This custom builder implements fine-grained instance node bounds
calculation by looking at all AABBs at tree depth 2.
Shaves off 0.3ms in the start scene for Indiana Jones: The Great Circle
on Deck (roughly 29.1ms->28.7ms).
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32797>
This is mostly adapted from radv's BVH building. This defines a common
"IR" for BVH trees, two algorithms for constructing it, and a callback
that the driver implements for encoding. The framework takes care of
parallelizing the different passes, so the driver just has to split the
encoding process into "stages" and implement just one part for each
stage.
The runtime changes are:
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
The radv changes are;
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31433>
DOOM Eternal builds acceleration structures with inactive primitives and
tries to make them active in later AS updates. This is disallowed by the
spec and triggers a GPU hang. Fix the hang by working around the bug.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27034>
Adds a cost field to radv_ir_node and uses it to model the cost of tree
depth. This improves framerates by 2% if my benchmarking is correct.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19756>
I need space ...
Furthermore, this only gets used during the build, and we can eat
the cost of generating the AABB a second time there.
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18692>
Optimizes code a tiny bit, and avoid the hack of encoding the root
node id in the low bits of the BLAS address in an instance node.
This is needed to adjust serialization/deserialization as the
instance address there has to be the base address, so this avoids
some wrapping/unwrapping.
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18692>
Use pointer arithmetic from GL_EXT_buffer_reference2 to replace explicit
struct sizes. For that to work we also need to fix BDA alignment by setting
it to the smallest possible value (default is 16).
We could also replace the INDEX macro with a[b] operators, but that's actually
a change in behavior because a[b] always uses 64bit arithmetic.
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18765>
This patch adds a header that is shared between the accel struct build
kernels and the dispatch code.
Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18376>
We only skipped the `accelerationStructureReference == 0` case
which could lead to reading uninitialized memory.
This patch introduces a NULL_NODE_ID constant and uses it
to handle inactive nodes properly.
I do not think, that this case is common in practice.
Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18141>
Thos are only included once, so just move the code into the
corresponding .comp files.
Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17963>
