e5122a5543
glsl: add a NIR based varying linker
...
With a NIR based linker we get better xfb packing, and we no longer
depend on the GLSL IR optimisations to be able to link shaders
with a large amount of dead input/outputs.
Acked-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
e3a45a4778
glsl: implement lower_packed_varyings() as a NIR pass
...
This is essentially the old GLSL IR packing pass rewritten as a
NIR based pass.
Doing this packing in NIR after we have preformed NIRs
optimisation passes can give us better packing results.
Acked-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
c1fbd0b8ab
nir: skip lowering io to scalar for must_be_shader_input
...
These varyings cannot be packed by the GLSL linkers packing pass
so we need to skip this lowering until later when we can properly
handle them.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
99ab530617
nir: abort io info gathering if location is not set or is a temp value
...
Unlike spirv glsl varyings might not have explicit locations set.
nir_shader_gather_info() was once only called at the end of linking
but these days it even gets called in NIR optimisation loops via
nir_opt_phi_precision.
In the following patches we implement a NIR version of the GLSL
varying linker which means we will have varyings with no location
set when nir_shader_gather_info() gets called the first few times,
and temp values set only for the purpose of removing unmatched
varyings between shaders for some calls after that.
Here rather than asserting we simply abort the io info gathering,
when we hit these values.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
cba2fd51a2
nir: add variable data fields required for NIR glsl varying linking
...
These will be used in the following patches that add a NIR based
varying linker.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
43a8454ea8
glsl: add new build program resource helpers
...
These will be used by a new nir based glsl varying linker that will
add varyings directly to the list before the are packed and we lose
the information.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
23ea24e11f
glsl/mesa: move parse_program_resource_name() to common linker_util code
...
This will be shared by a new NIR varying linking pass in following patches
but probably fits better here anyway considering its also used by
shader_query.cpp
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
5d57bd0345
nir/glsl: wrap component_slots_aligned()
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
6dbe075f92
nir/glsl: wrapper field_index()
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
42a97a0aef
nir/glsl: wrapper contains_{double,interger}()
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
7af9459670
nir/glsl: add glsl_record_compare() wrapper
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15731 >
2022-05-16 03:33:18 +00:00
98cc4c3a20
nir: Use nir_shader_instructions_pass in nir_lower_input_attachments
...
This simplifies things a bit and also fixes metadata handling.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16482 >
2022-05-13 22:51:38 +00:00
a170448a18
nir: Put the builder first in lower_input_attachments helpers
...
This is more idiomatic.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16482 >
2022-05-13 22:51:38 +00:00
5410f4ee89
mesa/st: Use lower_indirect_var_derefs in st_nir_lower_builtin
...
Instead of having a special NIR helper for GL stuff, we can now use the
more generic helper and do so directly.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16482 >
2022-05-13 22:51:38 +00:00
e16197c46e
nir: Add a var set version of lower_indirect_derefs
...
This version takes a set of variables and totally lowers indirects on
any variable in the set. We also rewrite the builtin_uniform version to
use the new helper internally.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16482 >
2022-05-13 22:51:38 +00:00
c23b20d43a
nir: Preserve metadata if remove_dead_derefs makes no progress
...
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16482 >
2022-05-13 22:51:38 +00:00
7446acf4b4
compiler: add VARYING_SLOT_CULL_PRIMITIVE
...
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16493 >
2022-05-13 09:43:02 +00:00
bc5c68fc08
nir/opt_algebraic: Optimize Doom Eternal's word extract by LSB.
...
Foz-db GFX10_3:
Totals from 419 (0.31% of 134913) affected shaders:
CodeSize: 4126032 -> 4121756 (-0.10%)
Instrs: 783608 -> 782541 (-0.14%)
Latency: 7889664 -> 7888521 (-0.01%); split: -0.02%, +0.00%
InvThroughput: 1315690 -> 1314863 (-0.06%); split: -0.06%, +0.00%
VClause: 11826 -> 11830 (+0.03%)
SClause: 27736 -> 27734 (-0.01%)
Copies: 50493 -> 50428 (-0.13%); split: -0.13%, +0.01%
PreSGPRs: 23264 -> 23265 (+0.00%)
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com >
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16436 >
2022-05-12 17:10:41 +00:00
938c9d9615
nir: Add a ray launch size addr intrinsic
...
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/15712 >
2022-05-12 15:04:31 +00:00
0f98ed4afe
nir: remove unreachable loop terminators
...
Remove the conditional break statements associated with all
terminators that are associated with a fixed iteration count,
except for the one associated with the limiting terminator.
This logic matches similiar functionality that exists in the
old GLSL IR unrolling code.
This change helps a piglit test pass on the r300 driver once
we switch off the old GLSL IR unrolling code.
