Some intrinsics are implemented by reading memory location that could
be rewritten by a further tracing calls. So we need to move those
reads prior to tracing operations in the shaders.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8979
Tested-by: Sviatoslav Peleshko <sviatoslav.peleshko@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34214>
The condition modifier on SEL means something completely different than
it means on MOV. On MOV it means to modify the flags based on the value
written to the destination. On SEL it means to compare the sources using
that mode and pick the result (i.e., as min() or max()) without
modifying the flags.
The resulting MOV should not have a condition modifier for the same
reason it (already) doesn't have a predicate. This bug was found by
inspection, so I added a unit test.
No shader-db or shader-db changes on any Intel platform.
Fixes: fab92fa1cb ("i965/fs: Optimize SEL with the same sources into a MOV.")
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34192>
v2: Add support for WE_all instructions... this already just worked, so
I only had to delete the check and the FINISHME comment.
v3: Use logic more like def_analysis::update_for_reads to determine when
to not insert LOAD_REG instructions. Based on a suggestion by Ken.
v4: Eliminate "store" from all the names since STORE_REG does not exist
anymore. Fold insert_load_reg into brw_insert_load_reg. Elminate extra
call to s.def_analysis.require() after progress. Pull a loop-invariant
check out of the inst->srouces loop. Drop call to
brw_opt_split_virtual_grfs after lowering load_reg. All suggested by
Caio.
v5: Assert that LOAD_REG doesn't already exist in
brw_insert_load_reg. Update comment before fully_defines. Both
suggested by Caio.
v6: Don't explicitly special-case SHADER_OPCODE_MEMORY_STORE_LOGICAL.
Move the inst->dst.file != VGRF check earlier to avoid the loop over
sources. Both suggested by Ken. Move the call the brw_insert_load_reg
a little bit later, and explain why it's at that location. Suggested
by Caio.
v7: Many changes to the for-each-source loop in brw_insert_load_reg.
Removes incorrect multiplication of s.alloc.sizes with reg_unit. Adds
checks for matching SIMD size and NoMask in the search for pre-existing
LOAD_REG of same value.
v8: Add some unit tests. Suggested by Caio.
shader-db:
Lunar Lake
total instructions in shared programs: 16923237 -> 16921895 (<.01%)
instructions in affected programs: 450565 -> 449223 (-0.30%)
helped: 251 / HURT: 377
total cycles in shared programs: 910428418 -> 889920590 (-2.25%)
cycles in affected programs: 719248184 -> 698740356 (-2.85%)
helped: 9076 / HURT: 9082
total fills in shared programs: 2242 -> 2218 (-1.07%)
fills in affected programs: 116 -> 92 (-20.69%)
helped: 2 / HURT: 0
total sends in shared programs: 848635 -> 848421 (-0.03%)
sends in affected programs: 810 -> 596 (-26.42%)
helped: 10 / HURT: 0
LOST: 82
GAINED: 78
Meteor Lake and DG2 had similar results. (Meteor Lake shown)
total instructions in shared programs: 19875784 -> 19871694 (-0.02%)
instructions in affected programs: 1050091 -> 1046001 (-0.39%)
helped: 251 / HURT: 2403
total cycles in shared programs: 905328238 -> 882446458 (-2.53%)
cycles in affected programs: 682736344 -> 659854564 (-3.35%)
helped: 7869 / HURT: 7911
