Binary : ptxas v13.0.88, 37.7 MB stripped ELF, ~40,000 functionsDocumented : 2,063 unique functions across 70 wiki pagesThis page : Top ~100 most cross-referenced functions, plus routing tablesComplete listings : Each wiki page has its own Function Map section with full details
This page is the central lookup index for identified functions in ptxas. It lists the functions that appear most frequently across the wiki (cross-cutting infrastructure and major entry points), and provides routing tables to find any function by address range or subsystem.
Confidence levels : CERTAIN = named in symbols or strings. HIGH = strong evidence from strings and call patterns (>90%). MEDIUM = structural analysis with partial string evidence (70-90%).
These functions appear in 10+ wiki pages -- they are the universal building blocks called by nearly every subsystem.
Address Identity Pages Callers Notes
0x424070pool_alloc(pool, size)19 3,809 Custom slab allocator, 8-byte aligned 0x4248B0pool_free(ptr)8 1,215 Coalescing free, boundary tags 0x4280C0get_thread_local_context10 3,928 Most-called function in ptxas; 280-byte TLS struct 0x42BDB0fatal_OOM_handler8 3,825 Called on every allocation failure 0x426150hashmap_put(map, key, value)11 2,800 Open-addressing + chaining, auto-resize 0x426D60hashmap_get(map, key)11 422 Returns value or 0 0x425CA0hashmap_create(hash_fn, cmp_fn, cap)7 127 Integer/pointer/custom hash modes 0x427630murmurhash3_x86_32(str)5 73 Constants: 0xcc9e2d51, 0x1b873593 0x42D850hashset_insert(set, key)4 282 Hash set variant 0x42FBA0diagnostic_emit(desc, loc, fmt...)7 2,350 Central error/warning reporter 0x42F590fatal_internal_error(desc, ...)8 3,825 Assertion handler 0x4279D0starts_with(str, prefix)4 185 Returns suffix pointer or 0 0x42CA60list_push_front(node, head_ptr)4 298 Pool-allocated linked list 0xBDBA60bitvector_allocate8 many (bits+31)>>5 word count 0xBDCDE0bitvector_or_assign (SSE2)5 many _mm_or_si128 on 128-bit chunks
Details : Memory Pools , Hash & Bitvector , Threading
Address Identity Pages Callers Notes
0x409460main5 1 Delegates to 0x446240 0x446240real_main (top-level driver)13 1 Orchestrates entire pipeline 0x4428E0ptx_input_setup6 1 Version/target validation 0x43CC70per_entry_compile_unit5 1 Processes each entry through pipeline 0x43F400function_abi_config4 1 Parameter regs, return addr, scratch 0x43A400compilation_target_config7 1 SM-specific defaults 0x43B660register_constraint_calculator5 1 Balances .maxnreg, occupancy 0x432A00option_registration9 1 CLI option definitions 0x434320option_parser9 1 Validates combinations, applies state
Details : Pipeline Entry , Pipeline Overview , CLI Options
Address Identity Pages Callers Notes
0x46E000instruction_table_builder9 1 93 KB, 1168 callees, one per PTX opcode 0x451730parser_setup (special register init)9 1 %ntid, %laneid, %clock, etc. 0x4CE6B0bison_parser (directive/decl)7 1 .local_maxnreg, .alias, .pragma 0x720F00flex_lexer (ptxlex / yylex)8 2 ~550 Flex rules, DFA scanner 0x4B2F20ptx_validator_general4 1 Validates texture, surface, cvt, call 0x4C5FB0ptx_validator_mma_wmma_tcgen054 1 MMA, WMMA, tensor core validation 0x71F630preprocessor_dispatch4 1 .MACRO, .ELSE, .INCLUDE 0x489050ptx_to_ori_converter5 1 PTX AST to ORI IR translation
Details : PTX Parser , PTX Directives , PTX to ORI
Address Identity Pages Callers Notes
0x4094C0ctor_001 -- thread infra init4 0 pthread_key_create, mutex 0x4095D0ctor_003 -- PTX opcode name table6 0 ~900 ROT13-encoded PTX mnemonics 0x40D860ctor_005 -- tuning knob registry6 0 80 KB, 2000+ ROT13 knob names 0x421290ctor_007 -- scheduler knob registry4 0 98 ROT13 scheduler knobs
Details : Pipeline Entry , Binary Layout
Address Identity Pages Callers Notes
0xC60D30phase_factory (159-case switch)12 1 Allocates phase objects 0xC62720PhaseManager_ctor10 2 159-entry phase table 0xC64F70phase_dispatch_loop5 2 Executes phases, reports timing 0xC64310per_phase_timing_reporter5 1 "[Total N KB] [Freeable N KB]" 0xC641D0phase_name_to_index_lookup5 3 Binary search, case-insensitive 0x7DDB50phase_run_dispatch14 many Vtable-based phase execution 0x9F4040NamedPhases_parse_and_build6 1 "shuffle", "OriCopyProp", etc. 0x798B60NamedPhases_parser4 2 PTXAS_DISABLE env var parsing 0x799250IsPassDisabled5 4 Checks knob index 185 0xA36360pass_sequence_builder6 1 Constructs NvOptRecipe pass list
Details : Phase Manager , Pass Inventory , Optimizer Pipeline
