aco: improve code for f2{i,u}{8,16}
[mesa.git] / src / amd / compiler /
2020-06-15 Rhys Perryaco: improve code for f2{i,u}{8,16}
2020-06-15 Rhys Perryaco: use num_opcodes instead of last_opcode
2020-06-12 Daniel Schürmannaco: fix shared subdword loads
2020-06-12 Samuel Pitoisetaco: implement radv_enable_mrt_output_nan_fixup workaround
2020-06-10 Rhys Perryaco: fix moving sub-dword values out of a register...
2020-06-10 Rhys Perryaco: use Info::definition_size instead of definition...
2020-06-10 Rhys Perryaco: add Info::{operand_size,definition_size}
2020-06-10 Rhys Perryaco: prefer 4-byte aligned definitions
2020-06-10 Rhys Perryaco: allow reading/writing upper halves/bytes when...
2020-06-10 Rhys Perryaco: p_extract_vector in 64-bit u2f16/i2f16
2020-06-10 Rhys Perryaco: validate instructions reading/writing upper halves...
2020-06-09 Daniel Schürmannaco: remove unnecessary split- and create_vector instru...
2020-06-09 Samuel Pitoisetaco: fix alignment of vectors with 4 elements
2020-06-09 Samuel Pitoisetaco: implement 8-bit/16-bit conversions on GFX6-GFX7
2020-06-09 Daniel Schürmannaco: optimize packing of 16bit subdword registers on...
2020-06-09 Daniel Schürmannaco: skip partial copies on first iteration when loweri...
2020-06-09 Daniel Schürmannaco: coalesce copies more aggressively when lowering...
2020-06-09 Daniel Schürmannaco: add and use scratch SGPR to lower subdword p_creat...
2020-06-09 Daniel Schürmannaco: adjust GFX6 subdword lowering workarounds for...
2020-06-09 Daniel Schürmannaco: Workarounds subdword lowering on GFX6/7
2020-06-09 Daniel Schürmannaco: use full-register instructions to implement subdwo...
2020-06-09 Daniel Schürmannaco: simplify statistics collection for copies
2020-06-09 Daniel Schürmannaco: fix register assignment for p_create_vector on...
2020-06-08 Rhys Perryaco: use v_xor3_b32
2020-06-05 Samuel Pitoisetaco: fix nir_intrinsic_quad_* with 8-bit in GFX6-GFX7
2020-06-05 Samuel Pitoisetaco: fix sign-extend 8-bit subgroup operations on GFX6...
2020-06-05 Samuel Pitoisetaco: use v_bfe_u32 for unsigned reductions sign-extensi...
2020-06-03 Samuel Pitoisetaco: sign-extend input/identity for 16-bit subgroup...
2020-06-03 Samuel Pitoisetaco: fix subdword copies on GFX6-GFX7
2020-06-03 Samuel Pitoisetaco: implement 16-bit nir_intrinsic_quad_* on GFX6...
2020-06-03 Samuel Pitoisetaco: implement 16-bit reduce operations on GFX6-GFX7
2020-06-02 Oschowaaco: Use correct reference type in for-range-loop.
2020-06-02 Oschowaaco: Don't std::move temporary object.
2020-06-02 Oschowaaco: Don't declare 'Block' as class, but define as...
2020-06-02 Timur Kristófaco: Implement subgroup shuffle on GFX6-7.
2020-06-02 Timur Kristófaco/gfx10: Refactor of GFX10 wave64 bpermute.
2020-06-01 Dylan Bakermeson: use gnu_symbol_visibility argument
2020-05-29 Samuel Pitoisetaco: sign-extend input/indentity for 32-bit reduce...
2020-05-29 Samuel Pitoisetaco: allow gfx10_wave64_bpermute with 8-bit/16-bit...
2020-05-29 Samuel Pitoisetaco: allocate a temp VGPR for some 8-bit/16-bit reducti...
2020-05-29 Samuel Pitoisetaco: implement 8-bit/16-bit reductions on GFX10
2020-05-29 Samuel Pitoisetaco: fix register allocation for subdword instructions...
2020-05-28 Rhys Perryaco: fix 64-bit shared_atomic_exchange
2020-05-28 Rhys Perryaco: don't reorder barriers in the scheduler
2020-05-28 Rhys Perryaco: preserve more fields when combining additions...
2020-05-28 Rhys Perryaco: check instruction format before waiting for a...
2020-05-28 Rhys Perryaco: consider SDWA during value numbering
2020-05-28 Rhys Perryaco: fix interaction with 3f branch workaround and...
2020-05-25 Samuel Pitoisetaco: add support for bias/lod with texture gather
2020-05-24 Samuel Pitoisetaco: implement nir_intrinsic_shader_clock with device...
