I had a single function for "does this do float input unpacking" with two
major flaws: It was missing the most common thing to try to copy propagate
a f32 input nunpack to (the VFPACK to an FP16 render target) along with
several other ALU ops, and also would try to propagate an f32 unpack into
a VFMUL which only does f16 unpacks.
instructions in affected programs: 659232 -> 655895 (-0.51%)
uniforms in affected programs: 132613 -> 135336 (2.05%)
and a couple of programs increase their thread counts.
The uniforms hit appears to be a pattern in generated code of doing (-a >=
a) comparisons, which when a is abs(b) can result in the abs instruction
being copy propagated once but not fully DCEed.
If you only bound rt 1+, we'd still emit a write to the rt0 that isn't
present (noticed while debugging an
ext_framebuffer_multisample-alpha-to-coverage-no-draw-buffer-zero
regression in another change).
This was copy-and-paste fail, that oddly showed up in the CTS's
reinterprets of r32f, rgba8, and srgba8 to rgba8i, but not r32ui and r32i
to rgba8i or reinterprets to other signed int formats.
Fixes: 6281f26f06 ("v3d: Add support for shader_image_load_store.")
We get a payload for the ivec3 workgroup and an int local invocation
index, and we use the core lowering to turn into the global invocation id
and the local invocation id ivec3s.
This is only exposed on V3D 4.1+, because we didn't have the TMU write
operations for images on 3.3 (To do GLES 3.1 there, you have to lower it
to SSBO load/stores, which is a problem to solve later).
We've been relying on linking splitting up our varying matrices into
separate vectors, but with SSO that doesn't happen. Supporting matrix
inputs isn't too hard, though.
V3D returns the texels in a different order in the resulting vec4 from
what GLSL wants, so we need to put in a swizzle. Fixes
dEQP-GLES31.functional.texture.gather.basic.2d.rgba8.base_level.level_1
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Even without any clever optimization on the unpack operations, this gives
us a useful value for the channels read field, which we can use to avoid
ldtmu instructions to the no-op register.
instructions in affected programs: 890712 -> 881974 (-0.98%)
We can pull a whole vector in a single indirect load. This saves a bunch
of round-trips to the TMU, instructions for setting up multiple loads,
references to the UBO base in the uniforms, and apparently manages to
reduce register pressure as well.
instructions in affected programs: 3086665 -> 2454967 (-20.47%)
uniforms in affected programs: 919581 -> 721039 (-21.59%)
threads in affected programs: 1710 -> 3420 (100.00%)
spills in affected programs: 596 -> 522 (-12.42%)
fills in affected programs: 680 -> 562 (-17.35%)
Improves 3dmmes performance by 2.29312% +/- 0.139825% (n=5)
In the process of adding support for SSBOs and CS shared vars, I ended up
needing a helper function for doing TMU general ops. This helper can be
that starting point, and saves us a bunch of round-trips to the TMU by
loading a vector at a time.
Before, I had per-stage entryoints with some helpers shared between them.
As I extended for compute shaders and shader-db, it turned out that the
other common code in the middle wanted to be shared too.
Loops will be trickier, since we need some analysis to figure out if the
breaks/continues inside are uniform. Until we get that in NIR, this gets
us some quick wins.
total instructions in shared programs: 6192844 -> 6174162 (-0.30%)
instructions in affected programs: 487781 -> 469099 (-3.83%)
I wanted to reuse the comparison stuff for nir_ifs, but for that I just
want the flags and no destination value. Splitting the conditions from
the destinations ended up cleaning the existing code up, anyway.
We can just look at the MSF flags -- if they're unset, then we're
definitely in a helper invocation. Fixes
dEQP-GLES31.functional.shaders.helper_invocation.* with GLES3.1 enabled.
This is what the GLSL ES 310 spec tells us to do, but apparently the
"gather mode" flag doesn't imply it in the HW. Fixes
dEQP-GLES31.functional.texture.gather.basic.2d.rgba8.filter_mode.min_nearest_mipmap_linear_mag_linear
The greedy comparison folding in bcsel means that we may have left the
original bool-generating NIR ALU instruction dead, but DCE wasn't
eliminating the VIR code for it because of the flags updates.
total instructions in shared programs: 5186024 -> 5100894 (-1.64%)
instructions in affected programs: 1448695 -> 1363565 (-5.88%)
This allows the original shader-db project's run.c runner to parse things
easily, and is probably a good thing to have for GL_ARB_debug_output in
general. I formatted it more like Intel's so I can mostly reuse their
report script.
On some GPUs, especially older Intel GPUs, some math instructions are
very expensive. On those architectures, don't reduce flow control to a
csel if one of the branches contains one of these expensive math
instructions.
This prevents a bunch of cycle count regressions on pre-Gen6 platforms
with a later patch (intel/compiler: More peephole select for pre-Gen6).
v2: Remove stray #if block. Noticed by Thomas.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
That flow control may be trying to avoid invalid loads. On at least
some platforms, those loads can also be expensive.
No shader-db changes on any Intel platform (even with the later patch
"intel/compiler: More peephole select").
v2: Add a 'indirect_load_ok' flag to nir_opt_peephole_select. Suggested
by Rob. See also the big comment in src/intel/compiler/brw_nir.c.
v3: Use nir_deref_instr_has_indirect instead of deref_has_indirect (from
nir_lower_io_arrays_to_elements.c).
v4: Fix inverted condition in brw_nir.c. Noticed by Lionel.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
We also enable it in all of the NIR drivers.
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Tested-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
This is a squash of a bunch of individual changes:
nir/builder: Generate 32-bit bool opcodes transparently
nir/algebraic: Remap Boolean opcodes to the 32-bit variant
Use 32-bit opcodes in the NIR producers and optimizations
Generated with a little hand-editing and the following sed commands:
sed -i 's/nir_op_ball_fequal/nir_op_b32all_fequal/g' **/*.c
sed -i 's/nir_op_bany_fnequal/nir_op_b32any_fnequal/g' **/*.c
sed -i 's/nir_op_ball_iequal/nir_op_b32all_iequal/g' **/*.c
sed -i 's/nir_op_bany_inequal/nir_op_b32any_inequal/g' **/*.c
sed -i 's/nir_op_\([fiu]lt\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]ge\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]ne\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]eq\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fi]\)ne32g/nir_op_\1neg/g' **/*.c
sed -i 's/nir_op_bcsel/nir_op_b32csel/g' **/*.c
Use 32-bit opcodes in the NIR back-ends
Generated with a little hand-editing and the following sed commands:
sed -i 's/nir_op_ball_fequal/nir_op_b32all_fequal/g' **/*.c
sed -i 's/nir_op_bany_fnequal/nir_op_b32any_fnequal/g' **/*.c
sed -i 's/nir_op_ball_iequal/nir_op_b32all_iequal/g' **/*.c
sed -i 's/nir_op_bany_inequal/nir_op_b32any_inequal/g' **/*.c
sed -i 's/nir_op_\([fiu]lt\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]ge\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]ne\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fiu]eq\)/nir_op_\132/g' **/*.c
sed -i 's/nir_op_\([fi]\)ne32g/nir_op_\1neg/g' **/*.c
sed -i 's/nir_op_bcsel/nir_op_b32csel/g' **/*.c
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Tested-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
These have all been floating in my head, and while I've thought about
encoding them in issues on gitlab once they're enabled, they also make
sense to just have in the area of the code you'll need to work in.
