Commit Graph

118 Commits

Author SHA1 Message Date
Bas Nieuwenhuizen
73749caf0e radv: calculate and emit GFX9 GS registers to pipeline state.
Reviewed-by: Dave Airlie <airlied@redhat.com>
2017-10-20 06:23:47 +01:00
Timothy Arceri
087e010b2b radv: copy indirect lowering settings from radeonsi
It looks the original indirect mask was probably copied from
ANV.

Sascha Willems demo results:

tessellation ~4000 -> ~4200 fps

V2: continue lowering local indirects due to llvm deficiencies.

Tested-by: Alex Smith <asmith@feralinteractive.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-10-20 08:01:26 +11:00
Bas Nieuwenhuizen
228325f4b7 radv: Modify rsrc1/rsrc2 generation for merged tess.
No OC_LDS_EN for HS, and the included LS vgpr_comp_cnt is at
a different offset.

Reviewed-by: Dave Airlie <airlied@redhat.com>
2017-10-19 22:25:44 +02:00
Bas Nieuwenhuizen
91b033f4f6 radv: Update GFX9 user data regs for GS/tess.
Reviewed-by: Dave Airlie <airlied@redhat.com>
2017-10-19 22:25:27 +02:00
Bas Nieuwenhuizen
ce03c119ce radv: Add code to compile merged shaders.
Reviewed-by: Dave Airlie <airlied@redhat.com>
2017-10-19 22:25:23 +02:00
Bas Nieuwenhuizen
a996ed1f9b ac/nir: Change interface to allow multiple source shaders.
Reviewed-by: Dave Airlie <airlied@redhat.com>
2017-10-19 22:24:47 +02:00
Bas Nieuwenhuizen
06f05040eb radv: Link shaders.
Here we make use of NIR the linking helpers to remove unused
varyings.

Sascha Willems demo results:

computecullandlod 39 -> 41 fps
pipelines ~6100 -> ~6200 fps

Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Signed-off-by: Timothy Arceri <tarceri@itsqueeze.com>
Acked-by: Dave Airlie <airlied@redhat.com>
2017-10-18 09:19:35 +11:00
Timothy Arceri
7664aaf331 radv: remove duplicate debug_flags field
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-10-12 08:52:38 +11:00
Dave Airlie
2c61594d84 radv: lower ffma in nir.
So it appears the Vulkan SPIR-V fma opcode can be equivalent to a
mad operation, and the fma hw opcode on AMD hw is issued like a double
opcode so is slower. Also the radeonsi stack does this.

This appears to improve performance on a number of games from Feral,
and thanks to Feral for noticing the problem.

I'm reposting this one as Marek indicated he thinks this is what
we should be doing on AMD hw.

Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2017-10-11 07:31:27 +10:00
Marek Olšák
7b697c8b78 amd: move r600d_common.h into r600g
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
2017-10-09 16:27:06 +02:00
Samuel Pitoiset
844ae722c4 radv: dump SPIRV when a GPU hang is detected
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-10-04 19:37:08 +02:00
Samuel Pitoiset
a2a350a3be radv: dump NIR when a GPU hang is detected
This looks a bit ugly to me, but the existing codepath
is not terribly elegant as well.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-10-04 19:37:08 +02:00
Samuel Pitoiset
80b8d9f7e7 radv: add radv_shader_dump_stats() helper
To dump the shader stats when a hang is detected.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-09-14 10:37:57 +02:00
Dave Airlie
64d9bd149a radv/nir: call opt_remove_phis after trivial continues.
With the shaders in the ssao demo, the nir_opt_if wasn't
working properly without this, after this the if gets optimised
so that loop unrolling gets called.

(loop unrolling fails due to instruction count, but at least
it gets to do that.)

Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Cc: "17.2" <mesa-stable@lists.freedesktop.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
2017-09-13 21:13:03 +01:00
Samuel Pitoiset
885d75760b radv: keep track of the disasm string in debug mode only
This will allow to dump the active shaders when a hang is
detected. Only the ASM will be dumped for now.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-09-08 17:18:17 +02:00
Samuel Pitoiset
92db23f3f9 radv: add shader_variant_create() helper function
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-09-08 17:17:40 +02:00
Samuel Pitoiset
47efc5264a radv: drop 'dump' parameters from some shader related functions
The device object contains the debug flags.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-09-08 17:17:40 +02:00
Samuel Pitoiset
d4d777317b radv: move shaders related code to radv_shader.c
Reduce size of radv_pipeline.c and improve code isolation. More
code can probably moved but it's a start.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2017-09-08 17:17:40 +02:00