[Simulation.Core] Refactor DefaultAnimationLoop + multithreading #3959

alxbilger · 2023-06-16T09:04:10Z

Split all pipeline steps into smaller functions

Each step of the animation loop now corresponds to a function. For example, the following code:

sofa::helper::AdvancedTimer::stepBegin("BehaviorUpdatePositionVisitor");
BehaviorUpdatePositionVisitor beh(params , dt);
gnode->execute ( beh );
sofa::helper::AdvancedTimer::stepEnd("BehaviorUpdatePositionVisitor");

is now in a function:

void DefaultAnimationLoop::behaviorUpdatePosition(const core::ExecParams* params, const SReal dt) const
{
    sofa::helper::ScopedAdvancedTimer timer("BehaviorUpdatePositionVisitor");
    BehaviorUpdatePositionVisitor beh(params, dt);
    gnode->execute(beh);
}

In the end, the pipeline is easier to read:

sofa::core::MechanicalParams mparams(*params);
mparams.setDt(dt);

behaviorUpdatePosition(params, dt);
updateInternalData(params);

resetConstraint(params);

beginIntegration(params, dt);
{
    collisionDetection(params);

    const core::ConstraintParams cparams;
    buildConstraintMatrix(cparams);
    accumulateMatrixDeriv(cparams);

    solve(params, dt);

    projectPositionAndVelocity(nextTime, mparams);
    propagateOnlyPositionAndVelocity(nextTime, mparams);
}
endIntegration(params, dt);
computeIsolatedForces(params, dt);

updateSimulationContext(params, dt, startTime);

AnimateVisitor

AnimateVisitor is no longer used. It was difficult to read its logic (the visitor called other visitors, it relies heavily on pruning which is easy to miss). Instead, I suggest to explicit all the steps from the AnimateVisitor into the DefaultAnimationLoop. This is closer to what is done in the FreeMotionAnimationLoop. Plus, we can find common steps, which makes the code more consistent.

Note also the introduction of ComputeIsolatedForceVisitor (see the comments in the code to understand its purpose).

Multithreading

Since AnimateVisitor is no longer used, SolveVisitor is used instead, providing optional multithreading, i.e. solving each ODE in parallel. Similar to what was done in the FreeMotionAnimationLoop.

Benchmark

Measured on examples/Component/SolidMechanics/FEM/TetrahedronFEMForceField.scn

Without parallelism

[INFO]    [BatchGUI] 1000 iterations done in 14.7031 s ( 68.0128 FPS).
 LEVEL   START    NUM      MIN     MAX   MEAN     DEV    TOTAL  PERCENT ID
   2       0.03    1      10.57   27.32   11.94    2.11   11.94   82.14 ..solve

With parallelism

[INFO]    [BatchGUI] 1000 iterations done in 7.13825 s ( 140.09 FPS).
 LEVEL   START    NUM      MIN     MAX   MEAN     DEV    TOTAL  PERCENT ID
   2       0.03    1       3.53    6.85    4.45    0.81    4.45   63.60 ..solve

The simulation is 2x faster. But there are 4 different objects, so we expected a speed-up of 4x. However, the speed-up is only about the solve function. Another time consuming function is the end event when computing the von Misses stresses. Still, the solve function only shows a speedup of 2.68x, instead of 4x. I cannot explain it for now...

[ci-depends-on https://github.com/sofa-framework/SofaPython3/pull/362]

By submitting this pull request, I acknowledge that
I have read, understand, and agree SOFA Developer Certificate of Origin (DCO).

Reviewers will merge this pull-request only if

it builds with SUCCESS for all platforms on the CI.
it does not generate new warnings.
it does not generate new unit test failures.
it does not generate new scene test failures.
it does not break API compatibility.
it is more than 1 week old (or has fast-merge label).

alxbilger · 2023-06-16T12:16:06Z

[ci-build][with-all-tests]

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp

fredroy · 2023-06-19T00:29:03Z

[ci-build][with-all-tests]

alxbilger · 2023-06-19T13:53:09Z

[ci-build][with-all-tests]

alxbilger · 2023-06-20T16:07:12Z

Another benchmark: examples/Demos/chainAll.scn

With parallelism

1000 iterations done in 41.8915 s ( 23.8712 FPS).

