This commit used to be "brw/emit: Allow scalar sources to 64-bit
3-source instructions". These instructions were fixed up in
brw_eu_emit. There seems to be some conflict with the <0,1,0> stride an
post-RA scheduling. The only difference between the passing code
generated by this commit and the failing code generated by the older
commit is some post-RA scheduling.
v2: Change the stride of a MAD even if the instruction isn't
lowered. MAD instructions that are already SIMD8 have to follow the same
rules. 🤦
v3: Pull the lowering out to its own pass. Update the comment in
brw_fs_validate. Suggested by Ken.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29884>
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>
The general idea is to be able to validate that certain instructions
were lowered and certain restrictions were already handled. Passes can
now assert their expectations, i.e. if a pass is mean to run after
certain lowerings or not.
The actual phases are a initial stab and as we re-organized the passes,
we may remove/add phases.
This commit just add some phase steps, later commits will make use of
them.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30496>
The long names were originally meant to map to the HW encoding but
nowadays the actual encoding values depend on gfx version, whether
instruction is 3src, etc.
Suggested by Ken.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30704>
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>
Indirect addressing(vx1 and vxh) not supported with UB/B datatype for
src0, so we need to change the data type for both dest and src0.
This fixes following tests cases on Xe2+
- dEQP-VK.spirv_assembly.instruction.compute.8bit_storage.push_constant_8_to_16*
- dEQP-VK.spirv_assembly.instruction.compute.8bit_storage.push_constant_8_to_32*
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29316>
We can do CSEL on F, HF, *W, and *D on Gfx11+. Gfx9 can only do F.
We can lower unsupported types to CMP+CSEL, allowing us to use CSEL
in the IR and not worry about the limitations.
Rework: (Sagar)
- Update validation pass for CSEL
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29316>
This has no impact on the generated shaders, but does have a small
(positive) impact on the amount of time spent in shader compilation.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29126>
The new def-based pass works better in many cases, and should be less
resource intensive. However, the limited visibility of the defs-based
pass due to many values not being SSA yet makes it unable to fully
replace the old pass. Try the new one, and if it can't make progress,
then try the old one. That way, things will mostly be handled by the
new pass, but everything that was being cleaned up still will be.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28666>
While the limited visibility due to partial SSA is a downside to the new
pass, it has a huge number of advantages that make it worth switching
over even now. It's much more efficient, can eliminate redundant memory
loads across blocks, and doesn't generate loads of unnecessary copies
that other passes have to clean up. This means we also eliminate the
infighting between the old CSE, coalescing, and copy propagation passes.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28666>
Our code to initialize gl_SubgroupInvocation uses multiple instructions
some of which are partial writes. This makes it difficult to analyze
expressions involving gl_SubgroupInvocation, which appear very
frequently in compute shaders.
To make this easier, we add a new virtual opcode which initializes
a full VGRF to the value of gl_SubgroupInvocation. (We also expand
it to UD for SIMD8 so there are not partial write issues.) We then
lower it to the original code later on in compilation, after we've
done the bulk of our optimizations.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28666>
Both of these helpers do the same thing. We now have brw_type_size_bits
and brw_type_size_bytes and can use whichever makes sense in that place.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28847>
The OPT macro will call validate() after each pass, so both cases
removed by this patch are just redundant calls. Will only affect
Debug builds since in Release builds validation is a no-op.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28534>
opt_copy_propagation() can sometimes propagate FIXED_GRF sources into
SHADER_OPCODE_SENDs as the message payload. For example, GS input
reads, which simply take a URB handle and have the offset in the
descriptor. For non-VGRFs, there isn't a payload to split, so just
skip past such send messages.
Fixes: 589b03d02f ("intel/fs: Opportunistically split SEND message payloads")
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28067>
The mlen tracking is in REG_SIZE units, but in Xe2 each GRF has
doubled the size. The optimization can only elide full GRFs, so
round down the amount of trailing zeros to ensure the optimization
will remove only full GRFs.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28279>
brw_fs_opt_eliminate_find_live_channel eliminates FIND_LIVE_CHANNEL
outside of control flow. None of our optimization passes generate
additional cases of that instruction, so once it's gone, we shouldn't
ever have to run the pass again. Moving it out of the loop should
save a bit of CPU time.
While we're at it, also clean adjacent BROADCAST instructions that
consume the result of our FIND_LIVE_CHANNEL. Without this, we have
to perform copy propagation to get the MOV 0 immediate into the
BROADCAST, then algebraic to turn it into a MOV, which enables more
copy propagation...not to mention CSE gets involved. Since this
FIND_LIVE_CHANNEL + BROADCAST pattern from emit_uniformize() is
really common, and it's trivial to clean up, we can do that. This
lets the initial copy prop in the loop see MOV instead of BROADCAST.
Zero impact on fossil-db, but less work in the optimization loop.
Together with the previous patches, this cuts compile time in
Borderlands 3 on Alchemist by -1.38539% +/- 0.1632% (n = 24).
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28286>
We don't necessarily want to split up MOVs for 64-bit addresses into
2x 32-bit MOVs right away, as this makes things like copy propagating
the whole address around harder. We should do this late, once, while
still doing other algebraic optimizations earlier.
fossil-db results for Alchemist show tiny improvements:
Totals:
Instrs: 161310502 -> 161310436 (-0.00%); split: -0.00%, +0.00%
Cycles: 14370605606 -> 14370605159 (-0.00%); split: -0.00%, +0.00%
Totals from 33 (0.01% of 652298) affected shaders:
Instrs: 15053 -> 14987 (-0.44%); split: -0.64%, +0.20%
Cycles: 196947 -> 196500 (-0.23%); split: -0.25%, +0.02%
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28286>
This copy propagation can create MADs with immediates in src1, which
need to be cleaned up by constant combining (which puts them back in
VGRFs).
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27876>
