Turn caching off by default when max_diff==1

[albi3ro]

Context:

Recent benchmarks (see #5211 (comment)) have shown that caching adds massive classical overheads, but often does not actually lead to a reduction in the the number of executions in normal workflows.

Because of this, we want to make smart choices about when to use caching. Higher-order derivatives often result in duplicate circuits, so we need to keep caching when calculating higher-order derivatives. But, we can make caching opt-in for normal workflows. This will lead to reduced overheads in the vast majority of workflows.

Description of the Change:

The QNode keyword argument cache defaults to None. This is interpreted as True if max_diff > 1 and False otherwise.

Benefits:

Vastly reduced classical overheads in most cases.

Possible Drawbacks:

Increased number of executions in a few edge cases. But these edge cases would be fairly convoluted. Somehow a transform would have to turn the starting tape into two identical tapes.

Related GitHub Issues:

Performance Numbers:

n_layers = 1

dev = qml.device('lightning.qubit', wires=n_wires)

shape = qml.StronglyEntanglingLayers.shape(n_wires=n_wires, n_layers=n_layers)
rng = qml.numpy.random.default_rng(seed=42)
params = rng.random(shape)

@qml.qnode(dev, cache=False)
def circuit(params):
    qml.StronglyEntanglingLayers(params, wires=range(n_wires))
    return qml.expval(qml.Z(0))

@qml.qnode(dev, cache=True)
def circuit_cache(params):
    qml.StronglyEntanglingLayers(params, wires=range(n_wires))
    return qml.expval(qml.Z(0))

For n_wires = 20

But for n_wires= 10:

For n_wires=20 n_layers=5, we have:

While the cache version does seem to be faster here, that does seem to be statistical fluctuations.

For n_wires=10 n_layers=20:

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.67%. Comparing base (eb03750) to head (ebcb269).
Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5243      +/-   ##
==========================================
- Coverage   99.68%   99.67%   -0.01%     
==========================================
  Files         399      399              
  Lines       36853    36575     -278     
==========================================
- Hits        36736    36457     -279     
- Misses        117      118       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[josh146]

Thanks @albi3ro!

Out of curiosity, I wonder if the caching affected any first-order gradient workflows that used qml.jacobian()? That is, any workflows where multiple identical tapes would be executed to build up the Jacobian?

[albi3ro]

Thanks @albi3ro!

Out of curiosity, I wonder if the caching affected any first-order gradient workflows that used qml.jacobian()? That is, any workflows where multiple identical tapes would be executed to build up the Jacobian?

We add an additional mandatory level of caching for autograd, so cache_execute was no influencing it.

[timmysilv]

i'm not sure what the current state of the benchmarking suite is, but perhaps this PR would be a good opportunity to test it out?

[albi3ro]

[sc-57404]

[AmintorDusko]

Good idea! Thank you for that!

[albi3ro]

i'm not sure what the current state of the benchmarking suite is, but perhaps this PR would be a good opportunity to test it out?

Current benchmarks are using a github runner and too noisy to extract reliable data.