Qiskit Aer (MPS) vs TensorCircuit
The same circuits, one host
Qiskit Aer (MPS) and TensorCircuit both finished 104 of the 175 workloads measured here. Qiskit Aer (MPS) was faster in 26 of them and TensorCircuit in 78. At the median TensorCircuit was 5.8× faster.
By circuit family
Who was faster where
| Circuit family | Workloads both finished | Qiskit Aer (MPS) faster in | TensorCircuit faster in | Faster at the median |
|---|---|---|---|---|
| Workloads both finished 12 | Qiskit Aer (MPS) faster in 5 | TensorCircuit faster in 7 | Faster at the median TensorCircuit 2.0× | |
| Workloads both finished 14 | Qiskit Aer (MPS) faster in 3 | TensorCircuit faster in 11 | Faster at the median TensorCircuit 8.8× | |
| Workloads both finished 18 | Qiskit Aer (MPS) faster in 4 | TensorCircuit faster in 14 | Faster at the median TensorCircuit 9.3× | |
| Workloads both finished 0 | Qiskit Aer (MPS) faster in 0 | TensorCircuit faster in 0 | Faster at the median – | |
| Workloads both finished 14 | Qiskit Aer (MPS) faster in 0 | TensorCircuit faster in 14 | Faster at the median TensorCircuit 10.3× | |
| Workloads both finished 11 | Qiskit Aer (MPS) faster in 3 | TensorCircuit faster in 8 | Faster at the median TensorCircuit 1.6× | |
| Workloads both finished 12 | Qiskit Aer (MPS) faster in 4 | TensorCircuit faster in 8 | Faster at the median TensorCircuit 5.3× | |
| Workloads both finished 12 | Qiskit Aer (MPS) faster in 4 | TensorCircuit faster in 8 | Faster at the median TensorCircuit 5.7× | |
| Workloads both finished 11 | Qiskit Aer (MPS) faster in 3 | TensorCircuit faster in 8 | Faster at the median TensorCircuit 3.6× |
Local circuit dynamics
Brickwork nearest-neighbour layers: lattice models, quench dynamics
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 942 µs | TensorCircuit 159 µs | Faster TensorCircuit 5.9× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 1.43 ms | TensorCircuit 249 µs | Faster TensorCircuit 5.8× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 1.83 ms | TensorCircuit 378 µs | Faster TensorCircuit 4.8× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 2.34 ms | TensorCircuit 2.17 ms | Faster TensorCircuit 1.1× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) 2.79 ms | TensorCircuit 63.5 ms | Faster Qiskit Aer (MPS) 22.8× |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) 3.14 ms | TensorCircuit 1.16 s | Faster Qiskit Aer (MPS) 369× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.08 ms | TensorCircuit 181 µs | Faster TensorCircuit 6.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 1.80 ms | TensorCircuit 285 µs | Faster TensorCircuit 6.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 2.48 ms | TensorCircuit 848 µs | Faster TensorCircuit 2.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 3.39 ms | TensorCircuit 33.5 ms | Faster Qiskit Aer (MPS) 9.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) 4.32 ms | TensorCircuit 735 ms | Faster Qiskit Aer (MPS) 170× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) 5.35 ms | TensorCircuit 16.2 s | Faster Qiskit Aer (MPS) 3037× |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 253 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 426 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 758 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 5.49 ms | Faster – |
Hamiltonian time evolution
Trotterised spin chain: ZZ + XX couplings with local fields
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 2.81 ms | TensorCircuit 267 µs | Faster TensorCircuit 10.5× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.89 ms | TensorCircuit 468 µs | Faster TensorCircuit 12.6× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 9.17 ms | TensorCircuit 749 µs | Faster TensorCircuit 12.2× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 12.3 ms | TensorCircuit 4.46 ms | Faster TensorCircuit 2.8× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 164 ms | Faster – |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit 6.20 s | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 2.95 ms | TensorCircuit 275 µs | Faster TensorCircuit 10.7× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 6.44 ms | TensorCircuit 522 µs | Faster TensorCircuit 12.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 10.0 ms | TensorCircuit 1.42 ms | Faster TensorCircuit 7.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 14.3 ms | TensorCircuit 45.6 ms | Faster Qiskit Aer (MPS) 3.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 1.43 s | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit 45.3 s | Faster – |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Qiskit Aer (MPS) 6.73 ms | TensorCircuit 593 µs | Faster TensorCircuit 11.4× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Qiskit Aer (MPS) 11.1 ms | TensorCircuit 2.12 ms | Faster TensorCircuit 5.3× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Qiskit Aer (MPS) 16.1 ms | TensorCircuit 78.8 ms | Faster Qiskit Aer (MPS) 4.9× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 2.65 s | Faster – |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Qiskit Aer (MPS) 6.26 ms | TensorCircuit 537 µs | Faster TensorCircuit 11.7× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Qiskit Aer (MPS) 10.0 ms | TensorCircuit 1.74 ms | Faster TensorCircuit 5.8× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Qiskit Aer (MPS) 13.8 ms | TensorCircuit 60.3 ms | Faster Qiskit Aer (MPS) 4.4× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 1.90 s | Faster – |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 442 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 962 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 1.61 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 10.7 ms | Faster – |
