Google qsim vs Qulacs
The same circuits, one host
Google qsim and Qulacs both finished 138 of the 175 workloads measured here. Google qsim was faster in 64 of them and Qulacs in 74. At the median Qulacs was 3.0× faster.
By circuit family
Who was faster where
| Circuit family | Workloads both finished | Google qsim faster in | Qulacs faster in | Faster at the median |
|---|---|---|---|---|
| Workloads both finished 12 | Google qsim faster in 6 | Qulacs faster in 6 | Faster at the median Qulacs 1.3× | |
| Workloads both finished 20 | Google qsim faster in 10 | Qulacs faster in 10 | Faster at the median Qulacs 2.5× | |
| Workloads both finished 23 | Google qsim faster in 11 | Qulacs faster in 12 | Faster at the median Qulacs 3.2× | |
| Workloads both finished 12 | Google qsim faster in 6 | Qulacs faster in 6 | Faster at the median Qulacs 2.3× | |
| Workloads both finished 20 | Google qsim faster in 10 | Qulacs faster in 10 | Faster at the median Qulacs 1.9× | |
| Workloads both finished 12 | Google qsim faster in 4 | Qulacs faster in 8 | Faster at the median Qulacs 3.3× | |
| Workloads both finished 15 | Google qsim faster in 7 | Qulacs faster in 8 | Faster at the median Qulacs 3.2× | |
| Workloads both finished 12 | Google qsim faster in 6 | Qulacs faster in 6 | Faster at the median Qulacs 1.8× | |
| Workloads both finished 12 | Google qsim faster in 4 | Qulacs faster in 8 | Faster at the median Qulacs 2.8× |
Local circuit dynamics
Brickwork nearest-neighbour layers: lattice models, quench dynamics
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 1.19 ms | Qulacs 32.1 µs | Faster Qulacs 37.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 2.03 ms | Qulacs 73.1 µs | Faster Qulacs 27.8× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 2.93 ms | Qulacs 585 µs | Faster Qulacs 5.0× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 4.20 ms | Qulacs 10.7 ms | Faster Google qsim 2.6× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 18.9 ms | Qulacs 206 ms | Faster Google qsim 10.9× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 555 ms | Qulacs 4.76 s | Faster Google qsim 8.6× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 1.28 ms | Qulacs 42.5 µs | Faster Qulacs 30.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 2.23 ms | Qulacs 119 µs | Faster Qulacs 18.7× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 3.12 ms | Qulacs 1.38 ms | Faster Qulacs 2.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 5.16 ms | Qulacs 26.2 ms | Faster Google qsim 5.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 33.4 ms | Qulacs 535 ms | Faster Google qsim 16.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 934 ms | Qulacs 10.7 s | Faster Google qsim 11.4× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 117 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 307 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 3.00 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 59.6 ms | Faster – |
Hamiltonian time evolution
Trotterised spin chain: ZZ + XX couplings with local fields
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 4.11 ms | Qulacs 104 µs | Faster Qulacs 39.5× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 8.72 ms | Qulacs 276 µs | Faster Qulacs 31.6× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 13.3 ms | Qulacs 1.85 ms | Faster Qulacs 7.2× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 19.2 ms | Qulacs 34.0 ms | Faster Google qsim 1.8× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 51.8 ms | Qulacs 711 ms | Faster Google qsim 13.7× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 1.04 s | Qulacs 21.1 s | Faster Google qsim 20.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 4.18 ms | Qulacs 113 µs | Faster Qulacs 37.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 8.90 ms | Qulacs 321 µs | Faster Qulacs 27.8× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 13.6 ms | Qulacs 2.51 ms | Faster Qulacs 5.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 19.9 ms | Qulacs 48.2 ms | Faster Google qsim 2.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 69.5 ms | Qulacs 1.02 s | Faster Google qsim 14.6× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 1.46 s | Qulacs 26.9 s | Faster Google qsim 18.4× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Google qsim 9.13 ms | Qulacs 349 µs | Faster Qulacs 26.2× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Google qsim 14.1 ms | Qulacs 3.17 ms | Faster Qulacs 4.4× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Google qsim 23.0 ms | Qulacs 61.5 ms | Faster Google qsim 2.7× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Google qsim 158 ms | Qulacs 1.28 s | Faster Google qsim 8.1× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Google qsim 9.07 ms | Qulacs 305 µs | Faster Qulacs 29.7× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Google qsim 14.0 ms | Qulacs 2.50 ms | Faster Qulacs 5.6× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Google qsim 20.7 ms | Qulacs 48.1 ms | Faster Google qsim 2.3× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Google qsim 95.0 ms | Qulacs 1.01 s | Faster Google qsim 10.6× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 424 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 1.17 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 9.19 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 181 ms | Faster – |
