Forum Replies Created
-
AuthorPosts
-
sharifBlocked
Here is my experience for compilation of FFMPEG on Linux, with the hope that it helps:
I tested FFMPEG with this patch on Linux (Ubuntu 14.04). The code was almost ready for compilation but for a small modifications. However the problem showed up during runtime. The problem was that FFMPEG distribution on Ubuntu PPA is static and not shared (only .a libraries are installed). I needed to compile FFMPEG as a shared library myself with the following config (.so libraries are compiled now):
sudo apt-get install yasm libvpx. libx264. cd ~ git clone git://source.ffmpeg.org/ffmpeg.git ffmpeg cd ffmpeg mkdir Build cd Build ../configure --enable-shared --enable-gpl --enable-version3 --enable-libx264 --cc="gcc -m64 -fPIC" --extra-ldflags="-ldl" make -j5 sudo make install sudo ldconfig
This script installs FFMPEG to a path that is easy to find for CMAKE without any configurations. So checking on
SOFA-EXTERNAL_FFMPEG
is adequate.Regards,
SharifsharifBlockedDear Talbot,
Thank you for your comprehensive response. Can you please explain more on the differentiation procedure? What should we do if our force field depends on location as a vector? In “VaccumSphereForceField”, force is computed according to the following (bold variables are vectors):
f = (dp/|dp|)*(|dp|-r)*stiffness-v*(|dp|-r)*damping
f
is the force vector,dp
the position vector of a particle relative to the center of the field,r
is the radius of the sphere where the forcefield applies,stiffness
is a user-input parameter specifying the intensity of the field,v
is the velocity of the particle, anddamping
is another user-input parameter which determines the viscosity of the field. As it is seen, our force field is best described by the following:f = f(dp, v)
According to what you said, we should compute
df/dt * dx(t+dt)
foraddDForce
. How is it when we deal with vectors? Dogradians
apply in this case?As another question, what is
mparams.kFactor
and how is it set in SOFA?As a suggestion, it would be perfect if this forum supports mathematical notations. It is hard to be precise with mathematics using normal notations!
Best regards,
SharifsharifBlockedHi Talbot,
Thank you for your quick response. Let’s assume that you have moved a mechanical object out of the visible part of the scene, or gravity (or another force field) has caused a mechanical object to go out of the screen. So, to be precise, I mean are there any functions that automatically change the zooming so that all mechanical objects fit into the visible part of the screen? In other words, can we order the camera to change its zooming level so that all the objects (and of course their bounding boxes) be in the view of the camera?
Thank you,
SharifsharifBlockedDear Talbot,
Thank you for this post. However the patch you have provided here has been subject to automatic changes and must be manually modified before can be applied. (For instance, look at — symbols that should have been three -‘s, or —). Can you put the original file here so that we can use it more comfortably?
Regards,
Sharif16 May 2015 at 12:37 in reply to: [SOLVED] Setting a convenient environment for SOFA development #3313sharifBlockedDear Talbot,
Thank you for your encouragement. In fact, implementation of SofaCUDA is a much greater job than reading manuals and compiling it, and what I did was nothing more than compiling it and trying to get it to work. It is great to know that many Sofa developers normally use Linux as the “incubator”. It can be a sign that bugs may occur in Windows or Mac OS more frequently (I know that developers try hard to keep sofa bug-free in all operating systems, but that is a possibility.) As a secondary choice to Ubuntu, do you also recommend Arch Linux?
Compiling SofaCUDA took me much time, and at last it was only possible by downgrading the version of CUDA SDK and NVIDIA card driver. So by VGA, I meant VGA card (or Video Card). I want to ask if there is any unwritten rules about SofaCUDA and its compatibility with different Video Cards, CUDA SDK versions, etc.
I also have a suggestion. Already, there is enough documentation for Sofa users and Sofa developers to start working with this framework. However, for more advanced usage of the framework and a steeper learning curve, more accurate documentation and tutorials is needed. This way, sofa developers can leverage a greater enthusiastic community of developers and more expedited growth of the framework.Sincerely
Sharif -
AuthorPosts