Intel hardware supports 8-bit arithmetic but it's tricky and annoying:
- Byte operations don't actually execute with a byte type. The
execution type for byte operations is actually word. (I don't know
if this has implications for the HW implementation. Probably?)
- Destinations are required to be strided out to at least the
execution type size. This means that B-type operations always have
a stride of at least 2. This means wreaks havoc on the back-end in
multiple ways.
- Thanks to the strided destination, we don't actually save register
space by storing things in bytes. We could, in theory, interleave
two byte values into a single 2B-strided register but that's both a
pain for RA and would lead to piles of false dependencies pre-Gen12
and on Gen12+, we'd need some significant improvements to the SWSB
pass.
- Also thanks to the strided destination, all byte writes are treated
as partial writes by the back-end and we don't know how to copy-prop
them.
- On Gen11, they added a new hardware restriction that byte types
aren't allowed in the 2nd and 3rd sources of instructions. This
means that we have to emit B->W conversions all over to resolve
things. If we emit said conversions in NIR, instead, there's a
chance NIR can get rid of some of them for us.
We can get rid of a lot of this pain by just asking NIR to get rid of
8-bit arithmetic for us. It may lead to a few more conversions in some
cases but having back-end copy-prop actually work is probably a bigger
bonus. There is still a bit we have to handle in the back-end. In
particular, basic MOVs and conversions because 8-bit load/store ops
still require 8-bit types.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7482>
We can't really support these directly on any platform. May as well let
NIR lower them. The NIR lowering is potentially one more instruction
for scan/reduce ops thanks to not being able to do the B->W conversion
as part of SEL_EXEC. For imax/imin exclusive scan, it's yet another
instruction thanks to the extra imax/imin NIR has to insert to deal with
the fact that the first live channel will contain the identity value
which, when signed, will cast wrong. However, it does let us drop some
complexity from our back-end so it's probably worth it.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/7482>
Instead, we do a limited indirect deref lowering and then use
nir_lower_vars_to_explicit_types and nir_lower_explicit_io to lower it
as if it were SSBO or global memory access. Among other things, this
should enable pointer arithmetic on local variables. Fun!
The only shader-db change from this change on ICL was a few tiny cycle
count changes in 7 Aztec Ruins compute shaders.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5909>
Instead of always lowering everything, we add a threshold such that if
the total indirected array size (AoA size) is above that threshold, it
won't lower. It's assumed that the driver will sort things out somehow
by, for instance, lowering to scratch.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5909>
Since everything flows through NIR and we're doing all of our indirect
deref lowering there now, there's no reason to keep making those
decisions in brw_compiler and stuffing them in the GLSL compiler
structs.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5909>
If no options are provided, existing intrinsics are used.
If the lowering pass indicates there should be offsets used for global
invocation ID or work group ID, then those instructions are lowered to
include the offset.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5891>
This manages to make some extra vec operations that would turn into movs
go away.
brw shader-db:
total instructions in shared programs: 3895037 -> 3893221 (-0.05%)
total cycles in shared programs: 113832759 -> 113792154 (-0.04%)
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6050>
This allows us to do API specific checks before removing variable
without filling nir_remove_dead_variables() with API specific code.
In the following patches we will use this to support the removal
of dead uniforms in GLSL.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4797>
I think we're unanimous in wanting to drop nir_lower_to_source_mods.
It's a bit of complexity to handle in the backend, but perhaps more
importantly, would be even more complexity to handle in nir_search.
And, it turns out that since we made other compiler improvements in the
last few years, they no longer appear to buy us anything of value.
Summarizing the results from shader-db from this patch:
- Icelake (scalar mode)
Instruction counts:
- 411 helped, 598 hurt (out of 139,470 shaders)
- 99.2% of shaders remain unaffected. The average increase in
instruction count in hurt programs is 1.78 instructions.
- total instructions in shared programs: 17214951 -> 17215206 (<.01%)
- instructions in affected programs: 1143879 -> 1144134 (0.02%)
Cycles:
- 1042 helped, 1357 hurt
- total cycles in shared programs: 365613294 -> 365882263 (0.07%)
- cycles in affected programs: 138155497 -> 138424466 (0.19%)
- Haswell (both scalar and vector modes)
Instruction counts:
- 73 helped, 1680 hurt (out of 139,470 shaders)
- 98.7% of shaders remain unaffected. The average increase in
instruction count in hurt programs is 1.9 instructions.
