Both per-primitive and per-vertex space is allocated in MUE in 8 dword
chunks and those 8-dword chunks (granularity of
3DSTATE_SBE_MESH.Per[Primitive|Vertex]URBEntryOutputReadLength)
are passed to fragment shaders as inputs (either non-interpolated
for per-primitive and flat vertex attributes or interpolated
for non-flat vertex attributes).
Some attributes have a special meaning and must be placed in separate
8/16-dword slot called Primitive Header or Vertex Header.
Primitive Header contains 4 such attributes (Cull Primitive,
ViewportIndex, RTAIndex, CPS), leaving 4 dwords (the rest of 8-dword
slot) potentially unused.
Vertex Header is similar - it starts with 3 unused dwords, 1 dword for
Point Size (but if we declare that shader doesn't produce Point Size
then we can reuse it), followed by 4 dwords for Position and optionally
8 dwords for clip distances.
This means we have an interesting optimization problem - we can put
some user attributes into holes in Primitive and Vertex Headers, which
may lead to smaller MUE size and potentially more mesh threads running
in parallel, but we have to be careful to use those holes only when
we need it, otherwise we could force HW to pass too much data to
fragment shader.
Example 1:
Let's assume that Primitive Header is enabled and user defined
12 dwords of per-primitive attributes.
Without packing we would consume 8 + ALIGN(12, 8) = 24 dwords of
MUE space and pass ALIGN(12, 8) = 16 dwords to fragment shader.
With packing, we'll consume 4 + 4 + ALIGN(12 - 4, 8) = 16 dwords of
MUE space and pass ALIGN(4, 8) + ALIGN(12 - 4, 8) = 16 dwords to
fragment shader.
16/16 is better than 24/16, so packing makes sense.
Example 2:
Now let's assume that Primitive Header is enabled and user defined
16 dwords of per-primitive attributes.
Without packing we would consume 8 + ALIGN(16, 8) = 24 dwords of
MUE space and pass ALIGN(16, 16) = 16 dwords to fragment shader.
With packing, we'll consume 4 + 4 + ALIGN(16 - 4, 8) = 24 dwords of
MUE space and pass ALIGN(4, 8) + ALIGN(16 - 4, 8) = 24 dwords to
fragment shader.
24/24 is worse than 24/16, so packing doesn't make sense.
This change doesn't affect vk_meshlet_cadscene in default configuration,
but it speeds it up by up to 25% with "-extraattributes N", where
N is some small value divisible by 2 (by default N == 1) and we
are bound by URB size.
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20407>
Instead of using 4 dwords for each output slot, use only the amount
of memory actually needed by each variable.
There are some complications from this "obvious" idea:
- flat and non-flat variables can't be merged into the same vec4 slot,
because flat inputs mask has vec4 stride
- multi-slot variables can have different layout:
float[N] requires N 1-dword slots, but
i64vec3 requires 1 fully occupied 4-dword slot followed by 2-dword slot
- some output variables occur both in single-channel/component split
and combined variants
- crossing vec4 boundary requires generating more writes, so avoiding them
if possible is beneficial
This patch fixes some issues with arrays in per-vertex and per-primitive data
(func.mesh.ext.outputs.*.indirect_array.q0 in crucible)
and by reduction in single MUE size it allows spawning more threads at
the same time.
Note: this patch doesn't improve vk_meshlet_cadscene performance because
default layout is already optimal enough.
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/20407>
Use a struct for various common parameters rather than per stage
structure or arguments to stage specific entrypoints.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Felix DeGrood <felix.j.degrood@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23942>
It should be providing a write mask that is all the channels. Drop the
one case for load_input where we stomp this for no good reason. Also,
make ALU write-masking AND with the existing mask. This prepares us for
the next patch where we convert to new-style registers.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24104>
Sampler always writes 4/8 register worth of data but for ld_mcs only
valid data is in first two register. So with 16-bit payload, we need to
split 2-32bit registers into 4-16-bit payload.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22802>
Convert it to an opt-in for backends to prefer and use nir_load_texture_scale
instead of txs for nir lowerings.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Suggested-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24054>
It doesn't do anything yet. We leave that to the subsequent patches so we can
keep the tree-wide refactor as simple as possible.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23089>
OpenCL 3.0 core requires __opencl_c_subgroups to be set, the OpenCL
cl_khr_subgroups extenions can only be enabled if and only if the driver
guarentees independent forward progress between subgroups.
See CL_DEVICE_SUB_GROUP_INDEPENDENT_FORWARD_PROGRESS for more information.
Signed-off-by: Karol Herbst <git@karolherbst.de>
Reviewed-by: Nora Allen <blackcatgames@protonmail.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22893>
GLSL booleans (and hence bool derefs) may be translated either as 1-bit or
32-bit NIR registers, depending whether the backend uses nir_lower_bool_to_int32
or not. Add a knob for this and choose the right type for different backends.
Fixes nir_validate failure on
dEQP-VK.subgroups.ballot_broadcast.graphics.subgroupbroadcast_bvec3 run under
lavapipe. That test indexes into a bvec3 array, and gallivm first lowers bools
and then lowers derefs to registers, resulting in random 1-bit booleans mixed in
with 32-bit bools.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23804>
Cuts down serialized size from 2850288 to 1377780 bytes.
Reduces clinfo with Rusticl time by 40% for debug builds.
(Old data, but the point stands)
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15996>
Code already exists to convert SHADER_OPCODE_SHUFFLE into a simple MOV
when either source is constant. However... the constants have to
actually get into those sources!
On a shader that I'm working on that multiplies very large matrices using
lots of subgroup operations,
-SIMD8 shader: 1378 instructions. 3 loops. 793896 cycles. 0:0 spills:fills, 23 sends, scheduled with mode non-lifo. Promoted 0 constants. Compacted 22048 to 21664 bytes (2%)
+SIMD8 shader: 346 instructions. 3 loops. 61742 cycles. 0:0 spills:fills, 23 sends, scheduled with mode top-down. Promoted 0 constants. Compacted 5536 to 5216 bytes (6%)
No changes in shader-db or fossil-db on any Intel platform.
v2: Merge a bunch of identical cases. Suggested by Ken.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com> [v1]
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23609>
This is described in Wa_14014617373 and a programming note has
been added to specification.
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23682>
For historical reasons, nir and vtn were compiled together,
and a bunch of vtn specific targets were defined in
src/compiler/meson.build.
Now that we can, make src/compiler/spirv produce an internal
library that depends on NIR, and is used by the drivers/tools.
Also move the vtn specific targets into that directory's
meson.build.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23668>
These were historically in the spirv+nir combo, but the common mesa clc
is a better home for them.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Nora Allen <blackcatgames@protonmail.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23667>
There is an actual external libclc and we do use it, so rename the
internal common library to avoid confusion.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Acked-by: Nora Allen <blackcatgames@protonmail.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23667>
There's plenty of places we can use these new and shiny helpers, so
let's clean up the code a bit.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23460>
In zink-on-anv fs-mod-dvec3-dvec3.shader_test, we were memsetting 2MB of
last_grf_write 2400 times, multiple times through the scheduler. Just
resetting for the processed instructions reduces runtime from 21s to 16s.
No change on steam shader-db runtime across several runs.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23635>