Without parallelism:

1000 iterations done in 91.9302 s ( 10.8778 FPS).

alxbilger · 2023-06-27T14:15:36Z

[ci-build][with-all-tests][force-full-build]

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp

examples/Component/SolidMechanics/FEM/TetrahedronFEMForceField.scn

hugtalbot · 2023-06-29T12:26:35Z

[ci-build][with-all-tests] (check for MacOS)

fredroy · 2023-07-12T07:44:36Z

[ci-build][with-all-tests][force-full-build]

hugtalbot · 2023-07-12T07:47:51Z

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp

+void DefaultAnimationLoop::buildConstraintMatrix(core::ConstraintParams cparams) const
+{
+    sofa::helper::ScopedAdvancedTimer timer("buildConstraintMatrix");
+    unsigned int constraintId = 0;
+    mechanicalvisitor::MechanicalBuildConstraintMatrix buildConstraintMatrix(&cparams, core::MatrixDerivId::constraintJacobian(), constraintId );
+    buildConstraintMatrix.execute(gnode);
+}


This function in a DefaultAnimationLoop does not make sense since only FF-based (weak) constraints can be handled with the DefaultAnimationLoop

What about removing it @alxbilger ?

I don't know what it is supposed to do. I guess we can try to remove it and see if CI complains

have you tested it @alxbilger ?

alxbilger · 2023-07-17T15:40:47Z

I don't understand why the dashboard says 2 scenes are failing. When I see the result, I see only one

hugtalbot · 2023-07-19T07:48:14Z

CI sees apparently twice an error in advancedtimer .. strange

[ERROR]   [AdvancedTimer::end] timer[Animate] does not correspond to last call to begin(cg_timer)

The logic was difficult to read in AnimateVisitor. Instead, all the steps of the animation loop are now explicit in the the code of the animation loop.

Co-authored-by: Hugo <hugo.talbot@sofa-framework.org>

alxbilger · 2023-07-21T12:56:29Z

There is an error in the test because timer ids changed in this PR

sofabot · 2023-07-24T08:58:14Z

[ci-depends-on] detected during build #13.

To unlock the merge button, you must

Merge or close Follow changes from #3959 SofaPython3#362
For this build, SofaPython3/ExternalProjectConfig.cmake.in will be edited by builders with
GIT_REPOSITORY https://github.com/sofa-framework/SofaPython3
GIT_TAG 51a2a5bfb15a74c930ab28b24e64eadc0a1752c8

# Conflicts: # Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp # Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.h

sofabot · 2023-07-27T09:45:26Z

[ci-depends-on] detected during build #14.

To unlock the merge button, you must

Merge or close Follow changes from #3959 SofaPython3#362
For this build, SofaPython3/ExternalProjectConfig.cmake.in will be edited by builders with
GIT_REPOSITORY https://github.com/sofa-framework/SofaPython3
GIT_TAG 2d5f82603ca0ff8a45abe813e9908b1b2944cd43

alxbilger · 2023-07-27T09:47:19Z

The crash that happened on Windows also happens on the master branch on my computer. But I don't understand how it is not detected by the CI

fredroy · 2023-07-30T23:46:19Z

The crash that happened on Windows also happens on the master branch on my computer. But I don't understand how it is not detected by the CI

Maybe it is random, it does not crash on mine (but I only tried like 2, 3 times) on msvc
Nevertheless, I dont think this scene is working as intended anyway.... (the particules do not interact with the static mesh)

hugtalbot · 2023-08-16T07:30:30Z

It seems strange that the factorization of the mapping propagation at the end of the time step makes the scene crashes..
Have you investigated it further @alxbilger ?

alxbilger · 2023-08-25T09:29:31Z

[ci-build][with-all-tests][force-full-build]

sofabot · 2023-08-25T09:29:40Z

[ci-depends-on] detected during build #15.

To unlock the merge button, you must

Merge or close Follow changes from #3959 SofaPython3#362
For this build, SofaPython3/ExternalProjectConfig.cmake.in will be edited by builders with
GIT_REPOSITORY https://github.com/sofa-framework/SofaPython3
GIT_TAG c0dfe37b6e7605a5a09c589a540e6296dd660d25

alxbilger · 2023-08-25T13:51:10Z

[ci-build][with-all-tests][force-full-build]

sofabot · 2023-08-25T13:51:18Z

[ci-depends-on] detected during build #16.

