Lines Matching full:nir
101 - [regression][bisected] "nir: Copy propagation between blocks"
182 - [nir] src/compiler/nir/nir_loop_analyze.c uninitialized variable
192 - [regression][bisected] "nir: Switch to using 1-bit Booleans for
257 - nir: remove unused variable
258 - nir/xfb: don't assert when xfb_buffer/stride is present but not
260 - nir/xfb: distinguish array of structs vs array of blocks
378 - nir: Account for atomics in copy propagation.
415 - nir: properly clear the entry sources in copy_prop_vars
416 - nir: properly find the entry to keep in copy_prop_vars
417 - nir: add a way to print the deref chain
418 - nir: remove dead code from copy_prop_vars
419 - nir: fix warning in nir_lower_io.c
423 - nir: check NIR_SKIP to skip passes by name
446 - nir: add lowering for ffloor
456 - nir/algebraic: Rewrite bit-size inference
457 - nir/algebraic: Add unit tests for bitsize validation
458 - nir: Fixup algebraic test for variable-sized conversions
467 - nir: add if opt opt_if_loop_last_continue()
484 - nir: move getting deref from var after we check deref type.
485 - nir: handle shared pointers in lowering indirect derefs.
543 - automake: Add include dir for nir src directory
551 - Revert "nir/opt_peephole_select: Don't peephole_select expensive math
740 - nir: Add some more consts to the nir_format_convert.h helpers.
741 - nir: Pull some of intel's image load/store format conversion to
744 - mesa/st: Expose compute shaders when NIR support is advertised.
745 - nir: Print the format of image variables.
748 - nir: Move intel's half-float image store lowering to to nir_format.h.
772 - nir: Fix clamping of uints for image store lowering.
819 - nir: Allow nir_format_unpack_int/sint to unpack larger values.
820 - nir: Add nir_lower_tex options to lower sampler return formats.
822 - nir: Add nir_lower_tex support for Broadcom's swizzled TG4 results.
824 - nir: Make nir_deref_instr_build/get_const_offset actually use
1040 - nir: Allow to skip integer ops in nir_lower_to_source_mods
1105 - nir/constant_folding: fix incorrect bit-size check
1106 - nir/from_ssa: fix bit-size of temporary register
1107 - Revert "nir/builder: Assert that intN_t immediates fit"
1114 - compiler/nir: add a nir_b2f() helper
1115 - compiler/nir: add nir_fadd_imm() and nir_fmul_imm() helpers
1125 - anv/pipeline_cache: fix incorrect guards for NIR cache
1126 - anv/pipeline_cache: free NIR shader cache
1136 - nir: Add a saturated unsigned integer add opcode
1138 - nir/phi_builder: Internal users should use
1142 - nir/phi_builder: Use per-value hash table to store [block] -> def
1144 - nir: Fix holes in nir_instr
1145 - nir: Release per-block metadata in nir_sweep
1151 - nir/opt_peephole_select: Don't try to remove flow control around
1154 - nir/opt_peephole_select: Don't peephole_select expensive math
1157 - Revert "nir/lower_indirect: Bail early if modes == 0"
1158 - nir/algebraic: Don't put quotes around floating point literals
1160 - nir: Silence zillions of unused parameter warnings in release builds
1226 - nir: Add a read_mask helper for ALU instructions
1231 - nir/lower_alu_to_scalar: Don't try to lower unpack_32_2x16
1232 - nir/builder: Assert that intN_t immediates fit
1233 - nir/builder: Add iadd_imm and imul_imm helpers
1234 - nir/builder: Add a nir_pack/unpack/bitcast helpers
1235 - nir/spirv: Force 32-bit for UBO and SSBO Booleans
1236 - nir/glsl: Force 32-bit for UBO and SSBO Booleans
1237 - nir/lower_io: Add shared to get_io_offset_src
