CodeUniverse · package
Cantera
many-reaction kinetic sensitivity from one pullback
AD speed-up on the same differentiation workload (against finite differences) Speed-up: 31.7×
High-dimensional demo
many-reaction kinetic sensitivity from one pullback
Supplied demo scenario wording.
FD components requested 1 – 256 · AD components 325
View demo measurements
| Measure | Value |
|---|---|
| AD median (s) | 0.00047691690269857645 |
| FD median (s) | 0.015114957932382822 |
| FD / AD | 31.693064026157735 |
| Max scaled error | 4.020002241655009e-17 |
| AD components | 325 |
| FD components requested | 256 |
| AD parameters | 256 |
| Repeats | 7 |
| FD components requested | AD median (s) | FD median (s) | FD / AD | Max scaled error | AD components |
|---|---|---|---|---|---|
| 1 | 325 | 0.0005128330085426569 | 0.00005495804361999035 | 0.10716557379207582 | 1.8728041291982752e-17 |
| 2 | 325 | 0.0004813749110326171 | 0.00012208300177007914 | 0.2536131380594678 | 1.8728041291982752e-17 |
| 4 | 325 | 0.000490125035867095 | 0.0002401659730821848 | 0.49000960062629717 | 1.8728041291982752e-17 |
| 8 | 325 | 0.0004909581039100885 | 0.00047791609540581703 | 0.9734355978638455 | 1.8728041291982752e-17 |
| 16 | 325 | 0.0004854579456150532 | 0.000949749955907464 | 1.9564000640759396 | 1.8728041291982752e-17 |
| 32 | 325 | 0.0004967919085174799 | 0.0018745000706985593 | 3.773209745489653 | 1.8728041291982752e-17 |
| 64 | 325 | 0.0004853330319747329 | 0.003818625002168119 | 7.868050906468946 | 1.8728041291982752e-17 |
| 128 | 325 | 0.00047433306463062763 | 0.007503249915316701 | 15.818525999572953 | 3.206104408652252e-17 |
| 256 | 325 | 0.00047691690269857645 | 0.015114957932382822 | 31.693064026157735 | 4.020002241655009e-17 |
One-direction performance audit
View audit measurements
| Demo scenario | Problem size | Repeats | AD median (s) | FD median (s) | FD / AD | Relative error | Supplied verdict |
|---|---|---|---|---|---|---|---|
| representative_reactor_direction | — | — | 0.000069625 | 0.000111875 | 1.606822262 | — | — |
FD/AD > 1 means automatic differentiation is faster; FD/AD < 1 means the finite-difference comparison is faster. One direction only, measured inside the package's own case.