MaxCut on a ring
QAOA over a cycle graph
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.84 ms | TensorCircuit 199 µs | Faster TensorCircuit 9.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.28 ms | TensorCircuit 387 µs | Faster TensorCircuit 11.1× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 19.9 ms | TensorCircuit 1.43 ms | Faster TensorCircuit 13.9× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 54.6 ms | TensorCircuit 21.2 ms | Faster TensorCircuit 2.6× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) 88.7 ms | TensorCircuit 596 ms | Faster Qiskit Aer (MPS) 6.7× |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) 123 ms | TensorCircuit 12.8 s | Faster Qiskit Aer (MPS) 104× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.95 ms | TensorCircuit 207 µs | Faster TensorCircuit 9.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.93 ms | TensorCircuit 428 µs | Faster TensorCircuit 11.5× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 32.2 ms | TensorCircuit 2.01 ms | Faster TensorCircuit 16.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 131 ms | TensorCircuit 58.3 ms | Faster TensorCircuit 2.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) 262 ms | TensorCircuit 1.66 s | Faster Qiskit Aer (MPS) 6.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) 424 ms | TensorCircuit 42.9 s | Faster Qiskit Aer (MPS) 101× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.72 ms | TensorCircuit 541 µs | Faster TensorCircuit 10.6× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Qiskit Aer (MPS) 56.4 ms | TensorCircuit 2.90 ms | Faster TensorCircuit 19.4× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Qiskit Aer (MPS) 274 ms | TensorCircuit 96.2 ms | Faster TensorCircuit 2.9× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 2.50 s | Faster – |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.03 ms | TensorCircuit 502 µs | Faster TensorCircuit 10.0× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Qiskit Aer (MPS) 35.0 ms | TensorCircuit 2.53 ms | Faster TensorCircuit 13.8× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Qiskit Aer (MPS) 141 ms | TensorCircuit 77.6 ms | Faster TensorCircuit 1.8× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 2.25 s | Faster – |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 372 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 858 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 4.31 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 104 ms | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.80 ms | TensorCircuit unsupported | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 23.8 ms | TensorCircuit unsupported | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 123 ms | TensorCircuit unsupported | Faster – |
MaxCut, arbitrary graph
QAOA over a random-edge graph
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 2.13 ms | TensorCircuit error | Faster – |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.01 ms | TensorCircuit error | Faster – |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 192 ms | TensorCircuit error | Faster – |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 4.03 s | TensorCircuit error | Faster – |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit error | Faster – |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 2.22 ms | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.54 ms | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 208 ms | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 5.23 s | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) timeout | TensorCircuit error | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit error | Faster – |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit error | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit error | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit error | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit error | Faster – |
Sampling / IQP
Diagonal phase circuit between Hadamard layers
