yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxxxxxx>]

On 05/03/15 15:37, Václav Šmilauer wrote:
> (sorry for PM)
>> I assume you are running your simulation in parallel. If not identical
>> simulations should give identical results (at least in Yade).
> That's what everybody says, but did you try to run the same sim and see?
This, at least, yes. Even with quite complex solid-fluid coupling:

yadedaily oedometer.py (it is in example folder)
Yade [1]: O.run(100,1); print O.bodies[10].state.vel
Vector3(0.000765388955990568,-0.0021344971932796193,0.0008870185413005638)

and again:
...
Vector3(0.000765388955990568,-0.0021344971932796193,0.0008870185413005638)

I think the register vs. RAM thing is decided at compile time in many
cases, though not always.
I've read something about that recently (kind of webpage explaining why
(1.2+1)==(1+1.2) could return false eventually).

Bruno