mesa/src/amd/common
Timur Kristóf 8341af5109 radv, aco, ac/nir: Tweak position export scheduling for NGG culling.
The result is about +5-ish fps in Doom Eternal.

It turns out that the location of position exports matters more
than we thought, and it's actually better to keep them at the bottom
for culling shaders rather than schedule it up to the top.

Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10525>
2021-07-13 23:56:33 +00:00
..
ac_binary.c ac,radeonsi: use correct VGPR granularity on Aldebaran 2021-03-10 18:02:28 +00:00
ac_binary.h amd/common: switch to 3-spaces style 2020-09-07 10:00:29 +02:00
ac_debug.c ac/debug: fix color printing PKT3 when count in header is too low 2021-06-08 11:19:00 +02:00
ac_debug.h amd/common: switch to 3-spaces style 2020-09-07 10:00:29 +02:00
ac_drm_fourcc.h ac: fix win32 build 2021-07-06 14:49:24 +00:00
ac_exp_param.h amd/common: switch to 3-spaces style 2020-09-07 10:00:29 +02:00
ac_gpu_info.c ac,radv: implement the cs_regalloc_hang HW bug workaround 2021-07-09 13:37:37 +00:00
ac_gpu_info.h ac,radv: implement the cs_regalloc_hang HW bug workaround 2021-07-09 13:37:37 +00:00
ac_msgpack.c ac/rgp: add ac_msgpack.h/c 2021-02-19 14:29:46 +00:00
ac_msgpack.h ac/rgp: add ac_msgpack.h/c 2021-02-19 14:29:46 +00:00
ac_nir.c ac: move ac_lower_indirect_derefs() outside of the LLVM dir 2021-04-23 11:52:01 +02:00
ac_nir.h radv, aco, ac/nir: Tweak position export scheduling for NGG culling. 2021-07-13 23:56:33 +00:00
ac_nir_cull.c ac/nir: Add a NIR port of ac_llvm_cull. 2021-07-13 23:56:33 +00:00
ac_nir_lower_esgs_io_to_mem.c ac: Fix emit_split_buffer_store modulus operation. 2021-03-23 19:48:15 -07:00
ac_nir_lower_ngg.c radv, aco, ac/nir: Tweak position export scheduling for NGG culling. 2021-07-13 23:56:33 +00:00
ac_nir_lower_tess_io_to_mem.c ac/nir: Update TCS output barriers with nir_var_mem_shared. 2021-06-21 08:27:14 +00:00
ac_perfcounter.c ac/perfcounters: add a GPU block ID to every block definitions 2021-06-22 06:38:55 +00:00
ac_perfcounter.h ac/perfcounters: add a GPU block ID to every block definitions 2021-06-22 06:38:55 +00:00
ac_rgp.c ac: add ac_thread_trace::data 2021-06-03 15:39:34 +00:00
ac_rgp.h ac: add ac_thread_trace::data 2021-06-03 15:39:34 +00:00
ac_rgp_elf_object_pack.c ac/rgp: BSD elf library compatibility 2021-04-12 22:50:52 +00:00
ac_rtld.c ac: handle bigger instruction prefetch for Aldebaran 2021-03-10 18:02:27 +00:00
ac_rtld.h ac/rtld: make ac_rtld_upload returns the code size 2021-03-05 13:10:11 +00:00
ac_shader_args.c ac: add shader return values into ac_shader_args 2020-12-09 20:13:24 +00:00
ac_shader_args.h radv: Add sbt descriptors user SGPR input. 2021-05-18 18:29:36 +00:00
ac_shader_util.c ac,radeonsi: move late alloc computation into common code and shader states 2021-07-08 18:37:41 +00:00
ac_shader_util.h ac,radeonsi: move late alloc computation into common code and shader states 2021-07-08 18:37:41 +00:00
ac_shadowed_regs.c amd: Stub sections that don't have _WIN32 support 2020-11-27 20:49:00 -08:00
ac_shadowed_regs.h ac: add ifdef __cplusplus guard to header 2021-01-14 10:33:10 +01:00
ac_sqtt.c ac/rgp: move radv/sqtt functions to ac 2021-03-05 13:10:11 +00:00
ac_sqtt.h ac: add ac_thread_trace::data 2021-06-03 15:39:34 +00:00
ac_surface.c ac/surface: Handle non-retiled displayable DCC correctly for modifiers. 2021-07-05 22:34:13 +00:00
ac_surface.h ac/surface: document more meta equation dependencies 2021-04-26 22:53:30 +00:00
ac_surface_meta_address_test.c ac/surface: rename gfx9_dcc_equation to gfx9_meta_equation 2021-04-14 09:31:13 +00:00
ac_surface_modifier_test.c ac/surface/tests: fix the ARM build 2021-07-08 18:37:41 +00:00
ac_surface_test_common.h ac/surface/tests: fix RB counts 2021-06-20 01:22:01 -04:00
amd_family.c amd: add Beige Goby support 2021-05-24 17:41:34 +00:00
amd_family.h amd: add Beige Goby support 2021-05-24 17:41:34 +00:00
amd_kernel_code_t.h amd/common: switch to 3-spaces style 2020-09-07 10:00:29 +02:00
gfx10_format_table.h amd/common: switch to 3-spaces style 2020-09-07 10:00:29 +02:00
gfx10_format_table.py amd/registers: rename IMG_FORMAT to GFX10_FORMAT to disambiguate the meaning 2021-04-17 02:37:49 +00:00
meson.build ac/nir: Add a NIR port of ac_llvm_cull. 2021-07-13 23:56:33 +00:00
sid.h radv: add support for emitting PS_DONE/CS_DONE on GFX6-8 2021-01-26 09:26:58 +00:00
sid_tables.py amd: Work around MSVC limit for string literals 2020-11-27 20:49:00 -08:00