use arrays instead of textures for general storage on this card (image textures
are still stored as texture). Textures were found to be faster on older cards,
but the limits on 1D texture size have not increased along with the memory size,
which meant that the full 6 GB could not be used.
The performance actually seems to be slightly better with arrays in some tests
on Titan. For older cards there seems to be a bit of a mix, some are better and
others not. We may change those to use arrays too, but more testing is needed,
only Titan and Tesla K20 (sm_35) is changed for now.
The fact that arrays are faster is a bit surprising, as others found textures
to be faster on Kepler. However even if they were, the memory limitation is
more important to solve anyway.
https://research.nvidia.com/publication/understanding-efficiency-ray-traversal-gpus-kepler-and-fermi-addendum
Global OpenMP threads override is not a good idea because this would affect
all possible OpenMP blocks running at the same time as simulation.
And that was actually a big on restoring number of threads: code needed to
store current number of threads, not maximal one.
I’m not close to understand why this has worked fine until today... :/
Only tested with Debian, but I would not expect any issue with Fedora/Suse/Arch, this is a quite simple change!
Problem was introduced with r54648, which determined the initial interval for the Newton-Raphson method using the "total_time" of the collision - but this info is only defined for regular collisions, not
for the raycasting used in boids to find the "ground object". To ensure correct behavior, now clear the collision info before using it (good practice in any case), then check the inv_total_time variable
and use the standard 0.001 step if not defined.
scale and rotation in mapping node, there would be shearing, and the only way
to avoid that was to add 2 mapping nodes. This is because to transform the
texture, the inverse transform needs to be done on the texture coordinate
Now the mapping node has Texture/Point/Vector/Normal types to transform the
vector for a particular purpose. Point is the existing behavior, Texture is
the new default that behaves more like you might expect.
was regression in r59665 (fix for [#35002]).
now when checking a segment that lies on the same plane as the triangle intersects,
clip the segment by the triangle bounds so we know the points remain inside the triangle.
button under the mouse, and so clicking the button failed without first
moving the mouse. Similar issues could happen with popups and switching
windows.
The problem is that a notifier was being used to deactivate possible other
activated buttons, but as notifiers are delayed it could undo the correct
activation too. If anyone notices hanging tooltips or other issues after this
fix, please report them.
seems that somehow not having this is causing keyboard events to be caught by
SDL. This was removed because it broke addons that could use SDL, now set the
environment variable only temporary during SDL initialization.
This may have been causing issues with keyboard events getting missed in the
game engine, but I couldn't confirm the issue here.
normal and point parameter types of OSL shaders are creating SOCK_VECTOR sockets in the script node. When these sockets are in turn used to define the fixed input values for these parameters they get
converted as OSL vector always, losing the distinction of vector/normal/point. To prevent OSL rejecting the value due to type mismatch, explicitly define the parameter defaults in the OSL script node
compiler function as vector, normal and point (unused types will simply be ignored).