Shader-db results IRIS (BDW):
total instructions in shared programs: 17538619 -> 17538595 (<.01%)
instructions in affected programs: 216 -> 192 (-11.11%)
helped: 3
HURT: 0
helped stats (abs) min: 7 max: 10 x̄: 8.00 x̃: 7
helped stats (rel) min: 10.00% max: 12.07% x̄: 11.38% x̃: 12.07%
total cycles in shared programs: 858674910 -> 858672810 (<.01%)
cycles in affected programs: 79540 -> 77440 (-2.64%)
helped: 3
HURT: 0
helped stats (abs) min: 620 max: 800 x̄: 700.00 x̃: 680
helped stats (rel) min: 2.45% max: 2.83% x̄: 2.63% x̃: 2.62%
Reviewed-by: Emma Anholt <emma@anholt.net >
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16399 >
2022-05-12 02:06:31 +00:00
4c3d138e5d
nir: always set the exact_trip_count_unknown loop terminator property
...
Previously we only cared if this was set for the limiting
terminator. However in the following patch we will make use of this
information on other terminators to decide if we can eliminate them.
Reviewed-by: Emma Anholt <emma@anholt.net >
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16399 >
2022-05-12 02:06:31 +00:00
7de3034897
ac/nir: Add I/O lowering for task and mesh shaders.
...
Task shaders store their output payload to VRAM where mesh
shaders read from. There are two ring buffers:
1. Draw ring: this is where mesh dispatch sizes and
the ready bit are stored.
2. Payload ring: this is where the optional payload
is stored (up to 16K per task workgroup).
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com >
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14929 >
2022-05-12 00:29:51 +00:00
df1876f615
nir: Mark negative re-distribution on fadd as imprecise
...
Otherwise, it would mutate `fneg(fadd(-0, 0))` into `fadd(0, -0)` which
isn't correct since -0 + (+0) = +0 + (-0) = +0.
This fixes the OpenCL contraction tests on Iris.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16041 >
2022-05-12 00:05:10 +00:00
25249e8be2
nir/lower_blend: Expand or shrink output variables as needed
...
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
1d22465362
nir/builder: Add a nir_resize_vector helper
...
We're about to use this a couple of places.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
352e32e5ba
nir/builder: Add a nir_trim_vector helper
...
This pattern pops up a bunch and the semantics of nir_channels() aren't
very convenient much of the time. Let's add a nir_trim_vector() which
matches nir_pad_vector().
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
244b654de6
nir/lower_blend: Support SNORM and integer formats for logic ops
...
This fixes 158 of the dEQP-VK.pipeline.logic_op.* tests, once we turn
the feature on.
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
730d2b7660
nir/lower_blend: Stop passing the whole options object around
...
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
dcfffdcad1
nir/lower_blend: Be more explicit about deref assumptions
...
Because we pull the RT from the variable location and use that to look
up formats, we need a constant RT index. To deal with arrays (possibly
of arrays), we would either need to handle array derefs (we don't today)
or we need to require the variables to be split into one variable per
RT. Given that we have to lower indirect derefs anyway (to get constant
indices), we may as well require the client to split output variables by
calling nir_lower_io_arrays_to_elements_no_indirect().
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16309 >
2022-05-11 14:47:33 +00:00
fd75be7986
glsl: Fix ir_quadop_vector validation
...
Some glcts tests have failed due to incorrect processing of `ir_quadop_vector` in
`ir_validation`. e.g:
`GLES31.functional.shaders.builtin_functions.integer.imulextended.int_highp_geometry`
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6461
Fixes: 23cde71bemma@anholt.net >
Signed-off-by: Mykhailo Skorokhodov <mykhailo.skorokhodov@globallogic.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16420 >
2022-05-10 23:53:33 +00:00
60c9a45562
nir/opt_algebraic: Simple xor/ishr optimizations.
...
The first pattern here removes the xor-swap pattern.
Foz-DB GFX10_3:
Totals from 305 (0.23% of 134913) affected shaders:
CodeSize: 1589040 -> 1585164 (-0.24%)
Instrs: 284344 -> 283375 (-0.34%)
Latency: 4205148 -> 4198472 (-0.16%); split: -0.16%, +0.00%
InvThroughput: 708745 -> 708739 (-0.00%)
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com >
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com >
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com >
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16411 >
2022-05-10 19:29:31 +00:00
66e917fff6
nir/opt_algebraic: Fix mask in shift by constant combining.
...
The comment above is correct, but the code to calculate the mask was broken.
No Foz-db changes outside of noise.
Fixes: 0e6581b87ddadschoorse@gmail.com >
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15990 >
2022-05-10 18:47:21 +00:00
7f189e3467
nir: Add upper bound for AMD shader arg intrinsics.
...
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com >
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13155 >
2022-05-10 17:16:03 +00:00
aea935264a
shader_info: Bump the number of images and textures supported
...