Address Identity Pages Callers Notes
0x9253C0instruction_operand_get11 many Operand accessor on ORI instructions 0x7E6090instruction_modifier_set10 many IR modification helper 0x781F80instruction_iterator12 many Doubly-linked list traversal 0x7DF3A0instruction_property_query5 many Instruction flag/attribute checker 0x91BF30register_type_query8 many Register class/type inspection 0x9314F0register_class_id_query7 1,547 Most-called non-trivial regalloc fn 0x931920register_class_compat_checker6 328 Pair register class handling 0x934630register_id_packer9 856 Packs reg#/class/type into 32-bit 0xB28E00ir_node_type_query5 many Node kind discrimination 0xB28E90ir_node_field_accessor6 many Generic field getter 0xA50650CodeObject_EmitRecords1 8 74 KB, ORI record serializer (56 section types) 0xA53840EmitRecords_wrapper1 1 Thin wrapper, adds type-44 header
Details : Instructions , Registers , Data Structures , CFG
Address Identity Pages Callers Notes
0x5D1660intrinsic_table_register (608 entries)7 1 Master name-to-ID table 0x5D4190intrinsic_dispatch_builder13 1 PTX opcode -> codegen handler mapping 0x5FF700intrinsic_prototype_emitter5 1 354 KB -- largest function in binary 0x5C7A50wmma_mma_codegen4 1 173 KB, all shapes/types/layouts 0x5C10A0mma_codegen (mma.sync)4 1 120 KB, m8n8k4 through m16n8k256 0x5BBC30tcgen05_mma_codegen (Blackwell)5 1 90 KB, 5th-gen tensor core 0x70FA00ocg_intrinsic_handler8 1 OCG-level intrinsic routing 0x6A97B0intrinsic_lowering_main4 1 26 KB, switch-based lowering 0x6C9EB0ocg_builtin_name_lookup5 1 Blackwell+ OCG name table
Details : Intrinsics Index , Math Intrinsics , Tensor Intrinsics , Sync & Warp
Address Identity Pages Callers Notes
0x9721C0regalloc_entry ("REGALLOC GUIDANCE")6 1 Top-level allocator entry 0x957160fatpoint_allocator_core7 1 Core fatpoint graph coloring 0x96D940spill_guidance_engine5 1 Determines spill strategy 0x971A90full_alloc_with_spill_retry4 1 "NOSPILL REGALLOC" path 0x9714E0regalloc_failure_reporter6 1 "Register allocation failed..." 0x926A30interference_graph_builder9 7 22 KB, SSE bitvectors 0x92C240liveness_bitvector_ops5 87 Set/clear/query with aliasing 0x917A60opcode_to_regclass_mapping4 221 Massive switch 0x910840ConvertMemoryToRegisterOrUniform5 1 Pass driver
Details : RegAlloc Overview , RegAlloc Algorithm , Spilling , ABI
Address Identity Pages Callers Notes
0x8D0640ScheduleInstructions (top-level)7 1 String: "ScheduleInstructions" 0x688DD0scheduler_engine (main BB loop)5 1 ReduceReg / DynBatch selection 0x8C9320scheduling_priority_function4 0 ~300 locals, core heuristic 0x68B9C0dependency_graph_builder4 1 RAW/WAR/WAW hazard analysis 0x6820B0build_ready_list5 1 Zero-dependency instructions 0x8CD6E0reverse_scheduling_driver4 1 Reverse post-order iteration 0x8CEE80register_budget_with_occupancy4 1 Pressure coeff default 0.045 0x8E4400hw_profile_table_init6 3 Encoding/latency property tables 0xA9CDE0scheduling_metadata_builder6 1 Per-instruction sched metadata 0xA9CF90scheduling_metadata_accessor5 many Sched metadata field queries 0xAED3C0scheduling_optimization_mega_pass4 0 137 KB, ~560 locals, largest vtable pass
Details : Scheduling Overview , Scheduling Algorithm , Latency Model , Scoreboards
Address Identity Pages Callers Notes
0x169B190isel_pattern_dispatch (master)5 1 280 KB, 65,999 insns -- largest function 0x143C440sm120_peephole_dispatch4 1 SM120 (RTX 50), 373-case switch 0x198BCD0sm100_peephole_dispatch4 1 SM100 (Blackwell), 1336 callees 0x83EF00main_peephole_pass6 0 29 KB, 392 callees 0x6D9690master_instruction_encoder7 1 94 KB, opcode switch 0x6E4110sass_codegen_main4 1 EmitSASSForFunction, FNV-1a BB hash 0x6F52F0SASS_pipeline_run_stages5 1 Mercury SASS compilation pipeline 0x9ED2D0MercConverter_entry6 1 ORI to Mercury IR conversion 0x9F1A90MercConverter_builder6 1 Mercury instruction construction
Details : ISel , Encoding , Peephole , Mercury , Templates
Address Identity Pages Callers Notes
0x7B9B80bitfield_insert(insn, off, wid, val)9 18,347 Most-called by caller count 0x7BC030encode_register_operand4 6,147 1-bit + 4-bit type + 10-bit reg 0x7B9D60encode_reuse_flags_predicate4 2,408 1-bit reuse + 5-bit predicate 0x7BC5C0encode_immediate_const_operand4 1,449 Const buffer index or immediate 0x7BCF00encode_predicate_register4 1,657 PT=14, 2-bit type + 3-bit condition 0x10B61801_bit_boolean_encoder3 8,091 .S/.U, .STRONG, etc.