- total instructions in shared programs: 14199527 -> 14202262 (0.02%)
- instructions in affected programs: 446499 -> 449234 (0.61%)
Cycles:
- 5253 helped, 5559 hurt
- total cycles in shared programs: 359996545 -> 360038731 (0.01%)
- cycles in affected programs: 155897127 -> 155939313 (0.03%)
Given that ~99% of shader-db remains unaffected, and the affected
programs are hurt by about 1-2 instructions - which are all cheap
ALU instructions - this is unlikely to be measurable in terms of
any real performance impact that would affect users.
So, drop them and simplify the backend, and hopefully enable other
future simplifications in NIR.
Reviewed-by: Eric Anholt <eric@anholt.net> [v1]
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4616>
This commit enables the I/O vectorization pass that was originally
written for ACO for Intel drivers. We enable it for UBOs, SSBOs, global
memory, and SLM. We only enable vectorization for the scalar back-end
because it vec4 makes certain alignment assumptions.
Shader-db results with iris on ICL:
total instructions in shared programs: 16077927 -> 16068236 (-0.06%)
instructions in affected programs: 199839 -> 190148 (-4.85%)
helped: 324
HURT: 0
helped stats (abs) min: 2 max: 458 x̄: 29.91 x̃: 4
helped stats (rel) min: 0.11% max: 38.94% x̄: 4.32% x̃: 1.64%
95% mean confidence interval for instructions value: -37.02 -22.80
95% mean confidence interval for instructions %-change: -5.07% -3.58%
Instructions are helped.
total cycles in shared programs: 336806135 -> 336151501 (-0.19%)
cycles in affected programs: 16009735 -> 15355101 (-4.09%)
helped: 458
HURT: 154
helped stats (abs) min: 1 max: 77812 x̄: 1542.50 x̃: 75
helped stats (rel) min: <.01% max: 34.46% x̄: 5.16% x̃: 2.01%
HURT stats (abs) min: 1 max: 22800 x̄: 336.55 x̃: 20
HURT stats (rel) min: <.01% max: 17.11% x̄: 2.12% x̃: 1.00%
95% mean confidence interval for cycles value: -1596.83 -542.49
95% mean confidence interval for cycles %-change: -3.83% -2.82%
Cycles are helped.
total sends in shared programs: 814177 -> 809049 (-0.63%)
sends in affected programs: 15422 -> 10294 (-33.25%)
helped: 324
HURT: 0
helped stats (abs) min: 1 max: 256 x̄: 15.83 x̃: 2
helped stats (rel) min: 1.33% max: 67.90% x̄: 21.21% x̃: 15.38%
95% mean confidence interval for sends value: -19.67 -11.98
95% mean confidence interval for sends %-change: -23.03% -19.39%
Sends are helped.
LOST: 7
GAINED: 2
Most of the helped shaders were in the following titles:
- Doom
- Deus Ex: Mankind Divided
- Aztec Ruins
- Shadow of Mordor
- DiRT Showdown
- Tomb Raider (Rise, I think)
Five of the lost programs are SIMD16 shaders we lost from dirt showdown.
The other two are compute shaders in Aztec Ruins which switched from
SIMD8 to SIMD16.
Vulkan pipeline-db stats on ICL:
Instructions in all programs: 296780486 -> 293493363 (-1.1%)
Loops in all programs: 149669 -> 149669 (+0.0%)
Cycles in all programs: 90999206722 -> 88513844563 (-2.7%)
Spills in all programs: 1710217 -> 1730691 (+1.2%)
Fills in all programs: 1931235 -> 1958138 (+1.4%)
By far the most help was in the Tomb Raider games. A couple of Batman
games with DXVK were also helped. In Shadow of the Tomb Raider:
Instructions in all programs: 41614336 -> 39408023 (-5.3%)
Loops in all programs: 32200 -> 32200 (+0.0%)
Cycles in all programs: 1875498485 -> 1667034831 (-11.1%)
Spills in all programs: 196307 -> 214945 (+9.5%)
Fills in all programs: 282736 -> 307113 (+8.6%)
Benchmarks of real games I've done on this patch:
- Rise of the Tomb Raider: +3%
- Shadow of the Tomb Raider: +10%
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4367>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4367>
This change incurs a small amount of hurt now, but it enables a lot of
benefit on vec4 shaders on the next commit. nir_opt_algebraic_late
converts dph, dot3, etc. to dhp_replicated, dot_replicated3, etc. In
the process, it introduces extra moves. If the original NIR contained
vec1 32 ssa_45 = fdot4 ssa_51, ssa_44
vec1 32 ssa_46 = fneg ssa_45
nir_opt_algebraic_late will produce
vec4 32 ssa_18 = fdot_replicated4 ssa_1, ssa_15
vec1 32 ssa_19 = mov ssa_18.x
vec1 32 ssa_17 = fneg ssa_19
The algebraic pass added in the next commit can't see through the move
to know that the fneg applies to a fdot_replicated4.
