This opcode doesn't write SCC, which gives later passes more freedom to
move instructions.
fossil-db (navi21):
Totals from 727 (0.92% of 79395) affected shaders:
Latency: 14943483 -> 14942704 (-0.01%); split: -0.01%, +0.00%
InvThroughput: 3225790 -> 3225766 (-0.00%); split: -0.00%, +0.00%
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29912>
I tested that v_swap_b16 can be encoded as VOP3, because the ISA doc doesn't list
it as a possible VOP3 opcode. VOP3 is nessecary to access v128+.
Foz-DB Navi31:
Totals from 32 (0.04% of 79395) affected shaders:
Instrs: 201865 -> 195168 (-3.32%)
CodeSize: 1082220 -> 1031228 (-4.71%); split: -4.71%, +0.00%
Latency: 2258198 -> 2238586 (-0.87%)
InvThroughput: 796731 -> 788934 (-0.98%)
Copies: 34514 -> 29220 (-15.34%)
VALU: 122457 -> 117163 (-4.32%)
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29143>
This didn't consider that there might be different programs using the same
SIMD.
fossil-db (navi31):
Totals from 68129 (85.81% of 79395) affected shaders:
Instrs: 23230924 -> 23388315 (+0.68%)
CodeSize: 120636544 -> 121272888 (+0.53%)
Latency: 115645106 -> 115683965 (+0.03%)
InvThroughput: 18804076 -> 18806912 (+0.02%); split: -0.00%, +0.02%
Branches: 404644 -> 407945 (+0.82%)
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/28633>
We can skip the v_or_b32 or use an instruction smaller than
v_alignbyte_b32.
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/19933>
fossil-db (gfx1100):
Totals from 52753 (39.07% of 135032) affected shaders:
CodeSize: 153603860 -> 153163384 (-0.29%)
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/19933>
Foz-DB Navi21:
Totals from 2403 (1.78% of 134913) affected shaders:
CodeSize: 25329156 -> 25311244 (-0.07%)
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19413>
I misread the ISA doc and got the order wrong.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Fixes: dae1629778 ("aco: disable sdwa on gfx11")
Fixes: e68e6c75ca ("aco: use v_perm_b32 to copy 0xff00/0x00ff/0xff/0x00")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19223>
We were using a 16-bit inline constant with a 32-bit instruction and the
test would have created
"v1: %_:v[0] = v_alignbyte_b32 0.5, %_:v[1][16:32], 2" instead.
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/16296>
This commit introduces a new struct SubdwordSel
in order to ease and clean up the usage of SDWA
selections. This includes removing the distinction
between register-allocated and fixed SDWA selections.
Instead, SDWA selections can now also access the high
bits of subdword variables. Alignment and sizes are
validated accordingly. Size, offset and sign_extend
can be evaluated via helper methods.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12640>
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>
We can copy any value into a 16-bit subregister with a 3 dword
v_pack_b32_f16 on GFX10 or a v_and_b32+v_or_b32 on GFX9.
Because the generated code can depend on the register assignment and to
improve constant propagation, Builder::copy creates a p_create_vector in
the case of sub-dword literals.
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/7111>
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Daniel Schürmann <daniel@schuermann.dev>
Acked-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3521>