2020-05-21 Samuel Pitoisetaco: sign-extend the input and identity for 8-bit subgr...
2020-05-21 Samuel Pitoisetaco: use a temporary SGPR for 8-bit/16-bit literal...
2020-05-21 Samuel Pitoisetaco: implement 8-bit/16-bit nir_intrinsic_quad_*
2020-05-21 Samuel Pitoisetaco: implement 8-bit/16-bit nir_intrinsic_{shuffle...
2020-05-21 Samuel Pitoisetaco: implement 8-bit/16-bit nir_intrinsic_read_first_in...
2020-05-21 Samuel Pitoisetaco: validate 8-bit/16-bit VGPR operands for readfirstl...
2020-05-21 Samuel Pitoisetaco: implement 8-bit/16-bit reductions
2020-05-21 Samuel Pitoisetaco: declare 8-bit/16-bit reduce operations
2020-05-21 Rhys Perryaco: fix typo in insert_waitcnt's kill()
2020-05-21 Daniel Schürmannaco: don't allow unaligned subdword accesses on GFX6/7
2020-05-21 Daniel Schürmannaco: fix corner case in register allocation
2020-05-21 Daniel Schürmannaco: don't move create_vector subdword operands to...
2020-05-21 Daniel Schürmannaco: restrict copying of create_vector operands to...
2020-05-21 Samuel Pitoisetaco: fix missing break in label_instruction()
2020-05-19 Samuel Pitoisetaco: fix off-by-one error with 16-bit MTBUF opcodes...
2020-05-19 Samuel Pitoisetaco: implement 16-bit interp
2020-05-19 Samuel Pitoisetaco: emit v_interp_*_f16 instructions as VOP3 instead...
2020-05-19 Samuel Pitoisetaco: validate v_interp_*_f16 as VOP3 instructions inste...
2020-05-19 Samuel Pitoisetaco: implement 16-bit vertex fetches with tbuffer_load_...
2020-05-19 Samuel Pitoisetaco: implement 8-bit/16-bit mov's with p_create_vector
2020-05-19 Samuel Pitoisetaco: allow to load/store 16-bit values in VMEM for...
2020-05-19 Samuel Pitoisetaco: convert 16-bit values before exporting MRTs
2020-05-19 Samuel Pitoisetaco: store 16-bit temporary outputs as v2b
2020-05-19 Rhys Perryaco: fix disassembly with LLVM 11
2020-05-14 Rhys Perryaco: split operations that use a swap's definition
2020-05-14 Daniel Schürmannaco: fix WQM coalescing
2020-05-14 Samuel Pitoisetaco: add support for texturing with clamped LOD
2020-05-14 Samuel Pitoisetaco: remove useless check for nir_tex_src_bias
2020-05-13 Jason Ekstrandnir: Make "divergent" a property of an SSA value
2020-05-13 Rhys Perryaco: improve phi affinities with p_split_vector
2020-05-13 Rhys Perryaco: consider affinities when creating v_mac_f32
2020-05-13 Rhys Perryaco: mark phi definitions as last-seen phi operands
2020-05-13 Rhys Perryaco: fix consecutively written vgprs from vmem instructions
2020-05-13 Rhys Perryaco: simplify consecutive ordered vmem/lds writes optim...
2020-05-12 Samuel Pitoisetaco: optimize add/sub(a, cndmask(b, 0, 1, cond)) -...
2020-05-12 Daniel Schürmannaco: coalesce parallelcopies during register allocation
2020-05-11 Samuel Pitoisetaco: prevent invalid loads/stores vectorization if...
2020-05-11 Samuel Pitoisetnir: do not vectorize load/store if offset can overflow...
2020-05-11 Samuel Pitoisetaco: fix 64-bit trunc with negative exponents on GFX6
2020-05-07 Daniel Schürmannaco: either copy-propagate or inline create_vector...
2020-05-05 Samuel Pitoisetaco: remove unecessary p_split_vector with v2b reg...
2020-05-02 Rhys Perryaco: add message to static_assert
2020-05-02 Rhys Perryaco: remove use of f-strings
2020-04-29 Timur Kristófradv: Use smaller esgs_itemsize for ACO.
2020-04-29 Timur Kristófaco: Use new default driver locations.
2020-04-29 Timur Kristófaco: Set config->lds_size when TES or VS is running...
2020-04-29 Timur Kristófaco: Calculate workgroup size of legacy GS.
2020-04-29 Timur Kristófaco: Remember VS/TCS output driver locations.
2020-04-29 Timur Kristófaco: Use context variables instead of calculating TCS...
2020-04-29 Timur Kristófradv: Refactor calculate_tess_lds_size and get_tcs_num_...
next