Haswell, Ivy Bridge, and Sandybridge had similar results. (Haswell shown)
total cycles in shared programs: 187077604 -> 187079858 (<.01%)
cycles in affected programs: 350132 -> 352386 (0.64%)
helped: 174
HURT: 194
helped stats (abs) min: 2 max: 124 x̄: 23.60 x̃: 16
helped stats (rel) min: 0.12% max: 15.88% x̄: 4.98% x̃: 3.86%
HURT stats (abs) min: 2 max: 164 x̄: 32.78 x̃: 16
HURT stats (rel) min: 0.17% max: 22.82% x̄: 6.46% x̃: 0.86%
95% mean confidence interval for cycles value: 2.04 10.21
95% mean confidence interval for cycles %-change: 0.17% 1.93%
Cycles are HURT.
No shader-db changes on any other Intel platform.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/1359>
This will avoid generating multiple identical fences in a row.
For Gen11+ we have multiple types of fences (affecting different
variable modes), but is still better to combine them in a single
scoped barrier so that the translation to backend IR have the option
of dispatching both fences in parallel.
This will clean up redundant barriers from various
dEQP-VK.memory_model.* tests.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3224>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3224>
../src/intel/compiler/brw_nir.c:979:40: runtime error: left shift of 1 by 31 places cannot be represented in type 'int'
#0 0x7f78f590d10b in brw_nir_apply_sampler_key ../src/intel/compiler/brw_nir.c:979
#1 0x7f78f590e07b in brw_nir_apply_key ../src/intel/compiler/brw_nir.c:1057
#2 0x7f78f5754b45 in brw_compile_fs ../src/intel/compiler/brw_fs.cpp:8347
#3 0x7f78f255c8e4 in brw_codegen_wm_prog ../src/mesa/drivers/dri/i965/brw_wm.c:123
#4 0x7f78f2565571 in brw_fs_precompile ../src/mesa/drivers/dri/i965/brw_wm.c:608
#5 0x7f78f24edd2c in brw_shader_precompile ../src/mesa/drivers/dri/i965/brw_link.cpp:56
#6 0x7f78f24f3af8 in brw_link_shader ../src/mesa/drivers/dri/i965/brw_link.cpp:381
#7 0x7f78f39a302a in _mesa_glsl_link_shader ../src/mesa/program/ir_to_mesa.cpp:3119
#8 0x7f78f3a43826 in create_new_program ../src/mesa/main/ff_fragment_shader.cpp:1133
#9 0x7f78f3a43d00 in _mesa_get_fixed_func_fragment_program ../src/mesa/main/ff_fragment_shader.cpp:1163
#10 0x7f78f325ddcd in update_program ../src/mesa/main/state.c:134
#11 0x7f78f325fe64 in _mesa_update_state_locked ../src/mesa/main/state.c:360
#12 0x7f78f32600f1 in _mesa_update_state ../src/mesa/main/state.c:394
#13 0x7f78f2b3e587 in clear ../src/mesa/main/clear.c:169
#14 0x7f78f2b3e587 in _mesa_Clear ../src/mesa/main/clear.c:242
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3825>
Gen12 does not support RENDER_SURFACE_STATE::SurfaceArray = true &&
RENDER_SURFACE_STATE::Depth = 0. SurfaceArray can only be set to true
if Depth >= 1.
We workaround this limitation by adding the max(value, 1) snippet in
the shaders on the 3 components for texture array sizes.
