(start_label) )

# Finally, if both Qubits measured to 0 (our sentinel), then we want to end the program # Otherwise, we try again. program += enforce_sentinel(measures[0], reset)

# We used pyQuil to construct some of the branches for us, those methods use label placeholders # to avoid conflicts with existing labels in the program, so we resolve those placeholders here. program.resolve_label_placeholders() return program

Testing our program

Now that we have our program, let's test it out. We'll use the sentinel_program function to construct the program and run it against a QVM for 1000 shots. We'll use numpy to assert that the measures register contains only 0s. Over 1000 trials, this result would be improbable if our program didn't work as intended.

sentinel-based-loop

qubits = (Qubit(0), Qubit(1))

qc = get_qc("2q-qvm") program = sentinel_program(qubits) program.wrap_in_numshots_loop(1000) print(program.out()) results = qc.run(program) measures = results.get_register_map()["measures"]

assert np.all(measures == 0)

sentinel-based-loop

DECLARE measures BIT[2] LABEL @start-loop H 0 CNOT 0 1 MEASURE 0 measures[0] MEASURE 1 measures[1] JUMP-WHEN @THEN_0 measures[0] HALT JUMP @END_0 LABEL @THEN_0 JUMP-WHEN @THEN_1 measures[0] JUMP @END_1 LABEL @THEN_1 X 0 LABEL @END_1 JUMP-WHEN @THEN_2 measures[1] JUMP @END_2 LABEL @THEN_2 X 1 LABEL @END_2 JUMP @start-loop LABEL @END_0

Pauli Operator Algebra

Many algorithms require manipulating sums of Pauli combinations, such as $\sigma = \frac{1}{2}I - \frac{3}{4}X_0Y_1Z_3 + (5-2i)Z_1X_2,$ where Gn indicates the gate G acting on qubit n. We can represent such sums by constructing PauliTerm and PauliSum. The above sum can be constructed as follows:

pauli-algebra

from pyquil.paulis import ID, sX, sY, sZ

# Pauli term takes an operator "X", "Y", "Z", or "I"; a qubit to act on, and # an optional coefficient. a = 0.5 * ID() b = -0.75 * sX(0) * sY(1) * sZ(3) c = (5-2j) * sZ(1) * sX(2)

# Construct a sum of Pauli terms. sigma = a + b + c print(f"sigma = {sigma}")

pauli-algebra

sigma = (0.5+0j)I + (-0.75+0j)X0*Y1*Z3 + (5-2j)Z1X2

Right now, the primary thing one can do with Pauli terms and sums is to construct the exponential of the Pauli term, i.e., exp [ − iβσ]. This is accomplished by constructing a parameterized Quil program that is evaluated when passed values for the coefficients of the angle β.

Related to exponentiating Pauli sums, we provide utility functions for finding the commuting subgroups of a Pauli sum and approximating the exponential with the Suzuki-Trotter approximation through fourth order.

When arithmetic is done with Pauli sums, simplification is automatically done.

The following shows an instructive example of all three.

pauli-algebra

from pyquil.paulis import exponential_map

sigma_cubed = sigma * sigma * sigma print(f"Simplified: {sigma_cubed}n")

# Produce Quil code to compute exp[iX] H = -1.0 * sX(0) print(f"Quil to compute exp[iX] on qubit 0:n" f"{exponential_map(H)(1.0)}")

pauli-algebra

Simplified: (32.46875-30j)I + (-16.734375+15j)X0*Y1*Z3 + (71.5625-144.625j)Z1X2

Quil to compute exp[iX] on qubit 0: H 0 RZ(-2) 0 H 0

exponential_map returns a function allowing you to fill in a multiplicative constant later. This commonly occurs in variational algorithms. The function exponential_map is used to compute exp [ − iαH] without explicitly filling in a value for α.

pauli-algebra

expH = exponential_map(H) print(f"0:n{expH(0.0)}n") print(f"1:n{expH(1.0)}n") print(f"2:n{expH(2.0)}")

pauli-algebra

0: H 0 RZ(0) 0 H 0

1: H 0 RZ(-2) 0 H 0

2: H 0 RZ(-4) 0 H 0

To take it one step further, you can use parametric_compilation with exponential_map. For instance:

pauli-algebra

from pyquil import Program

pauli-algebra

ham = sZ(0) * sZ(1) prog = Program() theta = prog.declare('theta', 'REAL') prog += exponential_map(ham)(theta)