These will let us make the SDWA optimizer much simpler than if we were to
recognize combinations of shift/and/bfe.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3151>
For example, v_cvt_f64_i32. LLVM doesn't seem to allow this either and it
doesn't seem to work correctly.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3151>
Be consistent with other usages in Vulkan and SPIR-V, and the recently
added workgroup_size field.
Acked-by: Emma Anholt <emma@anholt.net>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11190>
This clarifies the semantics of the index variables compared to the previous
version, which used the same variables in a slightly different way depending
on whether they were used for downwards moves or upwards ones.
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10885>
Refactoring total_demand and total_demand_clause to cover non-overlapping
instruction intervals makes the code easier to follow and allows the register
demand to be updated more efficiently in some cases.
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10885>
nir_cf_reinsert() can re-create the block, invalidating dominance
metadata.
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/9808>
Instead of just reporting the failed statements, print where they
originated. This is useful for tests which have a number of similar
checks.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10898>
A simple post-RA optimization which takes advantage of the
s_cbranch_vccz and s_cbranch_vccnz instructions.
It works on the following pattern:
vcc = v_cmp ...
scc = s_and vcc, exec
p_cbranch scc
The result looks like this:
vcc = v_cmp ...
p_cbranch vcc
Fossil DB results on Sienna Cichlid:
Totals from 4814 (3.21% of 149839) affected shaders:
CodeSize: 15371176 -> 15345964 (-0.16%)
Instrs: 3028557 -> 3022254 (-0.21%)
Latency: 21872753 -> 21823476 (-0.23%); split: -0.23%, +0.00%
InvThroughput: 4470282 -> 4468691 (-0.04%); split: -0.04%, +0.00%
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/7779>
This commit adds the skeleton of a new ACO post-RA optimizer,
which is intended to be a simple pass called after RA, and
is meant to do code changes which can only be done
after RA.
It is currently empty, the actual optimizations will be added
in their own commits. It only has a DCE pass, which deletes
some dead code generated by the spiller.
Fossil DB results on Sienna Cichlid:
Totals from 375 (0.25% of 149839) affected shaders:
CodeSize: 2933056 -> 2907192 (-0.88%)
Instrs: 534154 -> 530706 (-0.65%)
Latency: 12088064 -> 12084907 (-0.03%); split: -0.03%, +0.00%
InvThroughput: 4433454 -> 4432421 (-0.02%); split: -0.02%, +0.00%
Copies: 81649 -> 78203 (-4.22%)
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/7779>
For fragment shaders that only contain a discard, the exec mask has
to be zero'd and everything discarded.
It seems unnecessary to emit an export here because if the FS has no
exports, the compiler already emits a null export at the end.
Fixes incorrect hair rendering in Detroit: Become Human.
fossil-db (Sienna Cichlid):
Totals from 3 (0.00% of 149839) affected shaders:
CodeSize: 2896 -> 2872 (-0.83%)
Instrs: 556 -> 553 (-0.54%)
Latency: 29266 -> 29214 (-0.18%)
InvThroughput: 3374 -> 3372 (-0.06%)
Cc: 21.1 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10955>
ds_swizzle_b32 requires a VGPR and DPP can't encode SGPR sources.
Fixes
dEQP-VK.graphicsfuzz.cov-derivative-uniform-vector-global-loop-count.
Cc: 21.1 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/10840>
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>
These don't really need the exec mask (and never have), but we haven't
needed to include them in needs_exec_mask yet.
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>
We need to know if there's a temporary in SCC before the instruction, not
after.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: 93c8ebfa78 ("aco: Initial commit of independent AMD compiler")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10459>
We don't want to have to deal with vector phis in freedreno, because
vectors are always split/unsplit around vectorized instructions anyways,
and the stated reason for not scalarising them (it hurting coalescing)
won't apply to us because we won't be using nir_from_ssa. Add this
option so that we don't have to do the equivalent thing while
translating from NIR.
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10809>
It could happen that we tested negative out-of-range
registers for p_create_vector operands resulting in a crash.
Fixes: 8962510e38 ('aco/ra: Conservatively refactor get_reg_specified to use PhysRegInterval')
Closes: #4697
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10799>
The initial value needs to be taken from the instruction that is being
moved over, not the one to be moved.
Additionally the parameter of this function was removed because it was
misleading. Setting it to any value other than source_idx would cause
register_demand to be initialized incorrectly. (Instead, the maximum
demand among the covered instructions would need to be determined.)
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10644>
By exposing a subgroupSize of 64, reductions with
cluster_size 32 in wave32 might be considered divergent,
and thus, result in a VGPR.
Fixes: dEQP-VK.subgroups.clustered.graphics.subgroupclustered* with wave32
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10769>
It could happen that VGPR spilling without SGPR spilling
calculated a negative spills_to_vgpr number and then
increasing the VGPR target demand above the limit.
Cc: mesa-stable
Reviewed-by: Tony Wasserka <tony.wasserka@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10756>