MaxCut on a ring
QAOA over a cycle graph
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 2.67 ms | Qulacs 63.1 µs | Faster Qulacs 42.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 5.15 ms | Qulacs 154 µs | Faster Qulacs 33.5× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 7.52 ms | Qulacs 1.14 ms | Faster Qulacs 6.6× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 11.0 ms | Qulacs 20.9 ms | Faster Google qsim 1.9× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 41.3 ms | Qulacs 433 ms | Faster Google qsim 10.5× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 1.07 s | Qulacs 10.6 s | Faster Google qsim 9.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 2.74 ms | Qulacs 72.0 µs | Faster Qulacs 38.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 5.23 ms | Qulacs 197 µs | Faster Qulacs 26.6× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 7.74 ms | Qulacs 1.91 ms | Faster Qulacs 4.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 12.0 ms | Qulacs 36.9 ms | Faster Google qsim 3.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 55.1 ms | Qulacs 740 ms | Faster Google qsim 13.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 1.48 s | Qulacs 16.6 s | Faster Google qsim 11.2× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Google qsim 5.68 ms | Qulacs 234 µs | Faster Qulacs 24.3× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Google qsim 8.49 ms | Qulacs 2.61 ms | Faster Qulacs 3.2× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Google qsim 15.9 ms | Qulacs 52.1 ms | Faster Google qsim 3.3× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Google qsim 151 ms | Qulacs 1.03 s | Faster Google qsim 6.8× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Google qsim 5.52 ms | Qulacs 191 µs | Faster Qulacs 29.0× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Google qsim 8.38 ms | Qulacs 1.95 ms | Faster Qulacs 4.3× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Google qsim 13.2 ms | Qulacs 38.6 ms | Faster Google qsim 2.9× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Google qsim 87.2 ms | Qulacs 762 ms | Faster Google qsim 8.7× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 255 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 681 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 6.05 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 120 ms | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 8 | Google qsim 5.15 ms | Qulacs 155 µs | Faster Qulacs 33.2× |
Statevector, global observable (all_z), depth 3 | Qubits 12 | Google qsim 7.61 ms | Qulacs 1.13 ms | Faster Qulacs 6.7× |
Statevector, global observable (all_z), depth 3 | Qubits 16 | Google qsim 12.0 ms | Qulacs 21.0 ms | Faster Google qsim 1.8× |
MaxCut, arbitrary graph
QAOA over a random-edge graph
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 2.89 ms | Qulacs 71.4 µs | Faster Qulacs 40.5× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 5.89 ms | Qulacs 188 µs | Faster Qulacs 31.4× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 9.80 ms | Qulacs 1.51 ms | Faster Qulacs 6.5× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 12.7 ms | Qulacs 25.8 ms | Faster Google qsim 2.0× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 65.3 ms | Qulacs 588 ms | Faster Google qsim 9.0× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 2.04 s | Qulacs 16.1 s | Faster Google qsim 7.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 2.95 ms | Qulacs 81.0 µs | Faster Qulacs 36.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 6.17 ms | Qulacs 239 µs | Faster Qulacs 25.8× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 10.0 ms | Qulacs 2.39 ms | Faster Qulacs 4.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 13.8 ms | Qulacs 46.3 ms | Faster Google qsim 3.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 79.9 ms | Qulacs 1.00 s | Faster Google qsim 12.6× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 2.44 s | Qulacs 24.2 s | Faster Google qsim 9.9× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 287 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 873 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 8.07 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 148 ms | Faster – |
Sampling / IQP
Diagonal phase circuit between Hadamard layers
