Rigidbodies: time dependent properties / and / rigid body outputs

[gpusph]

Rigidbodies: time dependent properties / and / rigid body outputs

[alda30]

By studying some simple examples of ODE I found the answer to the second questions.
At the end of g_callback class, I added this:

if (t>15.5 && t<16) {
        outputData << t << " " << dBodyGetForce(cylinder.m_ODEBody)[2] << " " << dBodyGetLinearVel(cylinder.m_ODEBody)[2] << endl;
}

It writes time, force and its velocity in z direction.
and I just needed the data on the time span between 15.5 to 16 seconds.

[Narcolessico]

Hello,

I would not recommend changing simulation parameters in the gravity callback: it is only designed to return a float3 and changing the simulator parameters might break the consistency of the simulation (e.g. the engine assumes ODE gravity is the same as GPUSPH gravity). Actually, class encapsulation itself should not allow for it.

This being said, it is a smart workaround that anticipates the callback for imposing forces, which is already implemented and will be released soon (sign up to stay up to date ;) ). In case you need to access the force from another part of the code (e.g. making your own Writer class and adding it with given frequency), you can find the resulting force acting on a body in GlobalData::s_hRbTotalForce[device_index][object_index].

[alda30]

Thanks Narcolessico for your comment.
When I use GlobalData::s_hRbTotalForce[device_index][object_index] I always get zero in all three directions.
I guess I fill device and object indices incorrectly. Could you help me how to fill them in?
How can I get the index of an specific object in the simulation? or the device? (is the latter cuda related parameter?)
Sorry for asking this, I am a freshman here!

---

Update: Ok, I think I found out the first parameter is the device number, since I have only one GPU, it should be set to zero, and for the second parameter is the index of rigid body, and again since I have only one rigid body, it should be set to zero again. Now, I have the foces!
Please correct me if I am wrong.

[Narcolessico]

Please note that s_hRbTotalForce only contains the force applied to the body by the fluid, so it should be 0 until the body hits the fluid.

You are correct about the indices of the array.

For any question about the simulator and the API, you're welcome to join #gpusph channel on Freenode.

[alda30]

Thanks.
I did so.