1238 - nir: Add alignment parameters to SSBO, UBO, and shared access
1239 - intel/compiler: Lower SSBO and shared loads/stores in NIR
1240 - intel,nir: Move gl_LocalInvocationID lowering to
1246 - anv/nir: Rework arguments to apply_pipeline_layout
1247 - nir/derefs: Add a nir_derefs_do_not_alias enum value
1249 - nir/opcodes: Pull in the type helpers from constant_expressions
1250 - nir/opcodes: Rename tbool to tbool32
1251 - nir/algebraic: Clean up some \__str_\_ cruft
1252 - nir/algebraic: Refactor codegen a bit
1253 - nir/algebraic: Add support for unsized conversion opcodes
1254 - nir/opt_algebraic: Simplify an optimization using the new search ops
1255 - nir/opt_algebraic: Drop bit-size suffixes from conversions
1256 - nir/opt_algebraic: Add 32-bit specifiers to a bunch of booleans
1257 - nir: Make boolean conversions sized just like the others
1261 - nir/lower_tex: Simplify lower_gradient logic
1262 - nir/lower_tex: Modify txd instructions instead of replacing them
1263 - nir/lower_tex: Add lowering for some min_lod cases
1268 - nir: Document the function inlining process
1269 - nir: Allow [iu]mul_high on non-32-bit types
1270 - nir/lower_int64: Add support for [iu]mul_high
1271 - nir: Add a pass for lowering integer division by constants
1274 - nir/lower_idiv: Use ilt instead of bit twiddling
1275 - nir/tgsi: Use nir_bany in ttn_kill_if
1276 - nir/constant_folding: Fix source bit size logic
1277 - nir/algebraic: Optimize x2b(xneg(a)) -> a
1278 - nir: Drop support for lower_b2f
1279 - nir/algebraic: Make an optimization more specific
1280 - nir: Rename Boolean-related opcodes to include 32 in the name
1281 - nir/constant_expressions: Rework Boolean handling
1282 - nir: Add support for 1-bit data types
1283 - nir/large_constants: Properly handle 1-bit bools
1284 - nir/algebraic: Generalize an optimization
1285 - nir: Add 1-bit Boolean opcodes
1286 - nir: Add a bool to int32 lowering pass
1287 - nir: Switch to using 1-bit Booleans for almost everything
1288 - nir/algebraic: Optimize 1-bit Booleans
1289 - nir/algebraic: Add some optimizations for D3D-style Booleans
1291 - st/nir: Use nir_src_as_uint for tokens
1295 - nir/propagate_invariant: Skip unknown vars
1296 - nir/linking_helpers: Look at derefs for modes
1297 - nir/lower_io_arrays_to_elements: Look at derefs for modes
1298 - nir/lower_io_to_scalar: Look at derefs for modes
1299 - nir/lower_wpos_center: Look at derefs for modes
1300 - nir/copy_prop_vars: Get modes directly from derefs
1301 - nir/dead_write_vars: Get modes directly from derefs
1311 - spirv: Emit NIR deref instructions on-the-fly
1312 - nir/builder: Add nir_i2i and nir_u2u helpers which take a bit size
1314 - nir/validate: Require array indices to match the deref bit size
1315 - nir: Allow storing to shader_storage
1316 - nir: Distinguish between normal uniforms and UBOs
1322 - nir: Move propagation of cast derefs to a new nir_opt_deref pass
1323 - nir: Add a ptr_as_array deref type
1324 - nir/validate: Don't allow derefs in if conditions
1325 - nir/opt_deref: Properly optimize ptr_as_array derefs
1326 - nir/deref: Support casts and ptr_as_array in comparisons
1327 - nir/deref: Skip over casts in fixup_deref_modes
1328 - nir/remove_dead_variables: Properly handle deref casts
1329 - nir/validate: Allow derefs in phi nodes
1330 - nir/intrinsics: Allow deref sources to consume anything
1331 - nir/intrinsics: Add access flags to load/store_deref
1332 - nir/validate: Allow array derefs on vectors in more modes