Tested on Gen9 with the following Vulkan CTS tests :
dEQP-VK.image.image_size.2d_array.*
v2: Drop debug print (Tapani)
Switch to GEN:BUG instead of Wa_
v3: Fix dEQP-VK.image.image_size.1d_array.* cases (Lionel)
v4: Fix dEQP-VK.glsl.texture_functions.query.texturesize.* cases
(Missing tex_op handling) (Lionel)
v5: Missing break statement (Lionel)
v6: Fixup comment (Tapani)
v7: Fixup comment again (Tapani)
v8: Don't use sample_dim as index (Jason)
Rename pass
Simplify control flow
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com> (v7)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3362>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3362>
This is required for the subgroupBroadcastDynamicId feature that was
added in Vulkan 1.2.
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
On gen7 and earlier the scratch space size is limited to 12kB.
By enabling this optimization we may easily exceed this limit
without having any fallback.
arb_compute_shader/linker/bug-93840.shader_test crashes with
this lowering on IVB due to exceeding scratch size limit.
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2092
Fixes: 69244fc7
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Shader-db results on Kaby Lake:
total instructions in shared programs: 14929212 -> 14880028 (-0.33%)
instructions in affected programs: 72428 -> 23244 (-67.91%)
helped: 6
HURT: 2
helped stats (abs) min: 2165 max: 15981 x̄: 8590.00 x̃: 7624
helped stats (rel) min: 56.06% max: 74.52% x̄: 67.55% x̃: 72.08%
HURT stats (abs) min: 1178 max: 1178 x̄: 1178.00 x̃: 1178
HURT stats (rel) min: 350.60% max: 361.35% x̄: 355.97% x̃: 355.97%
95% mean confidence interval for instructions value: -11947.03 -348.97
95% mean confidence interval for instructions %-change: -125.72% 202.37%
Inconclusive result (%-change mean confidence interval includes 0).
total cycles in shared programs: 368585300 -> 342557344 (-7.06%)
cycles in affected programs: 28144921 -> 2116965 (-92.48%)
helped: 6
HURT: 2
helped stats (abs) min: 1404978 max: 7766106 x̄: 4353922.00 x̃: 3890682
helped stats (rel) min: 82.01% max: 95.57% x̄: 89.95% x̃: 92.28%
HURT stats (abs) min: 47778 max: 47798 x̄: 47788.00 x̃: 47788
HURT stats (rel) min: 278.20% max: 282.98% x̄: 280.59% x̃: 280.59%
95% mean confidence interval for cycles value: -5900438.73 -606550.27
95% mean confidence interval for cycles %-change: -140.79% 146.16%
Inconclusive result (%-change mean confidence interval includes 0).
total spills in shared programs: 9243 -> 8901 (-3.70%)
spills in affected programs: 2718 -> 2376 (-12.58%)
helped: 4
HURT: 4
total fills in shared programs: 21831 -> 10141 (-53.55%)
fills in affected programs: 11804 -> 114 (-99.03%)
helped: 6
HURT: 2
total sends in shared programs: 815912 -> 815912 (0.00%)
sends in affected programs: 0 -> 0
helped: 0
HURT: 0
LOST: 1
GAINED: 3
The helped shaders are all compute shaders in Aztec Ruins. There is
also a compute shader in synmark2 OglCSDof that's helped but it doesn't
show up in above shader-db results because it went from SIMD8 to SIMD16.
That shader improves enough to yield an 15-20% performance boost to the
benchmark as a whole on my KBL laptop. The hurt shaders are a couple
shaders in Kerbal Space Program and a couple in Aztec Ruins.
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Set of opcodes doesn't have enough flexibility in certain cases. E.g.
Utgard PP has vector conditional select operation, but condition is always
scalar. Lowering all the vector selects to scalar increases instruction
number, so we need a way to filter only those ops that can't be handled
in hardware.
Reviewed-by: Qiang Yu <yuq825@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
v2: add to series
v3: update Makefile.sources
v4: don't remove a comment and break statement
v4: use nir_can_move_instr
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
The rules for gl_SubgroupSize in Vulkan require that it be a constant
that can be queried through the API. However, all GL requires is that
it's a uniform. Instead of always claiming that the subgroup size in
the shader is 32 in GL like we have to do for Vulkan, claim 8 for
geometry stages, the maximum for fragment shaders, and the actual size
for compute.
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Instead of lowering the subgroup size so early, wait until we have more
information. In particular, we're going to want different subgroup
sizes from different stages depending on the API. We also defer
lowering of subgroup masks because the ge/gt masks require the subgroup
size to generate a subgroup mask.
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
