aboutsummaryrefslogtreecommitdiffstats
path: root/src/amd
Commit message (Expand)AuthorAgeFilesLines
* radv: add the trace BO to the BO list at submit timeSamuel Pitoiset2020-02-242-7/+4
* radv: use RADEON_FLAG_ZERO_VRAM when creating the trace BOSamuel Pitoiset2020-02-201-3/+2
* radv/winsys: add a new flag that requests zerovram allocationsSamuel Pitoiset2020-02-202-2/+6
* aco: improve GFX9 1D ddx/ddy assertionRhys Perry2020-02-201-1/+1
* aco: add some helpers for filling/testing register rangesRhys Perry2020-02-191-148/+106
* aco: add RegisterFileRhys Perry2020-02-191-12/+27
* radv: Squelch possibly-undefined warningEric Anholt2020-02-181-1/+1
* radv: Advertise VK_KHR_shader_non_semantic_infoCaio Marcelo de Oliveira Filho2020-02-181-0/+1
* radv: add a comment about VK_AMD_mixed_attachment_samples on GFX6-GFX7Samuel Pitoiset2020-02-171-0/+1
* radv: enable VK_NV_compute_shader_derivatives on GFX6-GFX7Samuel Pitoiset2020-02-171-1/+1
* radv: enable VK_EXT_sampler_filter_minmax on GFX6Samuel Pitoiset2020-02-172-2/+2
* radv: enable shaderStorageImageMultisample on GFX6-GFX7Samuel Pitoiset2020-02-171-2/+2
* radv: remove unnecessary RADV_DEBUG=nobatchchain optionSamuel Pitoiset2020-02-145-15/+11
* radv: remove unused RADV_HASH_SHADER_IS_GEOM_COPY_SHADERSamuel Pitoiset2020-02-131-6/+5
* radv: remove RADV_DEBUG=nosisched and RADV_PERFTEST=sischedSamuel Pitoiset2020-02-135-43/+32
* radv: remove LLVM sicheduler enable for The Talos PrincipleSamuel Pitoiset2020-02-131-9/+1
* radv: implement VK_EXT_line_rasterizationSamuel Pitoiset2020-02-135-7/+112
* radv: fix line width range and granularitySamuel Pitoiset2020-02-131-2/+2
* aco: improve SCC handling in some SALU combinesRhys Perry2020-02-121-6/+4
* aco: disable some instruction combining if it could change an exec operandRhys Perry2020-02-121-2/+13
* Rename nir_lower_constant_initializers to nir_lower_variable_initalizersArcady Goldmints-Orlov2020-02-121-3/+3
* aco: fix creating v_madak if v_mad_f32 has two sgpr literalsSamuel Pitoiset2020-02-111-1/+1
* radv: set the chip name to GCN-NOOP when RADV_FORCE_FAMILY is setSamuel Pitoiset2020-02-111-0/+1
* radv: make sure to not submit any IBs when RADV_FORCE_FAMILY is setSamuel Pitoiset2020-02-115-0/+9
* radv: Do not redundantly set the RB+ regs on pipeline switch.Bas Nieuwenhuizen2020-02-112-1/+19
* radv: Optimize emitting index buffer changes.Bas Nieuwenhuizen2020-02-111-2/+7
* aco: fix waiting for scalar stores before "writing back" data on GFX8-GFX9Samuel Pitoiset2020-02-101-6/+9
* aco: do not use ds_{read,write}2 on GFX6Samuel Pitoiset2020-02-071-4/+6
* aco: gfx10_wave64_bpermute reduce op to print_irRhys Perry2020-02-061-0/+1
* aco: fix gfx10_wave64_bpermuteRhys Perry2020-02-061-1/+6
* aco/optimizer: Don't combine uniform bool s_and to s_andn2.Timur Kristóf2020-02-051-0/+3
* radv: Do not set SX DISABLE bits for RB+ with unused surfaces.Bas Nieuwenhuizen2020-02-041-6/+7
* radv: Allow non-dedicated linear images and buffer.Bas Nieuwenhuizen2020-02-022-6/+13
* aco: fix image_atomic_cmp_swapDaniel Schürmann2020-01-312-2/+3
* aco: fix MUBUF VS input loads when expanding vec3 to vec4 on GFX6Samuel Pitoiset2020-01-311-1/+6
* radv: refactor physical device propertiesSamuel Pitoiset2020-01-291-250/+233
* aco: Fix the meaning of is_atomic.Timur Kristóf2020-01-292-7/+7
* aco: simplify gathering of MIMG address componentsDaniel Schürmann2020-01-291-129/+79
* aco: simplify adjust_sample_index_using_fmask() & get_image_coords()Daniel Schürmann2020-01-291-26/+27
* aco: fix register allocation with multiple live-range splitsDaniel Schürmann2020-01-291-7/+13
* aco: reorder VMEM operands in ACO IRDaniel Schürmann2020-01-2911-132/+153
* aco: fix exec mask consistency issuesRhys Perry2020-01-291-3/+7
* aco: parallelcopy exec mask before s_wqmRhys Perry2020-01-291-0/+7
* aco: fill reg_demand with sensible information in add_coupling_code()Rhys Perry2020-01-291-14/+22
* aco: improve assertion at the end of spillerRhys Perry2020-01-291-1/+1
* aco: set exec_potentially_empty after continues/breaks in nested IFsRhys Perry2020-01-292-15/+51
* aco: error when block has no logical preds but VGPRs are live at the startRhys Perry2020-01-291-1/+2
* aco: don't always add logical edges from continue_break blocks to headersRhys Perry2020-01-291-1/+2
* aco: only create parallelcopy to restore exec at loop exit if neededRhys Perry2020-01-291-2/+29
* aco: don't update demand in add_coupling_code() for loop headersRhys Perry2020-01-291-3/+9