total spills in shared programs: 5512 -> 5032 (-8.71%)
spills in affected programs: 1830 -> 1350 (-26.23%)
helped: 8 / HURT: 0
total fills in shared programs: 5648 -> 4782 (-15.33%)
fills in affected programs: 3312 -> 2446 (-26.15%)
helped: 8 / HURT: 0
total sends in shared programs: 1032942 -> 1032722 (-0.02%)
sends in affected programs: 572 -> 352 (-38.46%)
helped: 10 / HURT: 0
LOST: 138
GAINED: 53
Tiger Lake
total instructions in shared programs: 19711930 -> 19715591 (0.02%)
instructions in affected programs: 1040623 -> 1044284 (0.35%)
helped: 317 / HURT: 2474
total cycles in shared programs: 862988990 -> 860573870 (-0.28%)
cycles in affected programs: 612392461 -> 609977341 (-0.39%)
helped: 7447 / HURT: 7686
total sends in shared programs: 1034763 -> 1034555 (-0.02%)
sends in affected programs: 784 -> 576 (-26.53%)
helped: 8 / HURT: 0
LOST: 56
GAINED: 143
Ice Lake and Skylake had similar results. (Ice Lake shown)
total instructions in shared programs: 20545461 -> 20545220 (<.01%)
instructions in affected programs: 422405 -> 422164 (-0.06%)
helped: 180 / HURT: 459
total cycles in shared programs: 872697345 -> 866874523 (-0.67%)
cycles in affected programs: 573117917 -> 567295095 (-1.02%)
helped: 6783 / HURT: 6980
total spills in shared programs: 4335 -> 4336 (0.02%)
spills in affected programs: 90 -> 91 (1.11%)
helped: 1 / HURT: 2
total fills in shared programs: 4194 -> 4196 (0.05%)
fills in affected programs: 463 -> 465 (0.43%)
helped: 1 / HURT: 2
total sends in shared programs: 1079446 -> 1079238 (-0.02%)
sends in affected programs: 784 -> 576 (-26.53%)
helped: 8 / HURT: 0
LOST: 117
GAINED: 37
fossil-db:
All Intel platforms had similar results. (Lunar Lake shown)
Totals:
Instrs: 209708136 -> 209695617 (-0.01%); split: -0.02%, +0.01%
Send messages: 10927753 -> 10927640 (-0.00%)
Cycle count: 30540172048 -> 30427084732 (-0.37%); split: -0.99%, +0.62%
Spill count: 511621 -> 510932 (-0.13%); split: -0.22%, +0.08%
Fill count: 621166 -> 618440 (-0.44%); split: -0.56%, +0.12%
Scratch Memory Size: 35574784 -> 35648512 (+0.21%); split: -0.06%, +0.26%
Max live registers: 65453860 -> 65453140 (-0.00%); split: -0.00%, +0.00%
Non SSA regs after NIR: 75374990 -> 35195764 (-53.31%)
Totals from 503284 (71.25% of 706391) affected shaders:
Instrs: 180203778 -> 180191259 (-0.01%); split: -0.02%, +0.01%
Send messages: 9699732 -> 9699619 (-0.00%)
Cycle count: 30080349592 -> 29967262276 (-0.38%); split: -1.01%, +0.63%
Spill count: 511584 -> 510895 (-0.13%); split: -0.22%, +0.08%
Fill count: 621120 -> 618394 (-0.44%); split: -0.56%, +0.12%
Scratch Memory Size: 35443712 -> 35517440 (+0.21%); split: -0.06%, +0.27%
Max live registers: 52566092 -> 52565372 (-0.00%); split: -0.01%, +0.00%
Non SSA regs after NIR: 70110949 -> 29931723 (-57.31%)
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31497>
The key change is to use a builder to write the expected shader result
and compare that. To make this less error prone, a few helper functions
were added
- a way to allocate VGRFs from both shaders in parallel, that way the
same brw_reg can be used in both of them;
- assertions that a pass will make progress or not, and proper output
when the unexpected happens;
- use a common brw_shader_pass_test class so to collect some of the helpers;
- make some helpers work directly with builder.
