Everything was fine if one batch is always used with instancing. But problem arise if the next drawcall for this batch is not using instancing as the attrib divisor stays set to 1 in th VAO.
As instancing is less used than normal drawing I prefer to reset the divisor after drawing as it is reset before drawing instances.
This was caused by small float precision being insuficient. The blue component of R11F_G11F_B10F has lower precision than the other 2 components. This resulted in colors drifting towards a yellowish tone.
Using RGBA16F for the concerned buffer. This double the memory usage of the framebuffers and add subsequent bandwidth usage.
This bug (explained here https://github.com/dfelinto/opengl-sandbox/blob/downsample/README.md) is breaking eevee beyond the point it's workable.
This patch workaround the issue by making sure every fbo have mipmaps that are strictly greater than 16px. This break the bloom visuals a bit but only for this setup.
This change affects CUDA GPUs not connected to a display or connected to a
display but supporting compute preemption so that the display does not
freeze. I couldn't find an official list, but compute preemption seems to be
only supported with GTX 1070+ and Linux (not GTX 1060- or Windows).
This helps improve small tile rendering performance further if there are
sufficient samples x number of pixels in a single tile to keep the GPU busy.
Best guess is that cuInit() somehow interferes with the AMD graphics driver
on Windows, and switching the initialization order to do OpenCL first seems
to solve the issue.
* Use common TextureInfo struct for all devices, except CUDA fermi.
* Move image sampling code to kernels/*/kernel_*_image.h files.
* Use arrays for data textures on Fermi too, so device_vector<Struct> works.
Instead of trying to be clever with swaps and lazy updating the weight
data, simply recalculate one single array. To improve performance, use
threading for that.
This was introduced on 9ad2c0b615.
Although this still doesn't fix the issue, it updates the preview
system to use COLLECTION_DISABLED as intended.
What is missing now is for the flushing to work effectively.
This add the possibility to add screen space raytraced shadows to fix light leaking cause by shadows maps.
Theses inherit of the same artifacts as other screenspace methods.
Two issues here:
- Checking table size to be non-zero is not a proper way to go here. This is
because we first resize the table and then fill it in. So it was possible that
non-initialized table was used.
Trickery with using temporary memory and then doing table.swap() might work,
but we can not guarantee that table size will be set after the data pointer.
- Mutex guard was useless, because every thread was using own mutex. Need to
make mutex guard static so all threads are using same mutex.
Tried 101 but it gives colisions.
I think 257 is enough now that we dont have thousands of uniforms.
This gives some noticeable performance improvement.
Could be refined further.
The issue was caused by light sample being evaluated to nan at some point.
This is root of the cause which is to be fixed, but is very hard to trace down
especially via ssh (the issue only happens on AVX2 release build). Will give it
a closer look when back to my AVX2 machine.
For until then this is a good check to have anyway, it corresponds to what's
happening in regular radiance sum.