1333 - nir/lower_io: Add "explicit" IO lowering
1334 - nir/vulkan: Add a descriptor type to vulkan resource intrinsics
1343 - nir: Add some more int64 lowering helpers
1348 - anv/pipeline_cache: Add support for caching NIR
1349 - anv/pipeline: Cache the pre-lowered NIR
1353 - intel/nir: Call nir_opt_deref in brw_nir_optimize
1354 - nir/gcm: Support deref instructions
1361 - nir: Add a bool to float32 lowering pass
1369 - nir: Mark deref UBO and SSBO access as non-scalar
1375 - nir: Preserve offsets in lower_io_to_scalar_early
1376 - nir/xfb: Fix offset accounting for dvec3/4
1377 - nir/xfb: Properly handle arrays of blocks
1385 - nir: Add load/store/atomic global intrinsics
1386 - nir/lower_io: Add a 32 and 64-bit global address formats
1387 - nir/lower_io: Add support for nir_var_mem_global
1388 - nir/validate: Allow array derefs of vectors for nir_var_mem_global
1389 - nir: Allow SSBOs and global to alias
1417 - nir/deref: Rematerialize parents in
1421 - nir/xfb: Properly align 64-bit values
1422 - nir: Rewrite lower_clip_cull_distance_arrays to do a lot less
1424 - nir/xfb: Work in terms of components rather than slots
1425 - nir/xfb: Handle compact arrays in gather_xfb_info
1426 - nir/lower_clip_cull: Fix an incorrect assert
1430 - intel,nir: Lower TXD with min_lod when the sampler index is not < 16
1434 - nir: Add a pass for lowering IO back to vector when possible
1435 - intel/nir: Vectorize all IO
1451 - nir: add fceil lowering
1460 - glsl/nir: int constants as float for native_integers=false
1461 - glsl/nir: ftrunc for native_integers=false float to int cast
1462 - glsl/nir: keep bool types when native_integers=false
1469 - freedreno: a2xx: NIR backend
1535 - nir: Fix assert in print_intrinsic_instr().
1544 - nir: add const_index parameters to system value builder function
1545 - nir: replace nir_load_system_value calls with appropiate builder
1547 - nir/spirv: cast shift operand to u32
1552 - nir: rename global/local to private/function memory
1556 - nir: rename nir_var_private to nir_var_shader_temp
1557 - nir: rename nir_var_function to nir_var_function_temp
1558 - nir: rename nir_var_ubo to nir_var_mem_ubo
1559 - nir: rename nir_var_ssbo to nir_var_mem_ssbo
1560 - nir: rename nir_var_shared to nir_var_mem_shared
1561 - nir/spirv: handle SpvStorageClassCrossWorkgroup
1564 - nir: replace more nir_load_system_value calls with builder functions
1565 - nir/validate: allow to check against a bitmask of bit_sizes
1566 - nir: add legal bit_sizes to intrinsics
1567 - nir: add bit_size parameter to system values with multiple allowed
1571 - nir/spirv: handle ContractionOff execution mode
1577 - nir/opt_if: don't mark progress if nothing changes
1583 - st/nir: Drop unused parameter from st_nir_assign_uniform_locations().
1587 - nir: Use nir_shader_get_entrypoint in nir_lower_clip_vs().
1588 - nir: Inline lower_clip_vs() into nir_lower_clip_vs().
1589 - nir: Save nir_variable pointers in nir_lower_clip_vs rather than
1591 - nir: Make nir_lower_clip_vs optionally work with variables.
1593 - i965: Use a 'nir' temporary rather than poking at brw_program
1594 - i965: Do NIR shader cloning in the caller.
1595 - intel/compiler: Use nir's info when checking uses_streams.
1605 - st/nir: Gather info after applying lowering FS variant features
1606 - st/nir: Drop unused gl_program parameter in VS input handling helper.
1607 - nir: Fix gl_nir_lower_samplers_as_deref's structure type handling.
