e.g. load X; load W; ==> load XYZW. Verified with a shader test.
This will be used by AMD drivers. See the code comments.
Reviewed-by: Simon Perretta <simon.perretta@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36098>
When we detect that the source is a conversion generated by the pass,
try to get the real source instead of doing a round-trip conversion.
Make sure that the nir_alu_type and the bit_size is the same between what
we need and what's before the detected conversion.
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35744>
The only last users of nir_link_opt_varyings are Vulkan drivers.
One linker error thrown by the optimizations is reimplemented
at the call site.
No interesting shader-db changes (other than random noise).
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36091>
A TF2 shader propagates 0 to the consumer, which eliminates 1 input
if we run algebraic opts and DCE before compaction.
This is a prerequisite for removing all IO var optimizations from the GLSL
linker that are redundant with nir_opt_varyings.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36091>
Will be used by etnaviv too.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35753>
It's a sysval in mesh shader, but it share the same
slot number with VARYING_SLOT_TESS_LEVEL_INNER.
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35931>
this is more explicit than vec2's and hence has fewer footguns. in particular
it's easier to handle with preambles in a sane way.
modelled on what ir3 does.
there's probably room for more clean up but for now this unblocks what I want to
do.
stats don't seem concerning.
Totals from 692 (1.29% of 53701) affected shaders:
MaxWaves: 441920 -> 442112 (+0.04%)
Instrs: 1588748 -> 1589304 (+0.03%); split: -0.05%, +0.08%
CodeSize: 11487976 -> 11491620 (+0.03%); split: -0.04%, +0.07%
ALU: 1234867 -> 1235407 (+0.04%); split: -0.06%, +0.10%
FSCIB: 1234707 -> 1235249 (+0.04%); split: -0.06%, +0.10%
IC: 380514 -> 380518 (+0.00%)
GPRs: 117292 -> 117332 (+0.03%); split: -0.08%, +0.11%
Preamble instrs: 314064 -> 313948 (-0.04%); split: -0.05%, +0.01%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35949>
it is sometimes useful to turn lowered bindless intrinsics into bound or vice
versa, and it is annoying to do so without this helper, so generalize the
helper.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Job Noorman <job@noorman.info>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35949>
Class represents an indexed "ideal" register class, where non-general classes
only allow defs that choose that class in the def_size callback.
nir_opt_preamble will try to assign specialized classes where possible, falling
back to the general class once the special-purpose classes are exhausted.
AGX will use this mechanism to promote bindless texture handles to bound texture
registers where possible, falling back to pushing the handle as a uniform where
not possible. Supporting multiple classes in nir_opt_preamble allows this
multi-level hoisting to work in a single nir_opt_preamble call with proper
global behaviour.
Add this concept to nir_opt_preamble so we can use it in AGX later in this MR.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Job Noorman <job@noorman.info>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35949>
This reduces GLSL compile times with the gallium noop driver by 0.6%.
This might decrease register usage and do less code reordering because
nir_lower_io_vars_to_temporaries is no longer called for inputs, which
moved most input loads to the top.
radeonsi+ACO shader-db results are noise.
More uniforms are identified as inlinable.
TOTALS FROM ALL SHADERS (58138):
VGPRs: 2152680 -> 2158032 (0.25 %)
Code Size: 71008908 -> 71064812 (0.08 %) bytes
Max Waves: 916943 -> 916924 (-0.00 %)
Inline Uniforms: 6395 -> 6414 (0.30 %)
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36018>
This can be used to move input loads to top after we stop using
nir_lower_io_vars_to_temporaries that does it unconditionally.
It's more flexible than what nir_lower_io_vars_to_temporaries was doing,
and can be extended to handle any instructions.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36018>
This is a partial replacement for nir_lower_io_vars_to_temporaries.
It supports all input and output loads. It doesn't handle stores.
The motivation is to improve compile times.