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 2.70 ms | Qulacs 68.8 µs | Faster Qulacs 39.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 5.02 ms | Qulacs 156 µs | Faster Qulacs 32.2× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 7.41 ms | Qulacs 1.04 ms | Faster Qulacs 7.1× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 10.8 ms | Qulacs 18.5 ms | Faster Google qsim 1.7× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 43.8 ms | Qulacs 392 ms | Faster Google qsim 9.0× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 1.11 s | Qulacs 10.7 s | Faster Google qsim 9.7× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 2.73 ms | Qulacs 78.2 µs | Faster Qulacs 34.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 5.21 ms | Qulacs 203 µs | Faster Qulacs 25.7× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 7.69 ms | Qulacs 1.97 ms | Faster Qulacs 3.9× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 11.8 ms | Qulacs 38.7 ms | Faster Google qsim 3.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 55.6 ms | Qulacs 797 ms | Faster Google qsim 14.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 1.50 s | Qulacs 19.1 s | Faster Google qsim 12.7× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 8 | Google qsim 5.64 ms | Qulacs 239 µs | Faster Qulacs 23.5× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 12 | Google qsim 8.40 ms | Qulacs 2.67 ms | Faster Qulacs 3.2× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 16 | Google qsim 15.6 ms | Qulacs 53.4 ms | Faster Google qsim 3.4× |
Expectation value, global observable (pauli_mix), depth 3 | Qubits 20 | Google qsim 147 ms | Qulacs 1.09 s | Faster Google qsim 7.4× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 8 | Google qsim 5.43 ms | Qulacs 196 µs | Faster Qulacs 27.7× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 12 | Google qsim 8.08 ms | Qulacs 2.04 ms | Faster Qulacs 4.0× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 16 | Google qsim 13.3 ms | Qulacs 39.9 ms | Faster Google qsim 3.0× |
Expectation value, global observable (zz_sum), depth 3 | Qubits 20 | Google qsim 88.2 ms | Qulacs 856 ms | Faster Google qsim 9.7× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 272 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 701 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 5.38 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 105 ms | Faster – |
Quantum kernel feature map
Data-encoding circuit for kernel methods
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, global observable (all_z), depth 2 | Qubits 8 | Google qsim 2.02 ms | Qulacs 107 µs | Faster Qulacs 19.0× |
Expectation value, global observable (all_z), depth 2 | Qubits 12 | Google qsim 2.92 ms | Qulacs 895 µs | Faster Qulacs 3.3× |
Expectation value, global observable (all_z), depth 2 | Qubits 16 | Google qsim 5.13 ms | Qulacs 16.9 ms | Faster Google qsim 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 8 | Google qsim 64.6 ms | Qulacs 3.36 ms | Faster Qulacs 19.2× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 12 | Google qsim 93.4 ms | Qulacs 28.3 ms | Faster Qulacs 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 16 | Google qsim 165 ms | Qulacs 539 ms | Faster Google qsim 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 8 | Google qsim 521 ms | Qulacs 26.9 ms | Faster Qulacs 19.4× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 12 | Google qsim 739 ms | Qulacs 226 ms | Faster Qulacs 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 16 | Google qsim 1.32 s | Qulacs 4.31 s | Faster Google qsim 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 8 | Google qsim 2.10 s | Qulacs 109 ms | Faster Qulacs 19.2× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 12 | Google qsim 2.98 s | Qulacs 915 ms | Faster Qulacs 3.3× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 16 | Google qsim 5.29 s | Qulacs 17.3 s | Faster Google qsim 3.3× |
Hardware-efficient VQE
RY/RZ layers with a CNOT ring
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 2.21 ms | Qulacs 52.5 µs | Faster Qulacs 42.2× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 4.06 ms | Qulacs 126 µs | Faster Qulacs 32.1× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 5.87 ms | Qulacs 1.03 ms | Faster Qulacs 5.7× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 8.94 ms | Qulacs 19.2 ms | Faster Google qsim 2.2× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 40.4 ms | Qulacs 386 ms | Faster Google qsim 9.6× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 1.09 s | Qulacs 8.90 s | Faster Google qsim 8.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 2.28 ms | Qulacs 63.7 µs | Faster Qulacs 35.8× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 4.17 ms | Qulacs 172 µs | Faster Qulacs 24.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 6.29 ms | Qulacs 1.97 ms | Faster Qulacs 3.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 10.0 ms | Qulacs 40.2 ms | Faster Google qsim 4.0× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 52.9 ms | Qulacs 808 ms | Faster Google qsim 15.3× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 1.51 s | Qulacs 17.1 s | Faster Google qsim 11.3× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 206 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 577 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 5.51 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 110 ms | Faster – |
