summaryrefslogtreecommitdiff
path: root/src/amd/compiler
AgeCommit message (Expand)AuthorFilesLines
2020-05-21aco: sign-extend the input and identity for 8-bit subgroup operationsSamuel Pitoiset1-0/+12
2020-05-21aco: use a temporary SGPR for 8-bit/16-bit literal reduction identitiesSamuel Pitoiset1-3/+5
2020-05-21aco: implement 8-bit/16-bit nir_intrinsic_quad_*Samuel Pitoiset1-0/+16
2020-05-21aco: implement 8-bit/16-bit nir_intrinsic_{shuffle,_read_invocation}Samuel Pitoiset1-1/+8
2020-05-21aco: implement 8-bit/16-bit nir_intrinsic_read_first_invocationSamuel Pitoiset1-1/+1
2020-05-21aco: validate 8-bit/16-bit VGPR operands for readfirstlane/readlane/writelaneSamuel Pitoiset1-1/+1
2020-05-21aco: implement 8-bit/16-bit reductionsSamuel Pitoiset4-24/+76
2020-05-21aco: declare 8-bit/16-bit reduce operationsSamuel Pitoiset2-13/+39
2020-05-21aco: fix typo in insert_waitcnt's kill()Rhys Perry1-1/+1
2020-05-21aco: don't allow unaligned subdword accesses on GFX6/7Daniel Schürmann2-16/+20
2020-05-21aco: fix corner case in register allocationDaniel Schürmann1-2/+2
2020-05-21aco: don't move create_vector subdword operands to unsupported register offsetsDaniel Schürmann1-1/+1
2020-05-21aco: restrict copying of create_vector operands to GFX9+Daniel Schürmann1-15/+26
2020-05-21aco: fix missing break in label_instruction()Samuel Pitoiset1-0/+1
2020-05-19aco: fix off-by-one error with 16-bit MTBUF opcodes on GFX10Samuel Pitoiset1-1/+1
2020-05-19aco: implement 16-bit interpSamuel Pitoiset1-4/+34
2020-05-19aco: emit v_interp_*_f16 instructions as VOP3 instead of VINTRPSamuel Pitoiset1-14/+42
2020-05-19aco: validate v_interp_*_f16 as VOP3 instructions instead of VINTRPSamuel Pitoiset1-2/+13
2020-05-19aco: implement 16-bit vertex fetches with tbuffer_load_format_d16_*Samuel Pitoiset1-17/+39
2020-05-19aco: implement 8-bit/16-bit mov's with p_create_vectorSamuel Pitoiset1-6/+8
2020-05-19aco: allow to load/store 16-bit values in VMEM for tess and geomSamuel Pitoiset1-3/+3
2020-05-19aco: convert 16-bit values before exporting MRTsSamuel Pitoiset1-2/+39
2020-05-19aco: store 16-bit temporary outputs as v2bSamuel Pitoiset1-1/+3
2020-05-19aco: fix disassembly with LLVM 11Rhys Perry1-0/+7
2020-05-14aco: split operations that use a swap's definitionRhys Perry2-7/+46
2020-05-14aco: fix WQM coalescingDaniel Schürmann1-1/+3
2020-05-14aco: add support for texturing with clamped LODSamuel Pitoiset1-3/+39
2020-05-14aco: remove useless check for nir_tex_src_biasSamuel Pitoiset1-4/+2
2020-05-13nir: Make "divergent" a property of an SSA valueJason Ekstrand2-31/+28
2020-05-13aco: improve phi affinities with p_split_vectorRhys Perry1-0/+19
2020-05-13aco: consider affinities when creating v_mac_f32Rhys Perry1-2/+8
2020-05-13aco: mark phi definitions as last-seen phi operandsRhys Perry1-14/+14
2020-05-13aco: fix consecutively written vgprs from vmem instructionsRhys Perry1-10/+26
2020-05-13aco: simplify consecutive ordered vmem/lds writes optimizationRhys Perry1-10/+2
2020-05-12aco: optimize add/sub(a, cndmask(b, 0, 1, cond)) -> addc/subbrev_co(0, a, b)Samuel Pitoiset1-13/+79
2020-05-12aco: coalesce parallelcopies during register allocationDaniel Schürmann1-6/+3
2020-05-11aco: prevent invalid loads/stores vectorization if robustness is enabledSamuel Pitoiset1-2/+10
2020-05-11nir: do not vectorize load/store if offset can overflow and robustness enabledSamuel Pitoiset1-1/+2
2020-05-11aco: fix 64-bit trunc with negative exponents on GFX6Samuel Pitoiset1-2/+3
2020-05-07aco: either copy-propagate or inline create_vector operandsDaniel Schürmann1-4/+6
2020-05-05aco: remove unecessary p_split_vector with v2b reg classSamuel Pitoiset1-80/+33
2020-05-02aco: add message to static_assertRhys Perry1-25/+25
2020-05-02aco: remove use of f-stringsRhys Perry1-1/+1
2020-04-29radv: Use smaller esgs_itemsize for ACO.Timur Kristóf1-8/+0
2020-04-29aco: Use new default driver locations.Timur Kristóf1-107/+17
2020-04-29aco: Set config->lds_size when TES or VS is running on HW ESGS.Timur Kristóf1-0/+1
2020-04-29aco: Calculate workgroup size of legacy GS.Timur Kristóf1-1/+5
2020-04-29aco: Remember VS/TCS output driver locations.Timur Kristóf2-10/+18
2020-04-29aco: Use context variables instead of calculating TCS inputs/outputs.Timur Kristóf2-11/+7
2020-04-29radv: Refactor calculate_tess_lds_size and get_tcs_num_patches.Timur Kristóf1-4/+9
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-24 16:52:44 +0000