Otherwise we run into problems by putting this optimization loop
before I/O lowering, where there might still be 8-bit values that
haven't been lowered to 16 or 32. Once that's done, any remaining
movs or vec ops will have higher bit sizes already.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
Add support for 16-bit UBO loads, delete handling of byte-addressed
UBO loads (which I think was never used anyway) and add handling
for the component const index to optimize out unneeded extractResults.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
Derefs have index-based access semantics, which means we don't need
custom intrinsics to encode an index instead of a byte offset.
Remove the "masked" store intrinsics and just emit the pair of atomics
directly. This massively reduces duplication between scratch, shared,
and constant, while also moving more things into nir so more optimizations
can be done.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
We still use shader_temp as a temporary variable mode to differentiate
which variables have simple deref patterns vs ones that need to be
lowered to ssbo, but then we put it back to mem_constant when we're
done to restore sanity.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
There's a few changes in here that are very inter-related.
First, we stop lowering load_deref on shader_temp to load_ptr_dxil,
and just leave it as load_deref. In order for that to work, we need
the derefs to be in a shape that's acceptable to DXIL, so the only
current producer of shader_temp loads (the CLC frontend) needs to
run some lowering passes on them first.
The DXIL backend is augmented to just write out deref indices while
walking a deref chain, which will get combined in the load op into
a GEP instruction. For non-mesh/raytracing shaders, these are required
to be single-level scalar arrays, but the complexity here is preparation
for when we don't need to do that anymore.
Additionally, the const lookups are changed from using a hash table
to just putting an index on the variable.
All of this together is enough to enable the authored-forever-ago test
which uses indirect array access into a const packed struct. The
load_ptr_dxil handling didn't deal with packed structs / unaligned
accesses, but now that we're in a logical address space with derefs
instead of physical, there's no alignment to deal with anymore and
the fact that it's packed goes out the window.
This removes one custom DXIL intrinsic.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
DXIL requires GEP chains to point to a global variable that's a flat
array of primitive types. If we're converting deref chains to GEP
chains, we're effectively in a logical address space, which means
we can do things like change sizes of variables, since we know
they won't alias with anything else. If they could alias, we'd be
lowering them to an explicit I/O op instead. That means we can
start disabling some of the low-bit-size lowering.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
Now that we try harder for memcpys, we can use nir's complex usage helper.
We also can just mark the vars instead of using a hash map, since location
doesn't mean anything for constant vars.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
For the case of memset, the SPIR-V translator produces a copy from
a byte array of 0s. If we wait to lower memcpys until after types
are sized, we can potentially turn those 0s into SSA zeros and remove
the entire constant array.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
I was seeing u2u64 still in my final shader after pack/unpack were
lowered, which sounds to me like some other optimizations are missing
for detecting the post-lowering pack/unpack patterns, but let's at
least add some patterns for the simple cases.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
We'd like to use this callback to adjust loads and stores from things
that are unsupported to things that are supported, but if the input
is already supported, we'd prefer not to change it. Rather than making
up a bit size that'd work and doing a bunch of pack/unpack bit math,
only return a different bit size if the input one doesn't work for us
(i.e. can't load enough memory or just an unsupported size entirely).
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
The var splitting pass can rearrange the variables as long as their
position in memory doesn't matter. For block-arranged variables,
or things like memcpys or casts, the layout matters, but atomics
don't imply anything about the layout of the overall variable, so
don't treat them as "complex" for this use case.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
Idiomatic DXIL has constants contained within global variables rather
than a big blob of data. Doing this allows us to have 16-bit and 64-bit
data as well, where normally bitcasts would be disallowed on variable
GEP chains.
Unfortunately, DXIL validation requires SOA to be turned into AOS,
which means we need to split structs. We want to be able to run this
on nir_var_mem_constant variables which have constant initializers,
so add a bit of logic to handle that case, and relax the mode validation.
There's nothing special about the modes it was set up to handle.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23173>
Instead of on Android. Which allows an end user to turn off expat
without breaking or disabling Intel support. I've additionally
refactored to separate expat and xmlconfig a bit more in the root
meson.build
This does make expat a hard dependency for building Intel tools, despite
the fact that only aubinator actually requires it. This simplifies the
build for the common case, and in the event that someone wants to build
the Intel tools and doesn't have libexpat, they can fall back to the
meson wrap for expat instead.
fixes: 75276deebc
closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8791
Reviewed-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23605>
Since Mesa drivers are now version-locked to the loader, that means that
we never need to support a newer hardware driver than the loader, and thus
don't need to generate dynamic dispatch stubs. This is great news, given
that we don't test those paths, and it involved delightful features like
arrays of hex for code to be pasted into executable memory.
More code removal will follow, this is the first cut of "don't generate,
and DCE generation code".
Fixes: #9158
Acked-by: Marek Olšák <marek.olsak@amd.com>
Acked-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23451>
While this is a generic bit twiddling ALU instruction, it's especially useful
for address calculations, since the architecture's tiled textures use Morton
coding within the tiles.
This will be used when lowering image_texel_address on AGX, as part of the image
atomics implementation. I don't know if there's any other neat uses I could
detect with opt_algebraic, this doesn't seem like an operation a shader would
open-code... Maybe useful for BVH building or something...
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23513>
Samuel Pitoiset