Details : Encoding , SASS Printing
Address Identity Pages Callers Notes
0x612DE0section_attr_builder11 1 76 KB, ELF section/attribute config 0x1C9F280master_elf_emitter9 1 Complete CUBIN assembly 0x1CB53A0elf_world_init7 1 672-byte ELFW context 0x1CB68D0symbol_table_builder5 1 .symtab from internal symbols 0x1CABD60master_section_allocator5 1 Shared/const/local memory 0x1CB3570add_function_section5 44 Creates .text.FUNCNAME + .rela 0x1CD48C0relocation_processor5 1 Relocation section emission 0x1C9B110mercury_capsule_builder4 1 Creates embedded .nv.merc ELF
Details : ELF Emitter , Sections , Relocations , Debug Info , Capsule Mercury
Address Identity Pages Callers Notes
0x79B240GetKnobIndex6 2 ROT13 name lookup, case-insensitive 0x79D070ReadKnobsFile5 1 Parses [knobs] section from file 0x79F540ParseKnobValue4 1 12-type switch: bool/int/float/string/... 0x79D990ProcessKnobs (top-level)4 1 File + pragma + numbered config 0xA0F020knob_conditional_evaluator5 many [WHEN condition] handler
Details : Knobs , Opt Levels
Address Identity Pages Callers Notes
0x6765E0target_profile_selector7 1 SM-dependent profile dispatch 0x607DB0target_feature_query7 many SM feature capability checks 0x896D50sass_mnemonic_table_init (ROT13)4 1 ~400+ SASS instruction names 0x89FBA0instruction_latency_init4 3 Encoding/latency property tables
Details : Targets Index , Turing-Ampere , Ada-Hopper , Blackwell , tcgen05
To find a specific function, locate it by address range or subsystem topic in this table. Each page contains a detailed Function Map section with complete listings.
Subsystem Primary Pages Functions
Memory allocator, pools memory-pools.md 30 Hash maps, bitvectors, sets hash-bitvector.md 51 Threading, TLS, jobserver threading.md 41 CLI parsing, option handling cli-options.md 17 Tuning knobs (2000+ knobs) knobs.md 56 Optimization levels opt-levels.md 14 DumpIR debug output dumpir.md 14 Compilation pipeline overview.md , entry.md 56+25 PTX lexer & parser ptx-parser.md 75 PTX directives ptx-directives.md 41 PTX-to-ORI translation ptx-to-ori.md 41 Optimizer pipeline optimizer.md 28 ORI instruction IR instructions.md 80 CFG construction cfg.md 18 Register representation registers.md 40 IR data structures data-structures.md 74 Phase manager (159 phases) phase-manager.md 26 Copy propagation, CSE, GVN copy-prop-cse.md 65 General optimization passes general-optimize.md 71 Loop optimization (unroll, LICM, SWP) loop-passes.md 92 Branch/switch optimization branch-switch.md 24 Strength reduction strength-reduction.md 25 Predication predication.md 28 Rematerialization rematerialization.md 55 Liveness analysis liveness.md 42 Sync barriers sync-barriers.md 66 Late legalization late-legalization.md 59 Hot/cold splitting hot-cold.md 10 GMMA pipelining gmma-pipeline.md 47 Uniform registers uniform-regs.md 22 Register allocator core algorithm.md 50 Spilling spilling.md 54 ABI handling abi.md 87 Scheduling overview overview.md 112 Scheduling algorithm algorithm.md 121 Latency model & HW profiles latency-model.md 78 Scoreboards & barriers scoreboards.md 56 ISel pattern matching isel.md 182 SASS encoding encoding.md 92 Peephole optimization peephole.md 67 Mercury IR conversion mercury.md 79 SASS templates templates.md 46 SASS printing / renderer sass-printing.md 96 Capsule Mercury capmerc.md 20 Intrinsic infrastructure index.md 159 Math intrinsics math.md 42 Tensor core intrinsics tensor.md 45 Sync & warp intrinsics sync-warp.md 65 SM targets & features index.md 70 ELF emitter elf-emitter.md 29 ELF sections sections.md 33 Debug info (DWARF) debug-info.md 33 Relocations relocations.md 19
Functions in the binary are clustered by subsystem. This table maps address ranges to the pages that document them.
