The new ac_hw_stage is going to be used by drivers as well.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Qiang Yu <yuq825@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23597>
We don't want to rely on any NIR structures in ACO, because
we would like to avoid the need to include nir.h in aco_ir.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/22241>
Dual source blending must be in strict WQM mode.
Cc: 22.3 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19643>
"Removing jumps" in ACO means skipping the jump instruction
at the beginning of a divergent branch (but still modify exec).
ACO already supports implicitly removing jumps when it decides
that executing a branch with empty exec mask is more beneficial
than a jump.
This commit adds the possibility to use this explicitly
through nir_selection_control. ACO will respect this
setting and remove the branch instructions when this is specified,
unless it decides that this would cause bugs (eg. exp instruction).
There are two cases that benefit from the new change:
1. When the application requests to "flatten" a branch (ie.
remove control flow), we now respect that.
2. When the compiler stack determines that a divergent branch
is always taken.
v2 by Georg Lehmann: fixed applying sel_ctrl to else blocks
Fossil DB stats on Navi 21:
Totals from 13 (0.01% of 134906) affected shaders:
CodeSize: 136616 -> 136496 (-0.09%)
Instrs: 26196 -> 26166 (-0.11%)
Latency: 417928 -> 417889 (-0.01%)
Branches: 1241 -> 1211 (-2.42%)
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-By: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17921>
Otherwise, in presence of p_exit_early_if the main FS will always
jump to the PS epilog regardless the exact mask.
This fixes dEQP-VK.draw.renderpass.shader_invocation.helper_invocation
and few vkd3d-proton regressions when PS epilogs are forced.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17617>
Because the data register of atomic VMEM instructions
is shared between src and dst, it might be necessary
to create live-range splits during RA.
Make the live-range splits explicit in WQM mode.
Totals from 7 (0.01% of 134913) affected shaders: (GFX10.3)
Latency: 17209 -> 17210 (+0.01%)
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15347>
After transitioning to exact after a discard, the exec stack might be:
[exact|global, wqm, exact]
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15244>
This was needed, so that in case of active helper lanes,
these contain the correct value. It is now handled implicitly.
Totals from 1004 (0.74% of 134913) affected shaders: (GFX10.3)
CodeSize: 7581020 -> 7580892 (-0.00%); split: -0.00%, +0.00%
Instrs: 1454940 -> 1454908 (-0.00%); split: -0.00%, +0.00%
Latency: 12984953 -> 12984894 (-0.00%); split: -0.00%, +0.00%
InvThroughput: 3173037 -> 3173049 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 47498 -> 47273 (-0.47%)
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14951>
This case doesn't seem to happen in practice.
No need to micro-optimize it.
This patch merges instruction selection for discard/discard_if.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14805>
Implement elect using a pseudo-op which is lowered during the
insert_exec_mask pass. This makes it possible to emit a more
optimal sequence when the exec mask is constant.
Fossil DB results on Sienna Cichlid:
Totals from 211 (0.16% of 128647) affected shaders:
CodeSize: 2254356 -> 2240468 (-0.62%); split: -0.62%, +0.00%
Instrs: 438471 -> 434996 (-0.79%); split: -0.80%, +0.01%
Latency: 2717082 -> 2709400 (-0.28%); split: -0.28%, +0.00%
InvThroughput: 566987 -> 566342 (-0.11%); split: -0.11%, +0.00%
Copies: 40058 -> 40162 (+0.26%)
Branches: 31209 -> 31211 (+0.01%)
PreSGPRs: 9927 -> 10125 (+1.99%)
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11458>
This migration was done with libclang-based automatic tooling, which
performed these replacements:
* Operand(uint8_t) -> Operand::c8
* Operand(uint16_t) -> Operand::c16
* Operand(uint32_t, false) -> Operand::c32
* Operand(uint32_t, bool) -> Operand::c32_or_c64
* Operand(uint64_t) -> Operand::c64
* Operand(0) -> Operand::zero(num_bytes)
Casts that were previously used for constructor selection have automatically
been removed (e.g. Operand((uint16_t)1) -> Operand::c16(1)).
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11653>
When exec is constant, we can remember the constant as the old exec,
and just copy the condition and use it as the new exec. There is no
need to save the constant.
Due to using p_parallelcopy which is lowered to s_mov_b64 (or 32),
many exec restores now become copies, hence the increase in the copy
stats.
Fossil DB changes on Sienna Cichlid:
Totals from 73969 (49.37% of 149839) affected shaders:
SpillSGPRs: 1768 -> 1610 (-8.94%)
CodeSize: 99053892 -> 99047884 (-0.01%); split: -0.02%, +0.01%
Instrs: 19372852 -> 19370398 (-0.01%); split: -0.02%, +0.01%
VClause: 515154 -> 515142 (-0.00%); split: -0.00%, +0.00%
SClause: 719236 -> 718395 (-0.12%); split: -0.14%, +0.02%
Copies: 1109770 -> 1254634 (+13.05%); split: -0.07%, +13.12%
Branches: 374338 -> 374348 (+0.00%); split: -0.00%, +0.00%
PreSGPRs: 1776481 -> 1653761 (-6.91%)
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10691>
Previously, we would store even the constant -1 exec mask from the
beginning of every merged shader. With this change it is no longer
necessary because we can restore to constant exec mask directly.
Hence, this frees up a register pair (single register for Wave32)
in every merged shader.
No Fossil DB changes.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10691>
This will enable us to store non-temporary values,
such as constant operands there.
No Fossil DB changes.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10691>
This should no longer be necessary since the mark_block_wqm() we use to
flag break conditions as WQM now adds block to the worklist. With them
added to the worklist, get_block_needs() will add WQM to block_needs.
Adding WQM to block_needs here without adding the block to the worklist
(like we do here) can cause issues because it does not ensure that the
predecessors' branches are in WQM (needed for it to be possible to
transition to WQM in the block). This happened in an Overwatch shader.
No fossil-db changes.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: 661922f6ac ("aco: add block to worklist in mark_block_wqm()")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/4066
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/8446>