The idea is to improve the signal in tests, so that the disasm comments
are not necessary anymore. For example
```
TEST_F(saturate_propagation_test, basic)
{
brw_reg dst1 = bld.vgrf(BRW_TYPE_F);
brw_reg src0 = bld.vgrf(BRW_TYPE_F);
brw_reg src1 = bld.vgrf(BRW_TYPE_F);
brw_reg dst0 = bld.ADD(src0, src1);
set_saturate(true, bld.MOV(dst1, dst0));
/* = Before =
*
* 0: add(16) dst0 src0 src1
* 1: mov.sat(16) dst1 dst0
*
* = After =
* 0: add.sat(16) dst0 src0 src1
* 1: mov(16) dst1 dst0
*/
brw_calculate_cfg(*v);
bblock_t *block0 = v->cfg->blocks[0];
EXPECT_EQ(0, block0->start_ip);
EXPECT_EQ(1, block0->end_ip);
EXPECT_TRUE(saturate_propagation(v));
EXPECT_EQ(0, block0->start_ip);
EXPECT_EQ(1, block0->end_ip);
EXPECT_EQ(BRW_OPCODE_ADD, instruction(block0, 0)->opcode);
EXPECT_TRUE(instruction(block0, 0)->saturate);
EXPECT_EQ(BRW_OPCODE_MOV, instruction(block0, 1)->opcode);
EXPECT_FALSE(instruction(block0, 1)->saturate);
}
```
becomes
```
TEST_F(saturate_propagation_test, basic)
{
brw_builder bld = make_shader(MESA_SHADER_FRAGMENT, 16);
brw_builder exp = make_shader(MESA_SHADER_FRAGMENT, 16);
brw_reg dst0 = vgrf(bld, exp, BRW_TYPE_F);
brw_reg dst1 = vgrf(bld, exp, BRW_TYPE_F);
brw_reg src0 = vgrf(bld, exp, BRW_TYPE_F);
brw_reg src1 = vgrf(bld, exp, BRW_TYPE_F);
bld.ADD(dst0, src0, src1);
bld.MOV(dst1, dst0)->saturate = true;
EXPECT_PROGRESS(brw_opt_saturate_propagation, bld);
exp.ADD(dst0, src0, src1)->saturate = true;
exp.MOV(dst1, dst0);
EXPECT_SHADERS_MATCH(bld, exp);
}
```
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33936>
The expectations are :
- no MSAA images
- a single tiling mode is used when not linear
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32676>
This was originally turned into a separate struct for reuse between vec4
and fs backends, that's not needed anymore.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33334>
Defaults to true. When set to false Iris and various tools can be
built without ELK support. In both cases this means supporting
only Gfx9+. This option must be true to build Crocus or Hasvk.
This allows skipping re-building ELK when developing for newer platforms
with tools/tests enabled.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11575
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33054>
In theory you can build a driver using OpenCL kernels with a
-Dmesa-clc=system. That shouldn't require any LLVM/Clang/etc...
But the checks to find the pre-compiled mesa_clc & vtn_bindgen
binaries are in meson files or conditions only triggered if you build
with LLVM (:
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Tested-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Dylan Baker <None>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33014>
As Asahi, Intel and soon Panfrost requires an offline compiler for their
respective internal shaders, this commit adds generic new options to
workaround meson current limitations around cross-compillation.
Signed-off-by: Mary Guillemard <mary.guillemard@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32719>
Borderlands 3 (both DX11 and DX12 renderers) have a common pattern
across many shaders:
con 32x4 %510 = (uint32)txf %2 (handle), %1191 (0x10) (coord), %1 (0x0) (lod), 0 (texture)
con 32x4 %512 = (uint32)txf %2 (handle), %1511 (0x11) (coord), %1 (0x0) (lod), 0 (texture)
...
con 32x4 %550 = (uint32)txf %2 (handle), %1549 (0x25) (coord), %1 (0x0) (lod), 0 (texture)
con 32x4 %552 = (uint32)txf %2 (handle), %1551 (0x26) (coord), %1 (0x0) (lod), 0 (texture)
A single basic block contains piles of texelFetches from a 1D buffer
texture, with constant coordinates. In most cases, only the .x channel
of the result is read. So we have something on the order of 28 sampler
messages, each asking for...a single uint32_t scalar value. Because our
sampler doesn't have any support for convergent block loads (like the
untyped LSC transpose messages for SSBOs)...this means we were emitting
SIMD8/16 (or SIMD16/32 on Xe2) sampler messages for every single scalar,
replicating what's effectively a SIMD1 value to the entire register.