Address Range Primary Subsystem Key Pages
0x400000-0x424000Entry, static init, main entry.md , binary-layout.md 0x424000-0x42E000Memory pools, hash maps, lists memory-pools.md , hash-bitvector.md 0x42E000-0x446000Diagnostics, CLI parsing cli-options.md , entry.md 0x446000-0x452000Compilation driver overview.md , entry.md 0x452000-0x4D5000PTX parser & validator ptx-parser.md , ptx-directives.md 0x4D5000-0x5AA000PTX-to-ORI, early IR ptx-to-ori.md , instructions.md 0x5AA000-0x612000Intrinsic infrastructure index.md , math.md , tensor.md 0x612000-0x67F000Section builder, target config sections.md , index.md 0x67F000-0x6E4000Scheduling engine, OCG lowering, encoding overview.md , encoding.md 0x6E4000-0x754000SASS codegen, SASS pipeline mercury.md , overview.md 0x754000-0x7C0000Liveness, knobs, bitfield encoding liveness.md , knobs.md , encoding.md 0x7C0000-0x8FE000Peephole, SASS mnemonics, scheduling upper peephole.md , algorithm.md 0x8FE000-0x9D3000Register allocator overview.md , algorithm.md , abi.md 0x9D3000-0xAA8000Post-regalloc, named phases, remat rematerialization.md , phase-manager.md 0xAA8000-0xC52000Mega-passes, sync barriers, dataflow sync-barriers.md , general-optimize.md 0xC52000-0xD27000Phase manager, phase factory phase-manager.md , optimizer.md 0xD27000-0x10B7000592 SASS encoder bodies encoding.md , isel.md 0x10B7000-0x1225000Field encoders, ISel helpers encoding.md , isel.md 0x1225000-0x13CF000Bitvector, ISel coordinators hash-bitvector.md , isel.md 0x13CF000-0x17F8000SM-specific ISel, pattern matchers, templates isel.md , templates.md 0x17F8000-0x1C21000SASS printing, peephole mega-dispatchers sass-printing.md , peephole.md 0x1C21000-0x1CE3000ELF emitter, capsule mercury, relocations elf-emitter.md , capmerc.md
Rank Address Identity Callers
1 0x7B9B80bitfield_insert 18,347 2 0x10B61801-bit boolean encoder 8,091 3 0x7BC030encode_register_operand 6,147 4 0x4280C0get_thread_local_context 3,928 5 0x42BDB0fatal_OOM_handler 3,825 6 0x424070pool_alloc 3,809 7 0x426150hashmap_put 2,800 8 0x7B9D30clear_const_buffer_slots 2,408 9 0x7B9D60encode_reuse_flags_predicate 2,408 10 0x42FBA0diagnostic_emit 2,350
Rank Address Identity Size
1 0x5FF700intrinsic_prototype_emitter 354 KB 2 0x169B190isel_pattern_dispatch 280 KB 3 0x198BCD0sm100_peephole_dispatch 233 KB 4 0x143C440sm120_peephole_dispatch 233 KB 5 0x5C7A50wmma_mma_codegen 173 KB
Rank Address Identity Pages
1 0x424070pool_alloc 19 2 0x7DDB50phase_run_dispatch 14 3 0x446240real_main 13 3 0x5D4190intrinsic_dispatch_builder 13 5 0x781F80instruction_iterator 12 5 0xC60D30phase_factory 12 7 0x9253C0instruction_operand_get 11 7 0x612DE0section_attr_builder 11 7 0x426150hashmap_put 11 7 0x426D60hashmap_get 11
Metric Count
Total unique functions documented 2,063 Wiki pages with function maps 70 Functions in 5+ pages (high cross-reference) 89 Functions in 1 page only (subsystem-internal) 1,324 Confidence CERTAIN ~40 Confidence HIGH ~1,400 Confidence MEDIUM ~620