Intel platforms will soon implement both bfi and bitfield_select. bfi is
more efficient for bitfield_insert.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37186>
The eliminated SENDs are from a single app that has a bunch of
fragment shaders with a sequence like:
con 32 %495 = fmul! %203.i, %1 (0.000000)
con 32 %496 = ffma! %203.j, %1 (0.000000), %495
con 32 %497 = ffma! %203.k, %1 (0.000000), %496
con 32 %498 = ffma! %203.l, %1 (0.000000), %497
con 32 %499 = @load_reloc_const_intel (param_idx=1, base=0)
con 32 %500 = @load_reloc_const_intel (param_idx=0, base=0)
con 32 %501 = f2u32 %498
con 32 %502 = umin %501, %172 (0x4)
con 32 %503 = ishl %502, %172 (0x4)
con 32 %504 = load_const (0x00000040 = 64)
con 32 %505 = umin %503, %504 (0x40)
con 32 %506 = iadd %500, %505
The `f2u` is replaced with 0, and that makes the `ffma` dot-product
sequence be unused. Since it is unused, most of the preceeding block
gets eliminated. A lot of instructions after the `f2u` are also
eliminated by other algebraic optimizations. Most importantly, %203 is
the result of a `load_ubo_uniform_block_intel` that is eliminated.
No shader-db changes on any Intel platform.
fossil-db:
All Intel platforms had similar results. (Lunar Lake shown)
Totals:
Instrs: 919895603 -> 919804051 (-0.01%); split: -0.01%, +0.00%
Send messages: 40892036 -> 40887569 (-0.01%)
Cycle count: 99176770712 -> 99174971806 (-0.00%); split: -0.00%, +0.00%
Max live registers: 190030365 -> 190030367 (+0.00%)
Max dispatch width: 47415040 -> 47415024 (-0.00%)
Non SSA regs after NIR: 228872538 -> 228863608 (-0.00%); split: -0.00%, +0.00%
Totals from 2234 (0.11% of 1955134) affected shaders:
Instrs: 1989743 -> 1898191 (-4.60%); split: -4.60%, +0.00%
Send messages: 44179 -> 39712 (-10.11%)
Cycle count: 25416114 -> 23617208 (-7.08%); split: -7.08%, +0.00%
Max live registers: 367357 -> 367359 (+0.00%)
Max dispatch width: 39184 -> 39168 (-0.04%)
Non SSA regs after NIR: 471173 -> 462243 (-1.90%); split: -1.90%, +0.00%
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37186>
If the source -1.0 < x < 0.0, simply removing the ftrun will introduce
undefined behavior. By chance of how at least Intel and NVIDIA GPUs
implement f2u, this has Just Worked.
No shader-db changes on any Intel platform.
fossil-db:
Lunar Lake
Totals:
Instrs: 913264354 -> 913264366 (+0.00%)
Cycle count: 104953995530 -> 104953996854 (+0.00%)
Max live registers: 189266026 -> 189266058 (+0.00%)
Non SSA regs after NIR: 227779417 -> 227779369 (-0.00%)
Totals from 24 (0.00% of 1984794) affected shaders:
Instrs: 4669 -> 4681 (+0.26%)
Cycle count: 50610 -> 51934 (+2.62%)
Max live registers: 1222 -> 1254 (+2.62%)
Non SSA regs after NIR: 1174 -> 1126 (-4.09%)
Meteor Lake, DG2, Tiger Lake, and Ice Lake had similar results. (Meteor Lake shown)
Totals:
Instrs: 1001288026 -> 1001288038 (+0.00%)
Cycle count: 92813392671 -> 92813392791 (+0.00%)
Max live registers: 121935383 -> 121935399 (+0.00%)
Max dispatch width: 19949928 -> 19949912 (-0.00%)
Totals from 2 (0.00% of 2284670) affected shaders:
Instrs: 1380 -> 1392 (+0.87%)
Cycle count: 18940 -> 19060 (+0.63%)
Max live registers: 136 -> 152 (+11.76%)
Max dispatch width: 32 -> 16 (-50.00%)
No fossil-db changes on Skylake.