Statevector, global observable (all_z), depth 3 | Qubits 8 | Google qsim 4.13 ms | Qulacs 129 µs | Faster Qulacs 32.0× |
Statevector, global observable (all_z), depth 3 | Qubits 12 | Google qsim 5.93 ms | Qulacs 1.00 ms | Faster Qulacs 5.9× |
Statevector, global observable (all_z), depth 3 | Qubits 16 | Google qsim 9.69 ms | Qulacs 18.9 ms | Faster Google qsim 2.0× |
Strongly-entangling ansatz
RX/RY/RZ layers with a CNOT ring
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, local observable, depth 3 | Qubits 4 | Google qsim 2.87 ms | Qulacs 68.2 µs | Faster Qulacs 42.1× |
Expectation value, local observable, depth 3 | Qubits 8 | Google qsim 5.35 ms | Qulacs 177 µs | Faster Qulacs 30.3× |
Expectation value, local observable, depth 3 | Qubits 12 | Google qsim 7.79 ms | Qulacs 1.56 ms | Faster Qulacs 5.0× |
Expectation value, local observable, depth 3 | Qubits 16 | Google qsim 11.4 ms | Qulacs 30.0 ms | Faster Google qsim 2.6× |
Expectation value, local observable, depth 3 | Qubits 20 | Google qsim 41.6 ms | Qulacs 595 ms | Faster Google qsim 14.3× |
Expectation value, local observable, depth 3 | Qubits 24 | Google qsim 1.05 s | Qulacs 13.2 s | Faster Google qsim 12.6× |
Expectation value, global observable (all_z), depth 3 | Qubits 4 | Google qsim 2.98 ms | Qulacs 77.7 µs | Faster Qulacs 38.4× |
Expectation value, global observable (all_z), depth 3 | Qubits 8 | Google qsim 5.57 ms | Qulacs 222 µs | Faster Qulacs 25.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 12 | Google qsim 8.01 ms | Qulacs 2.50 ms | Faster Qulacs 3.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 16 | Google qsim 12.3 ms | Qulacs 50.6 ms | Faster Google qsim 4.1× |
Expectation value, global observable (all_z), depth 3 | Qubits 20 | Google qsim 59.3 ms | Qulacs 1.02 s | Faster Google qsim 17.2× |
Expectation value, global observable (all_z), depth 3 | Qubits 24 | Google qsim 1.44 s | Qulacs 21.4 s | Faster Google qsim 14.8× |
Gradient, local observable, depth 3 | Qubits 4 | Google qsim unsupported | Qulacs 278 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 8 | Google qsim unsupported | Qulacs 822 µs | Faster – |
Gradient, local observable, depth 3 | Qubits 12 | Google qsim unsupported | Qulacs 8.63 ms | Faster – |
Gradient, local observable, depth 3 | Qubits 16 | Google qsim unsupported | Qulacs 180 ms | Faster – |
Data re-uploading QML
Repeated encode-and-rotate classifier
| Workload | Qubits | Google qsim | Qulacs | Faster |
|---|---|---|---|---|
Expectation value, global observable (all_z), depth 2 | Qubits 8 | Google qsim 3.99 ms | Qulacs 185 µs | Faster Qulacs 21.6× |
Expectation value, global observable (all_z), depth 2 | Qubits 12 | Google qsim 5.75 ms | Qulacs 2.06 ms | Faster Qulacs 2.8× |
Expectation value, global observable (all_z), depth 2 | Qubits 16 | Google qsim 9.07 ms | Qulacs 41.8 ms | Faster Google qsim 4.6× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 8 | Google qsim 128 ms | Qulacs 6.97 ms | Faster Qulacs 18.3× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 12 | Google qsim 186 ms | Qulacs 67.1 ms | Faster Qulacs 2.8× |
Expectation value, global observable (all_z), depth 2, batch of 32 | Qubits 16 | Google qsim 289 ms | Qulacs 1.33 s | Faster Google qsim 4.6× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 8 | Google qsim 1.02 s | Qulacs 48.4 ms | Faster Qulacs 21.1× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 12 | Google qsim 1.47 s | Qulacs 529 ms | Faster Qulacs 2.8× |
Expectation value, global observable (all_z), depth 2, batch of 256 | Qubits 16 | Google qsim 2.35 s | Qulacs 10.6 s | Faster Google qsim 4.5× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 8 | Google qsim 4.09 s | Qulacs 193 ms | Faster Qulacs 21.2× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 12 | Google qsim 5.89 s | Qulacs 2.13 s | Faster Qulacs 2.8× |
Expectation value, global observable (all_z), depth 2, batch of 1024 | Qubits 16 | Google qsim 9.24 s | Qulacs 42.7 s | Faster Google qsim 4.6× |
Gradient, local observable, depth 2 | Qubits 8 | Google qsim unsupported | Qulacs 606 µs | Faster – |
Gradient, local observable, depth 2 | Qubits 12 | Google qsim unsupported | Qulacs 5.81 ms | Faster – |
Gradient, local observable, depth 2 | Qubits 16 | Google qsim unsupported | Qulacs 117 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 8 | Google qsim unsupported | Qulacs 19.3 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 12 | Google qsim unsupported | Qulacs 189 ms | Faster – |
Gradient, local observable, depth 2, batch of 32 | Qubits 16 | Google qsim unsupported | Qulacs 3.76 s | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 8 | Google qsim unsupported | Qulacs 153 ms | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 12 | Google qsim unsupported | Qulacs 1.47 s | Faster – |
Gradient, local observable, depth 2, batch of 256 | Qubits 16 | Google qsim unsupported | Qulacs 30.0 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.