summaryrefslogtreecommitdiff
path: root/src/amd/compiler/aco_instruction_selection.cpp
AgeCommit message (Expand)AuthorFilesLines
2019-11-29aco: propagate p_wqm on an image_sample's coordinate p_create_vectorRhys Perry1-9/+12
2019-11-26aco: set dlc/glc correctly for image loadsRhys Perry1-0/+3
2019-11-25aco: Make unused workgroup id's 0Connor Abbott1-3/+3
2019-11-25aco: Use common argument handlingConnor Abbott1-75/+110
2019-11-25aco: Make num_workgroups and local_invocation_ids one argument eachConnor Abbott1-9/+13
2019-11-25aco: Split vector arguments at the beginningConnor Abbott1-1/+16
2019-11-25aco: Constify radv_nir_compiler_options in iselConnor Abbott1-1/+1
2019-11-19nir: move data.image.access to data.accessMarek Olšák1-2/+2
2019-11-15aco: implement VK_KHR_shader_float_controlsRhys Perry1-8/+135
2019-11-15aco: fix 64-bit fsign with 0Rhys Perry1-1/+1
2019-11-14aco: Treat all booleans as per-lane.Timur Kristóf1-233/+167
2019-11-14aco: use s_and_b64 exec to reduce uniform booleans to one bitDaniel Schürmann1-1/+3
2019-11-12aco: combine read_invocation and shuffle implementationsRhys Perry1-29/+7
2019-11-12aco: fix read_invocation with VGPR lane indexRhys Perry1-2/+1
2019-11-12aco: fix shuffle with uniform operandsRhys Perry1-6/+6
2019-11-06radv: determine shaders wavesize at pipeline levelSamuel Pitoiset1-2/+2
2019-10-30aco: always set scratch_offset in startpgmDaniel Schürmann1-11/+11
2019-10-30aco: remove potential critical edge on loops.Daniel Schürmann1-12/+21
2019-10-30aco: add can_reorder flags to load_ubo and load_constantDaniel Schürmann1-5/+9
2019-10-28aco: Introduce vgpr_limit to keep track of available VGPRs.Timur Kristóf1-0/+5
2019-10-28aco: Implement subgroup shuffle in GFX10 wave64 mode.Timur Kristóf1-6/+30
2019-10-22aco: add missing bld.scc()Rhys Perry1-1/+1
2019-10-22aco: use ds_read2_b64/ds_write2_b64Rhys Perry1-7/+24
2019-10-22aco: fix sparse store_lds()Rhys Perry1-8/+40
2019-10-22aco: create load_lds/store_lds helpersRhys Perry1-176/+195
2019-10-21aco: ensure that uniform booleans are computed in WQM if their uses happen in...Daniel Schürmann1-1/+2
2019-10-21aco/gfx10: Fix PS exports for SPI_SHADER_32_AR.Timur Kristóf1-1/+7
2019-10-16aco: emit_split_vector() s_memtime resultsRhys Perry1-0/+1
2019-10-16aco: don't CSE s_memtimeRhys Perry1-1/+1
2019-10-11aco: don't use p_as_uniform for vgpr sampler/image indicesRhys Perry1-1/+3
2019-10-11aco: implement divergent vulkan_resource_indexRhys Perry1-3/+13
2019-10-11aco: readfirstlane vgpr pointers in convert_pointer_to_64_bit()Rhys Perry1-0/+2
2019-10-10aco: rework scratch resource codeRhys Perry1-22/+28
2019-10-10aco/gfx10: disable GFX9 1D texture workaroundsRhys Perry1-5/+5
2019-10-10aco/gfx10: fix inline uniform blocksRhys Perry1-3/+9
2019-10-10aco: Set GFX10 dimensionality on the instructions that need it.Timur Kristóf1-0/+21
2019-10-10aco: Use ac_get_sampler_dim, delete duplicate code.Timur Kristóf1-44/+5
2019-10-10aco: Set GFX10 DLC bit properly.Timur Kristóf1-0/+17
2019-10-09aco: move s_andn2_b64 instructions out of the p_discard_ifRhys Perry1-0/+1
2019-10-04aco: fix load_constant with multiple arraysRhys Perry1-3/+3
2019-10-04radv/aco,aco: set lower_fmodRhys Perry1-29/+0
2019-09-28android: aco: fix undefined template 'std::__1::array' build errorsMauro Rossi1-0/+1
2019-09-27aco: set loop_info::has_discard for demotesRhys Perry1-2/+2
2019-09-25aco: implement 64-bit inegRhys Perry1-0/+16
2019-09-19aco: Initial commit of independent AMD compilerDaniel Schürmann1-0/+7621
generated by cgit v1.2.3 (git 2.39.1) at 2024-04-30 11:00:59 +0000