The main differences compared to nir_lower_io_vars_to_temporaries are:
- it only lowers indirect loads to temps and doesn't touch direct loads
which improves compile times and removes the need for nir_lower_vars_to_ssa
afterward because indirect temp access can't be lowered to SSA
- it doesn't move all input loads to the top; it only moves those input
loads to the top whose indirect loads are lowered (which improves
register usage because direct loads are not moved)
- it doesn't have to deal with complexities of variables
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36018>
It doesn't do anything since IO variables are lowered to intrinsics,
which simplifies and eliminates a lot of variable-specific stuff
like declared but dead builtin varyings and unused components
of builtin varying arrays.
This reduces GLSL compile times by 2.4% with the gallium noop driver.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Timothy Arceri <tarceri@itqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36023>
nir_opt_varyings reduces the number of varyings. Check against limits after
that, so that old and limited GPUs don't fail linking when nir_opt_varyings
is able to reduce varyings to or below the limit.
The previous code only checked FS inputs, which is glaringly obvious
from the removed var_counts_against_varying_limit function.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Timothy Arceri <tarceri@itqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36023>
The fmul+fadd -> fma rules in nir_opt_algebraic are marked imprecise,
because they are a contraction. However, they respect signed zero/Inf/NaN rules.
As such, it is legal to do this fusion with shader float controls as long as the
exact bit is not set (mapping to SPIR-V NoContract).
Unfortunately, NIR's imprecise rules do not distinguish between contraction
issues versus float special case issues, forcing nir_search to skip all
imprecise rules when any shader float control modes are used. This notably
affects DXVK, which sets shader float controls to get D3D11 float behaviour and
hence loses FMA fusing.
Therefore, we plumb in the exact bit to express NoContract independent of the
float controls, and weaken the requirement for fma fusion to allowable
contraction. For fma splitting, it's a similar issue, as inexact GLSL fma in
SPIR-V is just a multiply add that we're allowed to contract rather than the
real deal.
Drivers that use their own FMA fusing passes (notably, Intel and AMD) are
unaffected, but DXVK-capable drivers using fuse_ffma should like this. Results
on hk shown:
Totals from 2194 (4.06% of 54019) affected shaders:
MaxWaves: 2174272 -> 2175936 (+0.08%); split: +0.08%, -0.01%
Instrs: 1173283 -> 1131494 (-3.56%); split: -3.57%, +0.01%
CodeSize: 8568168 -> 8381724 (-2.18%); split: -2.18%, +0.01%
Spills: 1094 -> 747 (-31.72%)
Fills: 988 -> 681 (-31.07%)
Scratch: 4444 -> 3820 (-14.04%)
ALU: 953032 -> 913149 (-4.18%); split: -4.19%, +0.01%
FSCIB: 953032 -> 913149 (-4.18%); split: -4.19%, +0.01%
IC: 215398 -> 215274 (-0.06%)
GPRs: 139865 -> 139032 (-0.60%); split: -1.56%, +0.96%
Uniforms: 414886 -> 414466 (-0.10%); split: -0.14%, +0.04%
Preamble instrs: 646398 -> 644017 (-0.37%); split: -0.43%, +0.07%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35989>
Similar to nir_lower_alu_width(), the callback can return the
desired number of components for a phi, or 0 for no lowering.
The previous behavior of nir_lower_phis_to_scalar() with lower_all=true
can be elicited via nir_lower_all_phis_to_scalar() while the previous
behavior with lower_all=false now corresponds to nir_lower_phis_to_scalar()
with NULL callback.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Mel Henning <mhenning@darkrefraction.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35783>
This check causes unnecessary overhead and can be replaced by simply
checking whether a phi_src is from a loop continue block.
Except for rare edge cases, the result will be the same.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35783>
It's done later in nir_lower_io_passes only for shader stages not
supporting indirect access.
Unfortunately we have add a hack into nir_lower_io_passes to get rid of
output loads. A later commit will remove it.
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35945>
