Open adamkewley opened 1 month ago
Speedup appears to be roughly 5 % on Rajagopal
, which uses MillardEquilibriumMuscle
:
Test Name | lhs [secs] | σ [secs] | rhs [secs] | σ [secs] | Speedup |
---|---|---|---|---|---|
ToyDropLanding_nomuscles | 0.04 | 0.00 | 0.04 | 0.00 | 1.11 |
RajagopalModel | 3.82 | 0.00 | 3.59 | 0.00 | 1.06 |
Gait2354 | 0.20 | 0.00 | 0.19 | 0.00 | 1.03 |
passive_dynamic_noanalysis | 1.22 | 0.00 | 1.23 | 0.00 | 1.00 |
passive_dynamic | 1.83 | 0.00 | 1.82 | 0.00 | 1.00 |
Arm26 | 0.17 | 0.00 | 0.16 | 0.00 | 1.03 |
ToyDropLanding | 0.04 | 0.00 | 0.04 | 0.00 | 1.02 |
… and integrate into Millard muscles and scalar actuators
Brief summary of changes
ComponentPath
overrides for path-dependent parts ofComponent
's public interfaceComponentPath
, rather thanstd::string
, so that the path is pre-parsed ahead of time, which can improve performanceTesting I've completed
Looking for feedback on...
CHANGELOG.md (choose one)
This change is![Reviewable](https://reviewable.io/review_button.svg)