This is hugely wasteful, both in terms of register pressure, and also in
back-and-forth sending and receiving memory messages.
The good news is we can take advantage of our explicit SIMD model to
handle this more efficiently. This patch adds a new optimization pass
that detects a series of SHADER_OPCODE_TXF_LOGICAL, in the same basic
block, with constant offsets, from the same texture. It constructs a
new divergent coordinate where each channel is one of the constants
(i.e <10, 11, 12, ..., 26> in the above example). It issues a new
NoMask divergent texel fetch which loads N useful channels in one go,
and replaces the rest with expansion MOVs that splat the SIMD1 result
back to the full SIMD width. (These get copy propagated away.)
We can pick the SIMD size of the load independently of the native shader
width as well. On Xe2, those 28 convergent loads become a single SIMD32
ld message. On earlier hardware, we use 2 SIMD16 messages. Or we can
use a smaller size when there aren't many to combine.
In fossil-db, this cuts 27% of send messages in affected shaders, 3-6%
of cycles, 2-3% of instructions, and 8-12% of live registers. On A770,
this improves performance of Borderlands 3 by roughly 2.5-3.5%.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32573>
Intel GPUs have a saturate destination modifier, and
brw_fs_opt_saturate_propagation tries to replace explicit saturate
operations with this destination modifier. That pass is limited in
several ways. If the source of the explicit saturate is in a different
block or if the source of the explicit saturate is live after the
explicit saturate, brw_fs_opt_saturate_propagation will be unable to
make progress.
This optimization exists to help brw_fs_opt_saturate_propagation make
more progress. It tries to move NIR fsat instructions to the same block
that contains the definition of its source. It does this only in cases
where it will not create additional live values. It also attempts to do
this only in cases where the explicit saturate will ultimiately be
converted to a destination modifier.
v2: Fix metadata_preserve when theres no progress and use
nir_metadata_control_flow when there is progress. All suggested by
Alyssa.
v3: Fix a typo in the file header comment. Noticed by Ken. Don't
require nir_metadata_instr_index. Use nir_def_rewrite_uses_after instead
of open-coding something slightly more specific. Both suggested by Ken.
shader-db:
All Intel platforms had similar results. (Meteor Lake shown)
total instructions in shared programs: 19733645 -> 19733028 (<.01%)
instructions in affected programs: 193300 -> 192683 (-0.32%)
helped: 246
HURT: 1
helped stats (abs) min: 2 max: 48 x̄: 2.51 x̃: 2
helped stats (rel) min: 0.18% max: 0.39% x̄: 0.33% x̃: 0.34%
HURT stats (abs) min: 1 max: 1 x̄: 1.00 x̃: 1
HURT stats (rel) min: 0.31% max: 0.31% x̄: 0.31% x̃: 0.31%
95% mean confidence interval for instructions value: -2.87 -2.13
95% mean confidence interval for instructions %-change: -0.34% -0.32%
Instructions are helped.
total cycles in shared programs: 916180971 -> 916264656 (<.01%)
cycles in affected programs: 30197180 -> 30280865 (0.28%)
helped: 194
HURT: 142
helped stats (abs) min: 1 max: 21251 x̄: 872.75 x̃: 19
helped stats (rel) min: <.01% max: 23.17% x̄: 2.59% x̃: 0.23%
HURT stats (abs) min: 1 max: 28058 x̄: 1781.68 x̃: 399
HURT stats (rel) min: <.01% max: 37.21% x̄: 4.85% x̃: 1.63%
95% mean confidence interval for cycles value: -196.84 694.97
95% mean confidence interval for cycles %-change: -0.17% 1.27%
Inconclusive result (value mean confidence interval includes 0).
