Timur Kristóf
07754a9c9e
aco/wave32: Replace hardcoded numbers in spiller with wave size.
...
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
746b9380bd
aco: rematerialize s_movk instructions
...
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
2019-11-12 15:59:48 +00:00
Daniel Schürmann
5c7dcb15e0
aco: only use single-dword loads/stores for spilling
...
Fixes: 8678699918
2019-11-04 20:14:14 +01:00
Daniel Schürmann
d97c0bdd55
aco: fix immediate offset for spills if scratch is used
...
Fixes: 8678699918
2019-11-04 20:14:14 +01:00
Daniel Schürmann
8678699918
aco: implement VGPR spilling
...
VGPR spilling is implemented via MUBUF instructions and scratch memory.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
b0de16b7de
aco: omit linear VGPRs as spill variables
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
aded548e66
aco: ensure that spilled VGPR reloads are done after p_logical_start
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
a7ff1bb5b9
aco: simplify calculation of target register pressure when spilling
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Rhys Perry
e73de4e1d8
aco: fix new_demand calculation for first instructions
...
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
93b42a1907
aco: don't add interferences between spilled phi operands
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
fdf8ad0256
aco: consider loop_exit blocks like merge blocks, even if they have only one predecessor
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
d48d72e98a
aco: don't insert the exec mask into set of live-out variables when spilling
...
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
cd20e29de1
aco: fix transitive affinities of spilled variables
...
Variables spilled on both branch legs need to be assigned to the same spilling slot.
These affinities can be transitive through multiple merge blocks.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:33 +00:00
Daniel Schürmann
0b8216b2cd
aco: Lower to CSSA
...
Converting to 'Conventional SSA Form' ensures correctness w.r.t. spilling of phi nodes.
Previously, it was possible that phi operands have intersecting live-ranges, and thus,
couldn't get spilled to the same spilling slot. For this reason, ACO tried to avoid to
spill phis, even if it was beneficial.
This patch implements a conversion pass which is currently only called if spilling is necessary.
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
2019-10-30 19:48:32 +00:00
Rhys Perry
fc04a2fc31
aco: take LDS into account when calculating num_waves
...
pipeline-db (Vega):
SGPRS: 344 -> 344 (0.00 %)
VGPRS: 424 -> 524 (23.58 %)
Spilled SGPRs: 84 -> 80 (-4.76 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Private memory VGPRs: 0 -> 0 (0.00 %)
Scratch size: 0 -> 0 (0.00 %) dwords per thread
Code Size: 52812 -> 52484 (-0.62 %) bytes
LDS: 135 -> 135 (0.00 %) blocks
Max Waves: 56 -> 53 (-5.36 %)
v2: consider WGP, rework to be clearer and apply the
"maximum 16 workgroups per CU" limit properly
v2: use "SIMD" instead of "EU"
v2: fix spiller by introducing "Program::max_waves"
v2: rename "lds_size" to "lds_limit"
v3: make max_waves actually independant of register usage
v3: fix issue where max_waves was way too high
v3: use DIV_ROUND_UP(a, b) instead of max(a / b, 1)
v3: rename "workgroups_per_cu" to "workgroups_per_cu_wgp"
v4: fix typo from "workgroups_per_cu" rename
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> (v3)
2019-10-23 19:11:21 +01:00
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
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
