PennyLane benchmark
Timed against 7 other simulators
PennyLane 0.45.1 finished 147 of the 175 workloads it was run on and was the fastest of the 8 simulators in 0. At the median it took 32.3× the time of the fastest simulator in the same workload. What it is and how to install it is in its own documentation.
By circuit family
Where it is fast and where it is not
| Circuit family | Workloads finished | Fastest in | Median time against the fastest |
|---|---|---|---|
| Workloads finished 14 of 16 | Fastest in 0 | Median time against the fastest 54.1× | |
| Workloads finished 22 of 24 | Fastest in 0 | Median time against the fastest 69.8× | |
| Workloads finished 25 of 27 | Fastest in 0 | Median time against the fastest 21.6× | |
| Workloads finished 14 of 16 | Fastest in 0 | Median time against the fastest 40.6× | |
| Workloads finished 22 of 24 | Fastest in 0 | Median time against the fastest 25.3× | |
| Workloads finished 8 of 12 | Fastest in 0 | Median time against the fastest 6.0× | |
| Workloads finished 17 of 19 | Fastest in 0 | Median time against the fastest 29.1× | |
| Workloads finished 14 of 16 | Fastest in 0 | Median time against the fastest 40.2× | |
| Workloads finished 11 of 21 | Fastest in 0 | Median time against the fastest 7.4× |
PennyLane against each other simulator
On the workloads both finished
| Comparison | Workloads both finished | PennyLane faster in | The other faster in | Faster at the median |
|---|---|---|---|---|
| Workloads both finished 147 | PennyLane faster in 6 | The other faster in 141 | Faster at the median PennyLane Lightning 3.4× | |
| Workloads both finished 119 | PennyLane faster in 50 | The other faster in 69 | Faster at the median Qiskit 1.7× | |
| Workloads both finished 119 | PennyLane faster in 10 | The other faster in 109 | Faster at the median Qiskit Aer 3.0× | |
| Workloads both finished 107 | PennyLane faster in 38 | The other faster in 69 | Faster at the median Qiskit Aer (MPS) 2.3× | |
| Workloads both finished 119 | PennyLane faster in 10 | The other faster in 109 | Faster at the median Google qsim 2.7× | |
| Workloads both finished 147 | PennyLane faster in 0 | The other faster in 147 | Faster at the median Qulacs 10.6× | |
| Workloads both finished 127 | PennyLane faster in 4 | The other faster in 123 | Faster at the median TensorCircuit 12.4× |
Where it produced no timing
The status the benchmark recorded
A simulator that did not finish a workload keeps the status the benchmark recorded for it, and is left out of that workload's ranking.
| Recorded status | Operation | Workloads |
|---|---|---|
skip | Operation Expectation value | Workloads 19 |
unsupported | Operation Gradient | Workloads 8 |
skip | Operation Gradient | Workloads 1 |
How this was measured
One host, one harness
Every simulator ran the same circuits on one host, single-threaded and core-pinned, driven through one common harness, so any overhead the harness adds is charged to the simulator. Each result was checked against a reference implementation: a simulator that ran fast but returned the wrong answer is left out of the ranking. Nothing is interpolated or rescaled. A time was measured, or it is not shown.
Run 2026-08-10, 8 cores, threads=1. numpy 2.4.6, scipy 1.18.0, torch 2.13.0, numba 0.66.0, pennylane 0.45.1, pennylane_lightning 0.45.0, qiskit 2.5.1, qulacs 0.6.14, tensorcircuit 1.8.0, cudaq 0.15.1. One host and one thread make the comparison reproducible. It is not a prediction for your laptop.
The interactive tool shows the same measurements for any workload you pick, with the quantum hardware that could run it and what a run would cost.