fossil-db:
Meteor Lake, DG2, and Tiger Lake had similar results. (Meteor Lake shown)
Totals:
Instrs: 151512021 -> 151511351 (-0.00%); split: -0.00%, +0.00%
Cycle count: 17209013596 -> 17209840995 (+0.00%); split: -0.02%, +0.02%
Max live registers: 32013312 -> 32013549 (+0.00%)
Max dispatch width: 5512304 -> 5512136 (-0.00%)
Totals from 774 (0.12% of 630172) affected shaders:
Instrs: 1559285 -> 1558615 (-0.04%); split: -0.05%, +0.01%
Cycle count: 1312656268 -> 1313483667 (+0.06%); split: -0.24%, +0.30%
Max live registers: 82195 -> 82432 (+0.29%)
Max dispatch width: 6664 -> 6496 (-2.52%)
Ice Lake
Totals:
Instrs: 151416791 -> 151416137 (-0.00%); split: -0.00%, +0.00%
Cycle count: 15162468885 -> 15163298824 (+0.01%); split: -0.00%, +0.01%
Max live registers: 32471367 -> 32471603 (+0.00%)
Max dispatch width: 5623752 -> 5623712 (-0.00%)
Totals from 733 (0.12% of 635598) affected shaders:
Instrs: 877965 -> 877311 (-0.07%); split: -0.09%, +0.01%
Cycle count: 190763628 -> 191593567 (+0.44%); split: -0.21%, +0.64%
Max live registers: 72067 -> 72303 (+0.33%)
Max dispatch width: 6216 -> 6176 (-0.64%)
Skylake
Totals:
Instrs: 140794845 -> 140794075 (-0.00%); split: -0.00%, +0.00%
Cycle count: 14665159301 -> 14665320514 (+0.00%); split: -0.00%, +0.01%
Max live registers: 31783341 -> 31783662 (+0.00%); split: -0.00%, +0.00%
Totals from 659 (0.11% of 625670) affected shaders:
Instrs: 829061 -> 828291 (-0.09%); split: -0.09%, +0.00%
Cycle count: 185478478 -> 185639691 (+0.09%); split: -0.33%, +0.41%
Max live registers: 67491 -> 67812 (+0.48%); split: -0.01%, +0.48%
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29774>
We can achieve most of what brw_fs_opt_predicated_break() does with
simple peepholes at NIR -> BRW conversion time.
For predicated break and continue, we can simply look at an IF ... ENDIF
sequence after emitting it. If there's a single instruction between the
two, and it's a BREAK or CONTINUE, then we can move the predicate from
the IF onto the jump, and delete the IF/ENDIF. Because we haven't built
the CFG at this stage, we only need to remove them from the linked list
of instructions, which is trivial to do.
For the predicated while optimization, we can rely on the fact that we
already did the predicated break optimization, and simply look for a
predicated BREAK just before the WHILE. If so, we move the predicate
onto the WHILE, invert it, and remove the BREAK.
There are a few cases where this approach does a worse job than the old
one: nir_convert_from_ssa may introduce load_reg and store_reg in blocks
containing break, and nir_trivialize_registers may decide it needs to
insert movs into those blocks. So, at NIR -> BRW time, we'll actually
emit some MOVs there, which might have been possible to copy propagate
out after later optimizations.
However, the fossil-db results show that it's still pretty competitive.
For instructions, 1017 shaders were helped (average -1.87 instructions),
while only 62 were hurt (average +2.19 instructions). In affected
shaders, it was -0.08% for instructions.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30498>
Instead of having a hardcoded list of endian-independent format aliases
in the header, generate them from the format definitions.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29649>
When the destination of both instructions is NULL and the conditional
modifier matches, operands_match (by way of instructions_match) will
only test the first two operands. This can result in bad CSE
happening.
This is a very, very narrow edge case.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29848>
This introduces a new analysis pass that opportunistically looks for
VGRFs which happen to satisfy the SSA definition properties.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28666>
