The idea is to initialize the vectorization table with one
entry from the previous blocks if it's the same for all predecessors.
In order to not speculatively load out-of-bounds, backends need to
set a new bounds_checked_modes option indicating variable modes
for which per-component bounds checks are supported.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39373>
By default, load/store vectorization uses nir_round_up_components()
to round up loads and possibly writemasked stores to the next valid
NIR vector width. However, some backends may not support load/stores
at all sizes. For example, older Intel supports only power-of-two
vector widths. Newer Intel also supports vec2 and vec3, but not
vec5/6/7. By providing a callback, backends can request promotion
to their next supported memory load/store vector width.
The existing "should we vectorize?" callback should continue to return
false for unsupported vector widths (i.e. beyond the maximum supported).
With this new callback, they do not need to say "no" to vectorization
that would normally produce an unsupported count (e.g. vec5/6/7) but
instead request that the component count be rounded up appropriately.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39250>
This adds a new option, round_up_store_components, which rounds up the
number of components for stores that support writemasking to the next
valid vector size. For example, vec4+vec2 stores would round up from
6 components (which wouldn't be supported) to a full supportable vec8
store, relying on writemasking to ensure the correct pieces are written.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39250>
This is helpful to identify shared mem access for writing more generic code
operating on nir intrinsics.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39219>
Unifies nir per instruction float control.
In the future this can be split into contract/reassoc/transform
like SPIR-V.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com> (except SPIR-V)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39103>
Lowers a shader to use a smaller workgroup to do the same work,
while it will still appear as a bigger workgroup to applications.
To achieve this, the pass augments the CF of the shader
so that each real subgroup will execute two or more logical
subgroups. A logical subgroup represents what the application
can observe as a subgroup.
The size of a logical subgroup is the same as a real subgroup.
Only one logical subgroup may be executed per real subgroup
at the same time. This ensures that all subgroup operations
keep working and the subgroup invocation ID stays the same.
- When the CF contains barriers, we can't just repeat
the code and we need to augment each CF node individually
so that they are aware of logical subgroups.
- In case parts of the CF don't contain any barriers, we can simply
repeat and predicate that CF for each logical subgroup.
It is technically not necessary to implement this strategy, but
in practice it helps reduce the amount of branches in the shader
and therefore improves compile times.
The pass is mainly intended for working around HW limitations,
for example when the HW has an upper limit on the workgroup size
or doesn't support workgroups at all, but the API requires a
certain minimum.
Notes:
- Only applicable to shader stages that use workgroups
- Hits an assertion when called on smaller workgroups
- Always flattens workgroup size to 1D
- Creates local variables
- Does not change subgroup size
- Variable workgroup size not supported yet, maybe later
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Anna Maniscalco <anna.maniscalco2000@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37985>
Introduce a new NIR pass called nir_remove_outputs which works on
lowered I/O intrinsics and can remove any output varying or sysval.
This is meant to replace custom solutions in drivers,
such as radv_remove_varyings and similar.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33928>
We don't need one bit per bitsize per instruction if only one actually
matters in the end.
First step towards moving NIR in the direction of full float_controls2
only.
Also rename this from fp_fast_math, because that name implied that 0 is
the no fast math mode, while the opposite was the case.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39026>
GLSL defines gl_SampleMaskIn as :
"a fragment language that indicates the set of samples covered
by the primitive generating the fragment during multisample
rasterization"
when variable rate shading is enabled, a single invocation might cover
multiple samples. The lowering done in nir_lower_single_sampled() does
not account for that case, so add an option to selectively disable it.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38641>
This is potentially nicer for some drivers. AMD drivers will use it.
mesa_frag_result_get_color_index will be used often.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38604>
nir_lower_clip_cull_distance_array_vars was sneakily updating
shader_info::clip/cull_distance_array_size.
This moves the gathering into a new function
nir_gather_clip_cull_distance_sizes_from_vars.
v2: remove assertions that prevented nir_lower_clip_cull_distance_array_vars
from being used with non-compact arrays
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com> (v1)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38465>
We used load_frag_coord_unscaled_ir3 for loading the fragment coord for
input attachments in GMEM, where the normal scaling for gl_FragCoord
shouldn't be used. However with custom resolve a different scaling will
apply to attachments in GMEM. Separate "unscaled" from "gmem" and rename
the NIR options, in preparation for this.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38451>
The last holdouts of the var options are gone so we can just emit the
system values. This is overall simpler as it confines all the sysval to
var logic to nir_lower_sysvals_to_varyings().
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38562>
All drivers use the same callback and it is unlikely that new drivers will use
this pass since it has better replacements today (lower_mem_bit_sizes for
memory, and it never worked for I/O). This should discourage as much.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38533>
Shaders might declare PLS vars as inout but might just use them as in
or out but not both. This pass detects those cases and adjusts the
variable/deref modes accordingly.
This pass should be called before nir_lower_io_vars_to_temporaries(),
otherwise the copy_derefs will be inserted, turning unused variables
into used ones.
This should ideally be called after DCE to make sure we don't leave
PLS inout variables behind.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Eric R. Smith <eric.smith@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37110>
Rather than adding another boolean to optionally lower PLS vars, pass
the types we want to lowers through a nir_variable_mode bitmask.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Eric R. Smith <eric.smith@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37110>
This adds some new call operations to handle various parts of the
reductions.
cmat_reduce: is the initial toplevel operation from SPIR-V
this is used after lowering for row/col operation on single hw
supported matrix sizes. The spir-v operation is lowered into
multiple of these on flex dimensions, but also can be lowered into
others.
cmat_reduce_finish:
after multiple reduction operations on a flexible dimension matrix,
there is often subsequent operations on the output matrices to
finish the operation.
cmat_reduce_2x2:
this takes 4 input matrices, and 1 dst to do a 2x2 reduction op.
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38389>
With coopmat2 a bunch of functions need a lot of lowering passes
to happen before they can be lowered, so mark them as to be lowered
later.
Drivers needing these should call the nir_remove_non_cmat_call_entrypoints
where they remove entrypoints now, and call the original nir_remove_non_entrypoints
after lowering coopmat2.
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38389>
This adds a new instruction type to handle cooperative matrix calls.
This clones the call instr, drops callee, and adds a single metadata
slot and a call operation (dummy only for now).
(Not NACKed by Alyssa)
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38389>
Only needed by zink. This clip/cull distance separation pass is needed
to remove nir_io_separate_clip_cull_distance_arrays, so that all shared
GLSL code only uses merged clip+cull distance outputs.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38452>
This aborts if a pass would make any progress. It can be used to assert that:
- our minimalist pass invocation loops in drivers are sufficient and don't
leave any unoptimized code in the shader
- our lowering is sufficient and other passes don't add instructions that
would cause lowering having to be repeated
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Acked-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/38406>
Based on existing softfloat64 support and Berkeley SoftFloat. This is
targeted at drivers that can't preserve denorms, so operations where
denorm support is irrelevant like conversions to/from integers aren't
handled.
Because the existing mechanism used by Gallium for softfloat64 doesn't
support includes, we unfortunately can't extract common code into a
header. This can be done later if we switch Gallium to using glslang and
spirv-to-nir.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37608>
