[Camino-users] Camino trajectories generation

[Hall, Matt]

Hiya,

Let me see if I can answer these:

Dear all,
I am a new user of Camino. I would like to use Camino to generate the trajectories (inside the closed cylinder PLY mesh).
1. From Camino tutorial: The simple command to generate trajectories is:
datasynth -walkers 100000 -tmax 1000 -p 0.0 -voxels 1 -initial uniform -duration 0.2 -trajfile mytraj.traj
I am wondering about the particles stepping size in this case. The particles stepping size should be sigma = sqrt(6*D0*dt), where D0 is diffusivity coefficient and dt is time step. But in the above command, there is no D0 value. So, how it works? Is there specify defaults value for D0?

The step length is indeed set via the time increment and the diffusivity (sqrt[6D0dt]). The diffusivity can be set from the commandline with the –diffusivity switch – like everything in camino it should be specified in SI units. The default value is 2.1E-9 and is given in tools/CL_Initializer.java (the variable’s called DIFF_CONST)


2. Now, I would like to generate trajectories inside only one closed cylinder (defined by PLY mesh). The command is:
datasynth -walkers ${WALKERS} -tmax ${TVAL} -p 0.0 -voxels 1 -initial intra -voxelsizefrac 1 -diffusivity 0.5E-9 -duration 0.05 -substrate ply -plyfile ${plyfile} -trajfile ${mytrajfile}

I would not recommend using a ply file for a single cylinder. There are default cylinder substrates which will do the job more efficiently. First of all, are you wanting to simulate spins inside the cylinder only or also outside? If it’s the latter the packing of the cylinders in space will make a difference – square or hex both use cylindes of a fixed size. If you are only simulating inside a cylinder, then you can use any parallel cylinder substrate and it will give you results from a single cylinder. You can specify these options with –initial intra or –initial uniform.


Where ${WALKERS}=131072 and ${TVAL}=1000. The header of my plyfile is
ply
comment closed surface
format ascii 1.0
element vertex 8071
property float x
property float y
property float z
element face 16138
property list uchar uint vertex_indices
end_header

The question are:
2.a. Is this command correct?
2.b. Why is the generation running very very slowly? (just 1% finished for more than 12 hours running)

Your command looks about right, although see above about not using PLY files. It is possible that your performance is poor because of network latency. If you are writing your PLY file to a network drive this will be very slow. Try writing to a local drive and then copying it over when it’s finished.


Thank you so much.

Hope this helps,
Matt.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.nitrc.org/pipermail/camino-users/attachments/20180131/c7bd4b68/attachment.html>