Suggested-by: Georg Lehmann
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37186>
There are no shader-db changes on ELK platforms because those platforms
don't support 8- or 16-bit integer types.
v2: Restrict patterns generated such that the integer limits are exactly
representable in the specified floating point format. With the exception
of the value 0, this requires that float_sz > int_sz. This had no impact
on shader-db or fossil-db on any Intel platform. Noticed by Georg.
v3: Add a missing is_a_number.
shader-db:
All Intel platforms had similar results. (Lunar Lake shown)
total cycles in shared programs: 889936056 -> 889934082 (<.01%)
cycles in affected programs: 65806 -> 63832 (-3.00%)
helped: 2 / HURT: 0
fossil-db:
Lunar Lake
Totals:
Instrs: 233284796 -> 233282917 (-0.00%); split: -0.00%, +0.00%
Cycle count: 32756399804 -> 32754972188 (-0.00%); split: -0.01%, +0.00%
Spill count: 519861 -> 519813 (-0.01%)
Fill count: 663650 -> 663626 (-0.00%); split: -0.01%, +0.01%
Max live registers: 71738626 -> 71738696 (+0.00%)
Non SSA regs after NIR: 67837902 -> 67837648 (-0.00%)
Totals from 1236 (0.16% of 790723) affected shaders:
Instrs: 2134504 -> 2132625 (-0.09%); split: -0.09%, +0.01%
Cycle count: 604922278 -> 603494662 (-0.24%); split: -0.48%, +0.25%
Spill count: 16509 -> 16461 (-0.29%)
Fill count: 32760 -> 32736 (-0.07%); split: -0.22%, +0.15%
Max live registers: 250112 -> 250182 (+0.03%)
Non SSA regs after NIR: 302368 -> 302114 (-0.08%)
Meteor Lake, DG2, and Tiger Lake had similar results. (Meteor Lake shown)
Totals:
Instrs: 264095370 -> 264094056 (-0.00%); split: -0.00%, +0.00%
Cycle count: 26554146277 -> 26553027268 (-0.00%); split: -0.01%, +0.01%
Spill count: 530603 -> 530615 (+0.00%)
Fill count: 613231 -> 613273 (+0.01%)
Max live registers: 46559041 -> 46559087 (+0.00%)
Totals from 1237 (0.14% of 905547) affected shaders:
Instrs: 2262517 -> 2261203 (-0.06%); split: -0.07%, +0.01%
Cycle count: 518219799 -> 517100790 (-0.22%); split: -0.59%, +0.37%
Spill count: 17518 -> 17530 (+0.07%)
Fill count: 32273 -> 32315 (+0.13%)
Max live registers: 128360 -> 128406 (+0.04%)
Ice Lake and Skylake had similar results. (Ice Lake shown)
Totals:
Instrs: 269849640 -> 269848198 (-0.00%); split: -0.00%, +0.00%
Cycle count: 26718329643 -> 26718289020 (-0.00%); split: -0.00%, +0.00%
Max live registers: 46878430 -> 46878462 (+0.00%)
Totals from 1233 (0.14% of 905427) affected shaders:
Instrs: 2324225 -> 2322783 (-0.06%); split: -0.06%, +0.00%
Cycle count: 531467501 -> 531426878 (-0.01%); split: -0.11%, +0.10%
Max live registers: 130782 -> 130814 (+0.02%)
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37186>
v2: Add a comment around has_f2[ui]_sat explaining which opcodes it
enables. Suggested by Georg. Cast u_uintN_max and friends to double in
nir_opcodes.py. This ensures that an exact conversion is made.
Eliminate duplicate conversions from half float to double. Both noticed
by Georg.
v3: Apply "NaN should be zero" fix suggested by Georg.
Co-authored-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37186>
The Intel Gfx12.x generation of GPU has an architecture feature called
EU fusion in which 2 subgroups run lock step. A typical case where
this happens is a compute shader with 1x1x1 local workgroup size and a
dispatch command of 2x1x1. In that case 2 threads will be run in lock
step for each of the workgroup.
