summaryrefslogtreecommitdiff
AgeCommit message (Expand)AuthorFilesLines
2024-02-27nir: Optimize uniform iadd, fadd, and ixor reduction operationsIan Romanick3-16/+98
2023-12-08intel/fs: Don't add MOV instructions to DO blocks in combine constantsIan Romanick1-4/+59
2023-09-21nir/rematerialize: Rematerialize ALUs used only by compares with zeroIan Romanick1-0/+154
2023-08-29intel/fs: New VGRF packing scheme for constant combiningIan Romanick2-13/+184
2023-08-29intel/fs: Combine constants for integer instructions tooIan Romanick2-23/+60
2023-08-29intel/fs: Combine constants for SEL instructions tooIan Romanick2-24/+140
2023-07-25intel/fs: Constant fold SHLIan Romanick2-1/+44
2023-06-14nir/algebraic: Simplify various trivial bfiIan Romanick2-0/+50
2023-06-14nir/algebraic: Optimize some u2f of bfiIan Romanick1-0/+9
2023-06-14intel/fs: Emit better code for bfi(..., 0)Ian Romanick1-1/+16
2023-06-14intel/fs: Use nir_opt_reassociate_bfiIan Romanick1-0/+7
2023-04-06nir/loop_analyze: Determine iteration counts for more kinds of loopsYevhenii Kolesnikov4-42/+307
2023-03-29nir/algebraic: Undistribute fsat from fmaxIan Romanick1-0/+1
2023-02-17glsl: copy prop vars before scalarizing alusTimothy Arceri1-2/+6
2022-12-14nir/builder: Handle i2b conversions specially in nir_type_convertIan Romanick1-6/+34
2022-12-02hasvk: Drop CCS_E supportJason Ekstrand7-637/+22
2022-11-09intel/compiler: Fix signed integer range analysis of imax and iminIan Romanick1-2/+4
2022-11-08intel/compiler: Signed integer range analysis for imul_32x16 generationIan Romanick1-7/+175
2022-11-08intel/compiler: Add and use a pass to generate imul_32x16 instructionsIan Romanick4-0/+126
2022-11-07anv: Rip out shadow surfacesJason Ekstrand4-307/+1
2022-09-27intel/compiler: Use an existing URB write to end TCS threads when viableKenneth Graunke2-23/+61
2022-09-08glsl: remove GLSL IR inverse comparison optimisationsTimothy Arceri1-30/+0
2022-07-26intel/fs: Make logical URB write instructions more like other logical instruc...Ian Romanick7-121/+176
2022-07-19anv: Do depth/stencil optimization for dynamic depth/stencilJason Ekstrand4-133/+100
2022-07-19anv: Stop recording sample locations per-sample-countJason Ekstrand5-85/+67
2022-07-08intel/fs: Lower URB messages to SENDIan Romanick3-10/+74
2022-05-16nir: Use nir_vector_extract to generate code for ir_binop_vector_extractIan Romanick2-9/+3
2022-04-07intel/fs: Better handle constant sources of FS_OPCODE_PACK_HALF_2x16_SPLITIan Romanick2-10/+21
2022-01-19iris: Do primitive ID overrides in 3DSTATE_SBE not SBE_SWIZKenneth Graunke1-12/+13
2021-10-28iris: Set MOCS on 3DSTATE_CONSTANT_XS on Gfx9+Kenneth Graunke1-1/+3
2021-10-28intel/genxml: Change 3DSTATE_CONSTANT_XS::MOCS to be MBZ on Gfx8.Kenneth Graunke1-5/+5
2021-08-03nir: add heuristic for instructions in loops with GCMTimothy Arceri1-22/+101
2021-07-21intel/compiler: Use GCM in nir_optimizeTimothy Arceri1-0/+1
2021-06-08anv: Make use_softpin compile-time in genX codeJason Ekstrand3-11/+32
2021-05-20nir/algebraic: Rearrange some logic-joined comparisons and reduceIan Romanick1-0/+10
2021-05-20nir/algebraic: Invert comparisons less oftenIan Romanick6-15/+11
2021-03-04nir/algebraic: Apply addition property of equality more conservativelyIan Romanick1-10/+10
2021-02-08nir/algebraic: add patterns for a >> #b << #b and a << #b >> #bIan Romanick1-0/+4
2020-05-11nir/algebraic: Eliminate useless extract before unpackIan Romanick1-0/+5
2020-05-11nir/algebraic: Add some half packing optimizations for pack_half_2x16_splitIan Romanick1-0/+7
2020-05-07nir/algebraic: Optimize some bfe patternsIan Romanick1-0/+18
2020-05-07nir/algebraic: Split ibfe and ubfe with two constant sourcesIan Romanick3-0/+36
2020-05-07nir/algebraic: Recognize open-coded byte or word extract from bfeIan Romanick1-7/+21
2020-04-01nir/algebraic: Distribute source modifiers into instructionsIan Romanick3-0/+43
2020-03-09intel/fs: Do cmod prop again after schedulingIan Romanick1-0/+11
2019-10-14iris: Implement the Gen < 9 tessellation quads workaroundKenneth Graunke1-0/+3
2019-10-08iris: Implement the Broadwell NP Z PMA Stall FixKenneth Graunke5-2/+209
2019-09-19nir/algebraic: Additional D3D Boolean optimizationIan Romanick1-0/+1
2019-08-29nir/range-analysis: Adjust result range of multiplication to account for flus...Ian Romanick1-31/+22
2019-08-20iris: Enable non coherent framebuffer fetch on broadwellSagar Ghuge3-4/+3
generated by cgit v1.2.3 (git 2.39.1) at 2024-04-27 03:15:08 +0000