aco: Fix signedness compare warning.
[mesa.git] / src / amd / compiler /
2019-12-04 Timur Kristófaco/wave32: Allow setting the subgroup ballot size...
2019-12-04 Timur Kristófaco/wave32: Use wave_size for barrier intrinsic.
2019-12-04 Timur Kristófaco/wave32: Fix load_local_invocation_index to support...
2019-12-04 Timur Kristófaco/wave32: Use lane mask regclass for exec/vcc.
2019-12-04 Timur Kristófaco/wave32: Add wave size specific opcodes to aco_builder.
2019-12-04 Timur Kristófaco/wave32: Introduce emit_mbcnt which takes wave size...
2019-12-04 Timur Kristófaco/wave32: Replace hardcoded numbers in spiller with...
2019-12-04 Timur Kristófaco/wave32: Change uniform bool optimization to work...
2019-12-04 Timur Kristófaco: Optimize load_subgroup_id to one bit field extract...
2019-12-04 Timur Kristófaco: Remove lower_linear_bool_phi, it is not needed...
2019-12-04 Timur Kristófaco: Remove superfluous argument from emit_boolean_logic.
2019-12-04 Timur Kristófaco: Fix operand of s_bcnt1_i32_b64 in emit_boolean_reduce.
2019-12-02 Rhys Perryaco: fix block_kind_discard s_andn2 definition to exec
2019-12-02 Rhys Perryaco: handle loop exit and IF merge phis with break...
2019-12-02 Rhys Perryaco: validate the CFG
2019-11-29 Samuel Pitoisetaco: drop useless lowering of deref operations for...
2019-11-29 Daniel Schürmannaco: fix a couple of value numbering issues
2019-11-29 Daniel Schürmannaco: don't split live-ranges of linear VGPRs
2019-11-29 Rhys Perryaco: implement global atomics
2019-11-29 Rhys Perryaco: improve FLAT/GLOBAL scheduling
2019-11-29 Rhys Perryaco: don't enable store_global for helper invocations
2019-11-29 Rhys Perryaco: fix SADDR with FLAT on GFX10
2019-11-29 Rhys Perryaco: fix assembly of FLAT/GLOBAL atomics
2019-11-29 Rhys Perryaco: fix GFX10 opcodes for some global/flat atomics
2019-11-29 Rhys Perryaco: improve WAR hazard workaround with >64bit stores
2019-11-29 Rhys Perryaco: add v_nop inbetween exec write and VMEM/DS/FLAT
2019-11-29 Rhys Perryaco: fix incorrect cast in parse_wait_instr()
2019-11-29 Rhys Perryaco: fix i2i64
2019-11-29 Rhys Perryaco: propagate p_wqm on an image_sample's coordinate...
2019-11-26 Daniel Schürmannaco: don't value-number instructions from within a...
2019-11-26 Rhys Perryaco: set dlc/glc correctly for image loads
2019-11-26 Rhys Perryaco: allow constant offsets for global/scratch instruct...
2019-11-25 Rhys Perryaco: enable load/store vectorizer
2019-11-25 Connor Abbottaco: Make unused workgroup id's 0
2019-11-25 Connor Abbottaco: Use common argument handling
2019-11-25 Connor Abbottradv: Replace supports_spill with explict_scratch_args
2019-11-25 Connor Abbottaco: Make num_workgroups and local_invocation_ids one...
2019-11-25 Connor Abbottaco: Split vector arguments at the beginning
2019-11-25 Connor Abbottaco: Use radv_shader_args in aco_compile_shader()
2019-11-25 Connor Abbottaco: Constify radv_nir_compiler_options in isel
2019-11-22 Rhys Perryaco: fix waitcnts for barriers at block ends
2019-11-21 Rhys Perryaco: fix copy+paste error
2019-11-21 Rhys Perryaco: improve waitcnt insertion around loops
2019-11-19 Marek Olšáknir: move data.image.access to data.access
2019-11-19 Rhys Perryaco: implement 64-bit integer reductions
2019-11-19 Rhys Perryaco: refactor reduction lowering helpers
2019-11-15 Rhys Perryaco: implement VK_KHR_shader_float_controls
2019-11-15 Rhys Perryaco: fix 64-bit fsign with 0
2019-11-15 Rhys Perryaco: don't combine literals into v_cndmask_b32/v_subb...
2019-11-14 Timur Kristófaco: Optimize out trivial code from uniform bools.
