Commit graph

28 commits

Author SHA1 Message Date
Daniel Schürmann
0680b258f4 aco: align subdword registers during RA when necessary
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Daniel Schürmann
031edbc4a5 aco: adapt register allocation for subdword registers
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Daniel Schürmann
2c74fc98b8 aco: create helper function to collect variables from register area
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Daniel Schürmann
aca2bbf975 aco: add notion of subdword registers to register allocator
To not having to split the register file into single bytes,
we maintain a map with registers which contain subdword variables.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Daniel Schürmann
90811554da aco: remove unnecessary reg_file.fill() operation in get_reg_create_vector()
No pipelinedb changes

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Daniel Schürmann
7de003473c aco: fix Temp and assignment of renamed operands during RA
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Rhys Perry
34424b81df aco: make PhysReg in units of bytes
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-By: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4002>
2020-04-03 23:13:15 +01:00
Rhys Perry
1872759f55 aco: add a late kill flag
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/3914>
2020-03-16 16:09:02 +00:00
Albert Astals Cid
760fe44e8c aco: pass vars by const &
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3935>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3935>
2020-03-02 13:18:49 +00:00
Rhys Perry
fe5c5507bd aco: add some helpers for filling/testing register ranges
We do this a lot

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3768>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3768>
2020-02-19 12:23:50 +00:00
Rhys Perry
43497e30e2 aco: add RegisterFile
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/3768>
2020-02-19 12:23:50 +00:00
Daniel Schürmann
3b323d6601 aco: fix image_atomic_cmp_swap
Fixes: 71440ba0f5 ('aco: reorder VMEM operands in ACO IR')
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3652>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3652>
2020-01-31 16:51:46 +00:00
Daniel Schürmann
99d032f3cd aco: fix register allocation with multiple live-range splits
This patch fixes register allocation if multiple live-range splits
occur to the same variable within one instruction.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3602>
2020-01-29 18:45:23 +00:00
Daniel Schürmann
71440ba0f5 aco: reorder VMEM operands in ACO IR
For all VMEM instructions, the resource constant is now
in operands[0]. For MIMG instructions, the sampler shares
operands[1] with write data in case this instruction writes memory.
Moving the VADDR to be the last operand for MIMG is the first step to
support Navi NSA encoding.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3602>
2020-01-29 18:45:23 +00:00
Rhys Perry
15a1cc00d3 aco: fix off-by-one error when initializing sgpr_live_in
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2394
Fixes: 93c8ebfa78 ('aco: Initial commit of independent AMD compiler')
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3511>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3511>
2020-01-22 17:23:30 +00:00
Daniel Schürmann
427e5eeb02 aco: handle phi affinities transitively through parallelcopies
This can coalesce most unnecessarily inserted parallelcopies
from lowering to CSSA.

v2: refactor loop a bit to make it more efficient and readable.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3385>
2020-01-16 16:01:59 +01:00
Rhys Perry
f9405ceb8a aco: don't move literal to reg when making an instruction VOP3 on GFX10
pipeline-db (Navi):
Totals from affected shaders:
SGPRS: 163398 -> 163398 (0.00 %)
VGPRS: 143820 -> 143820 (0.00 %)
Spilled SGPRs: 0 -> 0 (0.00 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Code Size: 13065744 -> 13044308 (-0.16 %) bytes
Max Waves: 18921 -> 18921 (0.00 %)
Instructions: 2514644 -> 2509285 (-0.21 %)

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/2883>
2020-01-14 12:56:28 +00:00
Daniel Schürmann
ffb4790279 aco: compact various Instruction classes
No pipelinedb changes.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3332>
2020-01-10 17:49:18 +00:00
Daniel Schürmann
6a586a6006 aco: split read/writelane opcode into VOP2/VOP3 version for SI/CI
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-12-07 11:23:11 +01:00
Timur Kristóf
e0bcefc3a0 aco/wave32: Use lane mask regclass for exec/vcc.
Currently all usages of exec and vcc are hardcoded to use s2 regclass.
This commit makes it possible to use s1 in wave32 mode and
s2 in wave64 mode.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
2019-12-04 10:36:01 +00:00
Daniel Schürmann
8861a82be7 aco: don't split live-ranges of linear VGPRs
Fixes: 93c8ebfa78 'aco: Initial commit of independent AMD compiler'

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-11-29 21:54:27 +01:00
Daniel Schürmann
b6f5085dfe aco: preserve kill flag on moved operands during RA
Fixes: 93c8ebfa78 aco: Initial commit of independent AMD compiler

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-11-12 15:59:48 +00:00
Daniel Schürmann
a2a6880743 aco: fix invalid access on Pseudo_instructions
Fixes: 93c8ebfa78 aco: Initial commit of independent AMD compiler

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-11-12 15:59:48 +00:00
Daniel Schürmann
8023dcd71e aco: fix live-range splits of phis
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Timur Kristóf
c52ebbcea4 aco: Introduce vgpr_limit to keep track of available VGPRs.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
2019-10-28 23:52:50 +00:00
Rhys Perry
08d510010b aco: increase accuracy of SGPR limits
SGPRs are allocated in groups of 16 on GFX8/GFX9. GFX10 allocates a fixed
number of SGPRs and has 106 addressable SGPRs.

pipeline-db (Vega):
SGPRS: 5912 -> 6232 (5.41 %)
VGPRS: 1772 -> 1780 (0.45 %)
Spilled SGPRs: 0 -> 0 (0.00 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 88228 -> 87904 (-0.37 %) bytes
LDS: 0 -> 0 (0.00 %) blocks
Max Waves: 559 -> 571 (2.15 %)

piepline-db (Navi):
SGPRS: 341256 -> 363384 (6.48 %)
VGPRS: 171536 -> 170960 (-0.34 %)
Spilled SGPRs: 832 -> 581 (-30.17 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 14207332 -> 14190872 (-0.12 %) bytes
LDS: 33 -> 33 (0.00 %) blocks
Max Waves: 18072 -> 18251 (0.99 %)

v2: unconditionally count vcc as an extra sgpr on GFX10+
v3: pass SGPRs rounded to 8

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
2019-10-23 19:11:21 +01:00
Mauro Rossi
c24ad565ae android: aco: fix undefined template 'std::__1::array' build errors
Fixes a few building errors similar to the following:

In file included from external/mesa/src/amd/compiler/aco_instruction_selection.cpp:26:
In file included from external/libcxx/include/algorithm:639:
external/libcxx/include/utility:321:9:
error: implicit instantiation of undefined template 'std::__1::array<aco::Temp, 4>'
    _T2 second;
        ^

Fixes: 93c8ebf ("aco: Initial commit of independent AMD compiler")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
2019-09-28 15:56:23 +02:00
Daniel Schürmann
93c8ebfa78 aco: Initial commit of independent AMD compiler
ACO (short for AMD Compiler) is a new compiler backend with the goal to replace
LLVM for Radeon hardware for the RADV driver.

ACO currently supports only VS, PS and CS on VI and Vega.
There are some optimizations missing because of unmerged NIR changes
which may decrease performance.

Full commit history can be found at
https://github.com/daniel-schuermann/mesa/commits/backend

Co-authored-by: Daniel Schürmann <daniel@schuermann.dev>
Co-authored-by: Rhys Perry <pendingchaos02@gmail.com>
Co-authored-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Co-authored-by: Connor Abbott <cwabbott0@gmail.com>
Co-authored-by: Michael Schellenberger Costa <mschellenbergercosta@googlemail.com>
Co-authored-by: Timur Kristóf <timur.kristof@gmail.com>

Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-09-19 12:10:00 +02:00