This has been the sources of some troubles in the backend because one
subgroup can run with all lanes disabled, requiring care for SEND
messages using the NoMask flag (execution regardless of the lane mask).
We found out that other things are happening when 2 subgroups run
together :
- the HW will use the surface/sampler handle from only one subgroup
- the HW will use the sampler header from only one subgroup
So one of the fused subgroup can access the wrong surface/sampler if
the value is different between the 2 subgroups and that can happen
even with subgroup uniform values.
Fortunately we can flag SEND instructions to disable the fusion
behavior (most likely at a performance cost).
This change introduce a new divergence mode that tries to compute
things divergent between subgroups so that we can flag instructions
accordingly.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37394>
nir_recompute_io_bases will modify i/o intrinsics, which is not the
expected behaviour when the keep_intrinsics flag is set.
Fixes: 83aecc8f3f ("mesa/st, nir: commonize unlower_io_to_vars pass")
Signed-off-by: Simon Perretta <simon.perretta@imgtec.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37725>
Sometimes we need to select between ishr/ushr based some condition; this
builder makes this less verbose.
Signed-off-by: Job Noorman <jnoorman@igalia.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37793>
When you're trying to figure out what shader some NIR pass broke, use
nir_shader_bisect_select() to decide between NIR pass behaviors, and then
nir_shader_bisect.py will help you automatically bisect down to which
source_blake3 is at fault. Once it's identified, it prints you a C call
you can use for selecting that shader specifically, which you can use for
continuing on in your debugging.
On a test I was looking at, this took 10 steps to bisect 134 shaders down
to the source_blake3 of the NIR shader in question.
This idea is heavily lifted from Job Noorman's ir3_shader_bisect.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37468>
Usage of '--outdir' argument in python scripts makes it very
complicated for tools like ninja-to-soong to generate the Android
equivalent build file.
This is because the option is less clear on what will be generated.
Instead, change it for '--out' where we give the full path of the file
to generate. This has the good point of deduplicating the locations of
the file name to have it only in 'meson.build'.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37741>
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36602>
This is so that passes and backends can tell if a coherent load/store is
atomic or not, instead of having to assume it could be either.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36602>
These instructions don't need a sampler.
This doesn't fix anything now because this helper isn't unused yet, but
it will help for descriptor heap.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37720>
For some reason we missed adding this. This prevents some asserts
from triggering when I call divergence analysis at certain points
in an upcoming patch.
Cc: mesa-stable
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36750>
- Implement load_interpolated_input and friends.
- Optimize load_barycentric_* cases that can be simplified.
- Initial support for non-standard sample locations.
Signed-off-by: Simon Perretta <simon.perretta@imgtec.com>
Acked-by: Frank Binns <frank.binns@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37540>
Using NaN doesn't usually allow any extra optimizations, and 0 is an inline
constant on AMD hw where this opcode is used with undef for fragment shader
exports.
Foz-DB GFX1201:
Totals from 889 (1.11% of 80287) affected shaders:
Instrs: 1676365 -> 1676348 (-0.00%)
CodeSize: 8827040 -> 8821760 (-0.06%)
Latency: 13346728 -> 13346699 (-0.00%)
InvThroughput: 1799283 -> 1799262 (-0.00%)
Copies: 108125 -> 108102 (-0.02%)
VALU: 974875 -> 974852 (-0.00%)
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37552>
This broke after divergence became metadata because the divergence
analysis pass does not support all instructions.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35069>
This would unconditionally invalid all metadata except
nir_metadata_control_flow and then invalidate that if opt_if_safe_cf_list
and opt_if_regs_cf_list made no progress.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Marek Olšák <maraeo@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35069>
Operands of an addition will be promoted to int making the a+b<a
kind of checks ineffective.
Use u_overflow.h helpers to perform the check correctly.
The commit would be simpler if it used __typeof__ like so:
util_add_check_overflow(__typeof__(src0), src0, src1)
But typeof only became a standard in C23 so this commit instead extends
nir_opcodes a bit to allow opcodes that need the dest_type to get it.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Dylan Baker <dylan.c.baker@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37331>