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.98 ms | TensorCircuit 209 µs | Faster TensorCircuit 9.5× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.58 ms | TensorCircuit 425 µs | Faster TensorCircuit 10.8× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 21.5 ms | TensorCircuit 1.49 ms | Faster TensorCircuit 14.4× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 61.9 ms | TensorCircuit 21.1 ms | Faster TensorCircuit 2.9× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 633 ms | Faster – |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit 13.7 s | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 2.11 ms | TensorCircuit 229 µs | Faster TensorCircuit 9.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.12 ms | TensorCircuit 464 µs | Faster TensorCircuit 11.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 36.1 ms | TensorCircuit 2.04 ms | Faster TensorCircuit 17.7× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 167 ms | TensorCircuit 59.5 ms | Faster TensorCircuit 2.8× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 1.53 s | Faster – |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) skip | TensorCircuit 45.0 s | Faster – |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.95 ms | TensorCircuit 568 µs | Faster TensorCircuit 10.5× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Qiskit Aer (MPS) 65.2 ms | TensorCircuit 2.94 ms | Faster TensorCircuit 22.2× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Qiskit Aer (MPS) 359 ms | TensorCircuit 96.1 ms | Faster TensorCircuit 3.7× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 2.76 s | Faster – |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Qiskit Aer (MPS) 5.31 ms | TensorCircuit 525 µs | Faster TensorCircuit 10.1× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Qiskit Aer (MPS) 39.0 ms | TensorCircuit 2.56 ms | Faster TensorCircuit 15.2× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Qiskit Aer (MPS) 179 ms | TensorCircuit 79.0 ms | Faster TensorCircuit 2.3× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Qiskit Aer (MPS) skip | TensorCircuit 2.22 s | Faster – |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 425 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 898 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 4.39 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 104 ms | Faster – |
Quantum kernel feature map
Data-encoding circuit for kernel methods
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, global observable (all_z), depth 2 | Qubits 8 | Qiskit Aer (MPS) 1.81 ms | TensorCircuit 288 µs | Faster TensorCircuit 6.3× |
Expectation value, global observable (all_z), depth 2 | Qubits 12 | Qiskit Aer (MPS) 2.58 ms | TensorCircuit 871 µs | Faster TensorCircuit 3.0× |
Expectation value, global observable (all_z), depth 2 | Qubits 16 | Qiskit Aer (MPS) 3.42 ms | TensorCircuit 30.7 ms | Faster Qiskit Aer (MPS) 9.0× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 8 | Qiskit Aer (MPS) 57.9 ms | TensorCircuit 1.28 ms | Faster TensorCircuit 45.3× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 12 | Qiskit Aer (MPS) 82.5 ms | TensorCircuit 52.5 ms | Faster TensorCircuit 1.6× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 16 | Qiskit Aer (MPS) 107 ms | TensorCircuit 1.09 s | Faster Qiskit Aer (MPS) 10.1× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 8 | Qiskit Aer (MPS) 463 ms | TensorCircuit 8.45 ms | Faster TensorCircuit 54.8× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 12 | Qiskit Aer (MPS) 657 ms | TensorCircuit 428 ms | Faster TensorCircuit 1.5× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 16 | Qiskit Aer (MPS) 855 ms | TensorCircuit 8.81 s | Faster Qiskit Aer (MPS) 10.3× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 8 | Qiskit Aer (MPS) 1.84 s | TensorCircuit 79.4 ms | Faster TensorCircuit 23.2× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 12 | Qiskit Aer (MPS) 2.64 s | TensorCircuit 1.79 s | Faster TensorCircuit 1.5× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 16 | Qiskit Aer (MPS) 3.44 s | TensorCircuit error | Faster – |
Hardware-efficient VQE
RY/RZ layers with a CNOT ring
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.58 ms | TensorCircuit 211 µs | Faster TensorCircuit 7.5× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 3.55 ms | TensorCircuit 386 µs | Faster TensorCircuit 9.2× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 18.4 ms | TensorCircuit 1.50 ms | Faster TensorCircuit 12.3× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 57.4 ms | TensorCircuit 20.7 ms | Faster TensorCircuit 2.8× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) 96.0 ms | TensorCircuit 488 ms | Faster Qiskit Aer (MPS) 5.1× |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) 134 ms | TensorCircuit 13.7 s | Faster Qiskit Aer (MPS) 102× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.69 ms | TensorCircuit 212 µs | Faster TensorCircuit 8.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.12 ms | TensorCircuit 454 µs | Faster TensorCircuit 9.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 34.1 ms | TensorCircuit 2.02 ms | Faster TensorCircuit 16.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 169 ms | TensorCircuit 53.9 ms | Faster TensorCircuit 3.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) 350 ms | TensorCircuit 1.09 s | Faster Qiskit Aer (MPS) 3.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) 578 ms | TensorCircuit 27.5 s | Faster Qiskit Aer (MPS) 47.7× |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 340 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 830 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 4.05 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 66.6 ms | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 3.85 ms | TensorCircuit unsupported | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 20.7 ms | TensorCircuit unsupported | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 125 ms | TensorCircuit unsupported | Faster – |