OpenCL requires up to 128 read-only images and up to 64 write images.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Reviewed-by: Karol Herbst <kherbst@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15988 >
2022-05-10 11:23:15 -05:00
b37831c606
nir: Gather samplers_used separately from textures
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Reviewed-by: Karol Herbst <kherbst@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15988 >
2022-05-10 11:23:12 -05:00
3c07c3e16d
shader_info: Make images_used a bitset
...
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Reviewed-by: Karol Herbst <kherbst@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15988 >
2022-05-10 11:23:11 -05:00
28f534350c
nir: Stop assuming shader_info::textures_used is 32-bit
...
This isn't a hot path. We don't need to be manually using the
INSIDE_WORD version which will assert if we ever get a bigger texture
index.
Reviewed-by: Karol Herbst <kherbst@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15988 >
2022-05-10 11:23:07 -05:00
625b352f14
nir: Set image_buffers and msaa_images in lower_samplers_as_deref
...
This is where we set images_used so it's less likely that things will
accidentally get out-of-sync.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15988 >
2022-05-10 11:21:39 -05:00
1c3e584dfa
nir/divergence: handle more *_intel intrinsics
...
v2: fix topo/btd (Lionel)
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Acked-by: Timur Kristóf <timur.kristof@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16421 >
2022-05-10 08:49:58 +00:00
f3df3d4c80
glsl: Make all drivers take the GLSLOptimizeConservatively path.
...
Now that all consumers of GLSL use NIR, make the remaining drivers take
the path that relies on NIR to really do optimization.
nouveau steam shader-db runtime -6.69631% +/- 1.29235% (n=12).
No change on shader-db there.
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Reviewed-by: Karol Herbst <kherbst@redhat.com >
Reviewed-by: Rob Clark <robdclark@chromium.org >
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16364 >
2022-05-10 05:03:34 +00:00
9c5fd100cc
nir: add a nir_remove_non_entrypoints helper
...
This code just got duplicated a lot. There is still more, but the
remaining instances do a bit more than just removing other functions.
Signed-off-by: Karol Herbst <kherbst@redhat.com >
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16348 >
2022-05-10 03:37:44 +00:00
4b67d70d22
nir: Fix constant folding for non-32-bit ifind_msb and clz
...
Signed-off-by: Karol Herbst <kherbst@redhat.com >
Reviewed-by: Alyssa Rosenzweig <alyssa@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16348 >
2022-05-10 03:37:44 +00:00
23cde71bb9
glsl: Stop lowering ir_quadop_vector.
...
Now that everybody goes through NIR, glsl_to_nir is happy to handle the
instruction and turn it into nir_op_vec4 instead of going to a temp
variable and back.
No changes on freedreno shader-db.
Reviewed-by: Rob Clark <robdclark@chromium.org >
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16363 >
2022-05-09 22:13:31 +00:00
35d82ecf1e
nir/lower_shader_calls: put inserted instructions into a dummy block
...
When moving code into the main block or loop blocks, put the code into
its own :
if(true) { ... }
block so that we avoid break/continue/return issues.
v2: Also take care of the main block with return instructions
v3: Make deletion more obvious with dummy if blocks (Jason)
v4: Fixup assert for loops (Lionel)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 8dfb240b1fjason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16036 >
2022-05-09 08:43:40 +00:00
9cf986dcff
nir/lower_shader_calls: don't insert code after break/continue
...
When moving code from below to the insertion cursor point, if the
cursor points to a jump instruction, don't bother inserting the code.
It would break the break/continue assumptions of NIR and would not be
executed anyway.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 8dfb240b1fjason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16036 >
2022-05-09 08:43:40 +00:00
51dea59eb4
nir/lower_shader_calls: don't use nop instructions as cursors
...
Stop using nop instructions which are causing issues with
break/continue, instead use a nir_cursor (which brings its share of
pain).
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 8dfb240b1fjason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16036 >
2022-05-09 08:43:40 +00:00
25661ea028
nir/cf: Return a cursor from nir_cf_extract as well
...
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16036 >
2022-05-09 08:43:40 +00:00
d65cf403f3
nir/cf: return cursor after insertion of cf_list
...
This will be useful to cut code from one location and paste it at
another place and later keep pasting after the previous insertions.
v2: update comment (Jason)
deal with stiching 2 empty blocks (Jason)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16036 >
2022-05-09 08:43:40 +00:00
ad864a7c15
nir/lower_tex: Copy more fields in lower_tex_to_txd and friends
...
Fixes NIR validation errors for OpenMW on Panfrost.
Fixes: 1f97819fbemichael.blumenkrantz@gmail.com >
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15939 >
2022-05-07 10:51:10 +00:00
5c24eb721a
nir/gather_info: flag fbfetch on subpass image loads
...
might not be able to determine which output is being read, but these
are definitely fbfetch uses (from lavapipe)
Reviewed-by: Dave Airlie <airlied@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16346 >
2022-05-06 17:04:34 +00:00
Timothy Arceri