1608 - nir: Make gl_nir_lower_samplers use gl_nir_lower_samplers_as_deref
1611 - nir: Allow a non-existent sampler deref in
1613 - st/nir: Lower TES gl_PatchVerticesIn to a constant if linked with a
1624 - nir: Don't reassociate add/mul chains containing only constants
1666 - nir/lower_tex: add alpha channel parameter for yuv lowering
1667 - nir/lower_tex: Add AYUV lowering support
1745 - radeonsi/nir: parse more information about bindless usage
1823 - radeonsi: always enable NIR for Civilization 6 to fix corruption
1868 - nir: Call fflush() at the end of nir_print_shader()
1882 - nir: Rework nir_lower_constant_initializers() to handle functions
1883 - nir: Tag entrypoint for easy recognition by
1885 - nir: Wire up int64 lowering functions
1886 - nir: Implement lowering of 64-bit shift operations
1887 - nir: Add and set \info::uses_64bit
1888 - nir: Create nir_builder in nir_lower_doubles_impl()
1889 - nir: Add lowering support for 64-bit operations to software
1890 - nir: Unset metadata debug bit if no progress made
1909 - nir: Optimize double-precision lower_round_even()
1949 - spirv/nir: handle location decorations on block interface members
2027 - nir: Don't lower the local work group size if it's variable.
2070 - nir: fix constness in nir_intrinsic_align()
2075 - ac/nir,radv,radeonsi/nir: use correct indices for interpolation
2077 - nir: fix copy-paste error in nir_lower_constant_initializers
2123 - nir: fix spelling typo
2124 - mesa/st/nir: fix missing nir_compact_varyings
2154 - nir/vtn: add caps for some cl related capabilities
2202 - nir: Add a new lowering option to lower 3D surfaces from txd to txl.
2246 - ac/nir: fix intrinsic name string size in visit_image_atomic()
2278 - ac/nir: remove the bitfield_extract workaround for LLVM 8
2283 - ac/nir: restrict fmask lookup to image load intrinsics
2290 - ac/nir: add get_cache_policy() helper and use it
2291 - ac/nir: set cache policy when loading/storing buffer images
2296 - ac/nir: don't trash L1 caches for store operations with writeonly
2302 - nir: do not remove varyings used for transform feedback
2303 - nir: fix lowering arrays to elements for XFB outputs
2333 - nir: Length of boolean vtn_value now is 1
2363 - nir: cleanup glsl_get_struct_field_offset, glsl_get_explicit_stride
2367 - nir: initialize value in copy_prop_vars_block
2388 - nir: allow propagation of if evaluation for bcsel
2389 - nir: fix condition propagation when src has a swizzle
2391 - nir: add new linking opt nir_link_constant_varyings()
2393 - nir: add glsl_type_is_integer() helper
2394 - nir: don't pack varyings ints with floats unless flat
2396 - nir: add support for removing redundant stores to copy prop var
2398 - nir: small tidy ups for nir_loop_analyze()
2399 - nir: clarify some nit_loop_info member names
2400 - nir: add a new nir_cf_list_clone_and_reinsert() helper
2401 - nir: make use of new nir_cf_list_clone_and_reinsert() helper
2402 - nir: factor out some of the complex loop unroll code to a helper
2403 - nir: rework force_unroll_array_access()
2404 - nir: in loop analysis track actual control flow type
2405 - nir: reword code comment
2406 - nir: detect more induction variables
2407 - nir: fix opt_if_loop_last_continue()
2414 - nir: rename nir_link_constant_varyings() nir_link_opt_varyings()
2415 - nir: add can_replace_varying() helper
2416 - nir: rework nir_link_opt_varyings()
2417 - nir: link time opt duplicate varyings
2418 - nir: make nir_opt_remove_phis_impl() static
2419 - nir: make use of does_varying_match() helper
2420 - nir: simplify does_varying_match()
2421 - nir: add rewrite_phi_predecessor_blocks() helper
2422 - nir: merge some basic consecutive ifs
2424 - nir: avoid uninitialized variable warning
2427 - radeonsi/nir: get correct type for images inside structs
2436 - radeonsi/nir: add missing piece for bindless image support
2504 - nir/algebraic: Make algebraic_parser_test.sh executable.