Strongly-entangling ansatz
RX/RY/RZ layers with a CNOT ring
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.83 ms | TensorCircuit 224 µs | Faster TensorCircuit 8.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.02 ms | TensorCircuit 418 µs | Faster TensorCircuit 9.6× |
Expectation value, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) 18.2 ms | TensorCircuit 1.46 ms | Faster TensorCircuit 12.5× |
Expectation value, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) 57.7 ms | TensorCircuit 20.7 ms | Faster TensorCircuit 2.8× |
Expectation value, local observable, depth 3 | Qubits 20 | Qiskit Aer (MPS) 92.3 ms | TensorCircuit 467 ms | Faster Qiskit Aer (MPS) 5.1× |
Expectation value, local observable, depth 3 | Qubits 24 | Qiskit Aer (MPS) 131 ms | TensorCircuit 13.0 s | Faster Qiskit Aer (MPS) 99.5× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Qiskit Aer (MPS) 1.93 ms | TensorCircuit 229 µs | Faster TensorCircuit 8.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Qiskit Aer (MPS) 4.52 ms | TensorCircuit 458 µs | Faster TensorCircuit 9.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Qiskit Aer (MPS) 32.6 ms | TensorCircuit 2.01 ms | Faster TensorCircuit 16.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Qiskit Aer (MPS) 164 ms | TensorCircuit 50.3 ms | Faster TensorCircuit 3.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Qiskit Aer (MPS) 347 ms | TensorCircuit 1.13 s | Faster Qiskit Aer (MPS) 3.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Qiskit Aer (MPS) 560 ms | TensorCircuit 26.2 s | Faster Qiskit Aer (MPS) 46.8× |
Gradient, local observable, depth 3 | Qubits 4 | Qiskit Aer (MPS) unsupported | TensorCircuit 395 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 986 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 4.28 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 69.9 ms | Faster – |
Data re-uploading QML
Repeated encode-and-rotate classifier
| Workload | Qubits | Qiskit Aer (MPS) | TensorCircuit | Faster |
|---|---|---|---|---|
Expectation value, global observable (all_z), depth 2 | Qubits 8 | Qiskit Aer (MPS) 3.31 ms | TensorCircuit 388 µs | Faster TensorCircuit 8.5× |
Expectation value, global observable (all_z), depth 2 | Qubits 12 | Qiskit Aer (MPS) 6.67 ms | TensorCircuit 1.20 ms | Faster TensorCircuit 5.5× |
Expectation value, global observable (all_z), depth 2 | Qubits 16 | Qiskit Aer (MPS) 11.3 ms | TensorCircuit 43.8 ms | Faster Qiskit Aer (MPS) 3.9× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 8 | Qiskit Aer (MPS) 106 ms | TensorCircuit 2.00 ms | Faster TensorCircuit 52.9× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 12 | Qiskit Aer (MPS) 216 ms | TensorCircuit 59.8 ms | Faster TensorCircuit 3.6× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 16 | Qiskit Aer (MPS) 359 ms | TensorCircuit 2.02 s | Faster Qiskit Aer (MPS) 5.6× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 8 | Qiskit Aer (MPS) 839 ms | TensorCircuit 14.7 ms | Faster TensorCircuit 57.1× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 12 | Qiskit Aer (MPS) 1.72 s | TensorCircuit 497 ms | Faster TensorCircuit 3.5× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 16 | Qiskit Aer (MPS) 2.87 s | TensorCircuit 16.9 s | Faster Qiskit Aer (MPS) 5.9× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 8 | Qiskit Aer (MPS) 3.37 s | TensorCircuit 107 ms | Faster TensorCircuit 31.4× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 12 | Qiskit Aer (MPS) 6.96 s | TensorCircuit 2.17 s | Faster TensorCircuit 3.2× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 16 | Qiskit Aer (MPS) 11.4 s | TensorCircuit error | Faster – |
Gradient, local observable, depth 2 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 596 µs | Faster – |
Gradient, local observable, depth 2 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 1.16 ms | Faster – |
Gradient, local observable, depth 2 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 7.10 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 3.46 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 19.9 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 424 ms | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 8 | Qiskit Aer (MPS) unsupported | TensorCircuit 26.0 ms | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 12 | Qiskit Aer (MPS) unsupported | TensorCircuit 243 ms | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 16 | Qiskit Aer (MPS) unsupported | TensorCircuit 3.51 s | Faster – |
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.