All dependencies are merged/closed and all ExternalProject pointers are up-to-date. Congrats! 👍

alxbilger added enhancement About a possible enhancement pr: status to review To notify reviewers to review this pull-request labels Jun 16, 2023

damienmarchal reviewed Jun 16, 2023

View reviewed changes

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp Show resolved Hide resolved

damienmarchal reviewed Jun 16, 2023

View reviewed changes

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp Show resolved Hide resolved

alxbilger added the pr: highlighted in next release Highlight this contribution in the notes of the upcoming release label Jun 20, 2023

hugtalbot reviewed Jun 28, 2023

View reviewed changes

Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp Outdated Show resolved Hide resolved

hugtalbot reviewed Jun 28, 2023

View reviewed changes

examples/Component/SolidMechanics/FEM/TetrahedronFEMForceField.scn Outdated Show resolved Hide resolved

damienmarchal reviewed Jun 28, 2023

View reviewed changes

examples/Component/SolidMechanics/FEM/TetrahedronFEMForceField.scn Outdated Show resolved Hide resolved

hugtalbot reviewed Jul 12, 2023

View reviewed changes

alxbilger and others added 8 commits July 21, 2023 14:19

Clean DefaultAnimationLoop

9d70257

Stop using AnimateVisitor

f3d1c46

The logic was difficult to read in AnimateVisitor. Instead, all the steps of the animation loop are now explicit in the the code of the animation loop.

Compute isolated forces

6786c65

Introduce data to trigger parallel solving

a0272f3

Default value to false

99dc628

Restore previous behavior

024b69b

Better comment

e7748d0

Co-authored-by: Hugo <hugo.talbot@sofa-framework.org>

Revert example

6133775

remove timer as it breaks events where timers are accessed

5653f87

alxbilger force-pushed the defaultanimloop branch from 794ec14 to 5653f87 Compare July 24, 2023 08:57

Merge remote-tracking branch 'upstream/master' into defaultanimloop

899a98e

# Conflicts: # Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.cpp # Sofa/framework/Simulation/Core/src/sofa/simulation/DefaultAnimationLoop.h

hugtalbot added pr: status ready Approved a pull-request, ready to be squashed and removed pr: status to review To notify reviewers to review this pull-request labels Aug 23, 2023

alxbilger merged commit e0d5432 into sofa-framework:master Aug 25, 2023

hugtalbot added this to the v23.12 milestone Nov 29, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Simulation.Core] Refactor DefaultAnimationLoop + multithreading #3959

[Simulation.Core] Refactor DefaultAnimationLoop + multithreading #3959

alxbilger commented Jun 16, 2023 •

edited

Loading

alxbilger commented Jun 16, 2023

fredroy commented Jun 19, 2023

alxbilger commented Jun 19, 2023

alxbilger commented Jun 20, 2023

alxbilger commented Jun 27, 2023

hugtalbot commented Jun 29, 2023

fredroy commented Jul 12, 2023

hugtalbot Jul 12, 2023

alxbilger Jul 12, 2023

hugtalbot Jul 19, 2023

alxbilger commented Jul 17, 2023

hugtalbot commented Jul 19, 2023 •

edited

Loading

alxbilger commented Jul 21, 2023

sofabot commented Jul 24, 2023

sofabot commented Jul 27, 2023

alxbilger commented Jul 27, 2023

fredroy commented Jul 30, 2023 •

edited

Loading

hugtalbot commented Aug 16, 2023

alxbilger commented Aug 25, 2023

sofabot commented Aug 25, 2023

alxbilger commented Aug 25, 2023

sofabot commented Aug 25, 2023

[Simulation.Core] Refactor DefaultAnimationLoop + multithreading #3959

[Simulation.Core] Refactor DefaultAnimationLoop + multithreading #3959

Conversation

alxbilger commented Jun 16, 2023 • edited Loading

Split all pipeline steps into smaller functions

AnimateVisitor

Multithreading

Benchmark

alxbilger commented Jun 16, 2023

fredroy commented Jun 19, 2023

alxbilger commented Jun 19, 2023

alxbilger commented Jun 20, 2023

alxbilger commented Jun 27, 2023

hugtalbot commented Jun 29, 2023

fredroy commented Jul 12, 2023

hugtalbot Jul 12, 2023

Choose a reason for hiding this comment

alxbilger Jul 12, 2023

Choose a reason for hiding this comment

hugtalbot Jul 19, 2023

Choose a reason for hiding this comment

alxbilger commented Jul 17, 2023

hugtalbot commented Jul 19, 2023 • edited Loading

alxbilger commented Jul 21, 2023

sofabot commented Jul 24, 2023

sofabot commented Jul 27, 2023

alxbilger commented Jul 27, 2023

fredroy commented Jul 30, 2023 • edited Loading

hugtalbot commented Aug 16, 2023

alxbilger commented Aug 25, 2023

sofabot commented Aug 25, 2023

alxbilger commented Aug 25, 2023

sofabot commented Aug 25, 2023

alxbilger commented Jun 16, 2023 •

edited

Loading

hugtalbot commented Jul 19, 2023 •

edited

Loading

fredroy commented Jul 30, 2023 •

edited

Loading