2019-11-14 Timur Kristófaco: Treat all booleans as per-lane.
2019-11-14 Daniel Schürmannaco: use s_and_b64 exec to reduce uniform booleans...
2019-11-14 Timur Kristófaco: Make sure not to mistakenly propagate 64-bit const...
2019-11-14 Daniel Schürmannaco: value number instructions using the execution...
2019-11-14 Daniel Schürmannaco: check if SALU instructions are predeceeded by...
2019-11-12 Rhys Perryaco: combine read_invocation and shuffle implementations
2019-11-12 Rhys Perryaco: don't propagate vgprs into v_readlane/v_writelane
2019-11-12 Rhys Perryaco: fix read_invocation with VGPR lane index
2019-11-12 Rhys Perryaco: fix shuffle with uniform operands
2019-11-12 Rhys Perryaco: use DPP instead of exec modification when lowering...
2019-11-12 Daniel Schürmannaco: rematerialize s_movk instructions
2019-11-12 Daniel Schürmannaco: preserve kill flag on moved operands during RA
2019-11-12 Daniel Schürmannaco: fix invalid access on Pseudo_instructions
2019-11-08 Timur Kristófac: Handle invalid GFX10 format correctly in ac_get_tbu...
2019-11-08 Rhys Perryaco: add Instruction::usesModifiers() and add more...
2019-11-07 Rhys Perryradv: adjust loop unrolling heuristics for int64
2019-11-07 Daniel Schürmannaco: workaround Tonga/Iceland hardware bug
2019-11-06 Samuel Pitoisetradv: determine shaders wavesize at pipeline level
2019-11-04 Daniel Schürmannaco: fix accidential reordering of instructions when...
2019-11-04 Daniel Schürmannaco: only use single-dword loads/stores for spilling
2019-11-04 Daniel Schürmannaco: fix immediate offset for spills if scratch is...
2019-10-30 Daniel Schürmannaco: implement VGPR spilling
2019-10-30 Daniel Schürmannaco: always set scratch_offset in startpgm
2019-10-30 Daniel Schürmannaco: omit linear VGPRs as spill variables
2019-10-30 Daniel Schürmannaco: ensure that spilled VGPR reloads are done after...
2019-10-30 Daniel Schürmannaco: simplify calculation of target register pressure...
2019-10-30 Rhys Perryaco: fix new_demand calculation for first instructions
2019-10-30 Daniel Schürmannaco: don't add interferences between spilled phi operands
2019-10-30 Daniel Schürmannaco: consider loop_exit blocks like merge blocks, even...
2019-10-30 Daniel Schürmannaco: don't insert the exec mask into set of live-out...
2019-10-30 Daniel Schürmannaco: fix transitive affinities of spilled variables
2019-10-30 Daniel Schürmannaco: fix live-range splits of phis
2019-10-30 Daniel Schürmannaco: remove potential critical edge on loops.
2019-10-30 Daniel Schürmannaco: improve live variable analysis
2019-10-30 Daniel Schürmannaco: Lower to CSSA
2019-10-30 Rhys Perryaco: rename README to README.md
2019-10-30 Rhys Perryaco: a couple loop handling fixes for GFX10 hazard...
2019-10-30 Rhys Perryaco: try to group together VMEM loads of the same resource
2019-10-30 Daniel Schürmannaco: don't schedule instructions through depending...
2019-10-30 Daniel Schürmannaco: add can_reorder flags to load_ubo and load_constant
2019-10-30 Daniel Schürmannaco: only skip RAR dependencies if the variable is...
2019-10-30 Daniel Schürmannaco: restrict scheduling depending on max_waves
2019-10-28 Timur Kristófaco: Introduce vgpr_limit to keep track of available...
2019-10-28 Timur Kristófaco: Implement subgroup shuffle in GFX10 wave64 mode.
2019-10-28 Rhys Perryaco: Remove dead code in reduction lowering.
2019-10-28 Rhys Perryaco: Fix reductions on GFX10.
2019-10-25 Timur Kristófaco: Refactor hazard mitigations, separate pass for...
2019-10-25 Timur Kristófaco/gfx10: Fix mitigation of VMEMtoScalarWriteHazard.
2019-10-25 Timur Kristófaco/gfx10: Mitigate LdsBranchVmemWARHazard.
2019-10-25 Timur Kristófaco/gfx10: Mitigate SMEMtoVectorWriteHazard.
next