summaryrefslogtreecommitdiff
path: root/zjit/bindgen/src
AgeCommit message (Collapse)Author
4 daysZJIT: remove unused rb_RSTRUCT_LEN()Benoit Daloze
4 daysZJIT: remove unused rb_RSTRUCT_SET()Benoit Daloze
2025-12-16ZJIT: Use rb_zjit_writebarrier_check_immediate() instead of ↵Benoit Daloze
rb_gc_writebarrier() in gen_write_barrier() * To avoid calling rb_gc_writebarrier() with an immediate value in gen_write_barrier(), and avoid the LIR jump issue.
2025-12-09ZJIT: Add codegen for FixnumDiv (#15452)Abrar Habib
Fixes https://github.com/Shopify/ruby/issues/902 This pull request adds code generation for dividing fixnums. Testing confirms the normal case, flooring, and side-exiting on division by zero.
2025-12-05ZJIT: Avoid binding to `rb_iseq_constant_body`Alan Wu
Its definition changes depending on e.g. whether there is YJIT in the build.
2025-12-03ZJIT: Optimize setivar with shape transition (#15375)Max Bernstein
Since we do a decent job of pre-sizing objects, don't handle the case where we would need to re-size an object. Also don't handle too-complex shapes. lobsters stats before: ``` Top-20 calls to C functions from JIT code (79.4% of total 90,051,140): rb_vm_opt_send_without_block: 19,762,433 (21.9%) rb_vm_setinstancevariable: 7,698,314 ( 8.5%) rb_hash_aref: 6,767,461 ( 7.5%) rb_vm_env_write: 5,373,080 ( 6.0%) rb_vm_send: 5,049,229 ( 5.6%) rb_vm_getinstancevariable: 4,535,259 ( 5.0%) rb_obj_is_kind_of: 3,746,306 ( 4.2%) rb_ivar_get_at_no_ractor_check: 3,745,237 ( 4.2%) rb_vm_invokesuper: 3,037,467 ( 3.4%) rb_ary_entry: 2,351,983 ( 2.6%) rb_vm_opt_getconstant_path: 1,344,740 ( 1.5%) rb_vm_invokeblock: 1,184,474 ( 1.3%) Hash#[]=: 1,064,288 ( 1.2%) rb_gc_writebarrier: 1,006,972 ( 1.1%) rb_ec_ary_new_from_values: 902,687 ( 1.0%) fetch: 898,667 ( 1.0%) rb_str_buf_append: 833,787 ( 0.9%) rb_class_allocate_instance: 822,024 ( 0.9%) Hash#fetch: 699,580 ( 0.8%) _bi20: 682,068 ( 0.8%) Top-4 setivar fallback reasons (100.0% of total 7,732,326): shape_transition: 6,032,109 (78.0%) not_monomorphic: 1,469,300 (19.0%) not_t_object: 172,636 ( 2.2%) too_complex: 58,281 ( 0.8%) ``` lobsters stats after: ``` Top-20 calls to C functions from JIT code (79.0% of total 88,322,656): rb_vm_opt_send_without_block: 19,777,880 (22.4%) rb_hash_aref: 6,771,589 ( 7.7%) rb_vm_env_write: 5,372,789 ( 6.1%) rb_gc_writebarrier: 5,195,527 ( 5.9%) rb_vm_send: 5,049,145 ( 5.7%) rb_vm_getinstancevariable: 4,538,485 ( 5.1%) rb_obj_is_kind_of: 3,746,241 ( 4.2%) rb_ivar_get_at_no_ractor_check: 3,745,172 ( 4.2%) rb_vm_invokesuper: 3,037,157 ( 3.4%) rb_ary_entry: 2,351,968 ( 2.7%) rb_vm_setinstancevariable: 1,703,337 ( 1.9%) rb_vm_opt_getconstant_path: 1,344,730 ( 1.5%) rb_vm_invokeblock: 1,184,290 ( 1.3%) Hash#[]=: 1,061,868 ( 1.2%) rb_ec_ary_new_from_values: 902,666 ( 1.0%) fetch: 898,666 ( 1.0%) rb_str_buf_append: 833,784 ( 0.9%) rb_class_allocate_instance: 821,778 ( 0.9%) Hash#fetch: 755,913 ( 0.9%) Top-4 setivar fallback reasons (100.0% of total 1,703,337): not_monomorphic: 1,472,405 (86.4%) not_t_object: 172,629 (10.1%) too_complex: 58,281 ( 3.4%) new_shape_needs_extension: 22 ( 0.0%) ``` I also noticed that primitive printing in HIR was broken so I fixed that. Co-authored-by: Aaron Patterson <tenderlove@ruby-lang.org>
2025-12-04ZJIT: Inline Kernel#class (#15397)Max Bernstein
We generally know the receiver's class from profile info. I see 600k of these when running lobsters.
2025-12-02ZJIT: Optimize GetIvar for non-T_OBJECTBenoit Daloze
* All Invariant::SingleRactorMode PatchPoint are replaced by assume_single_ractor_mode() to fix https://github.com/Shopify/ruby/issues/875 for SingleRactorMode patchpoints.
2025-12-01ZJIT: Specialize String#<< with FixnumMax Bernstein
Append a codepoint.
2025-11-25ZJIT: Specialize setinstancevariable when ivar is already in shape (#15290)Max Bernstein
Don't support shape transitions for now.
2025-11-20Name the `iseq->body->param` struct and update bindings for JITsAlan Wu
This will make reading the parameters nicer for the JITs. Should be no-op for the C side.
2025-11-18Extract `KW_SPECIFIED_BITS_MAX` for JITs (GH-15039)Jacob
Rename to `VM_KW_SPECIFIED_BITS_MAX` now that it's in `vm_core.h`.
2025-11-06ZJIT: Untag block handler (#15085)Max Bernstein
Storing the tagged block handler in profiles is not GC-safe (nice catch, Kokubun). Store the untagged block handler instead. Fix bug in https://github.com/ruby/ruby/pull/15051
2025-11-05ZJIT: Profile specific objects for invokeblock (#15051)Max Bernstein
I made a special kind of `ProfiledType` that looks at specific objects, not just their classes/shapes (https://github.com/ruby/ruby/pull/15051). Then I profiled some of our benchmarks. For lobsters: ``` Top-6 invokeblock handler (100.0% of total 1,064,155): megamorphic: 494,931 (46.5%) monomorphic_iseq: 337,171 (31.7%) polymorphic: 113,381 (10.7%) monomorphic_ifunc: 52,260 ( 4.9%) monomorphic_other: 38,970 ( 3.7%) no_profiles: 27,442 ( 2.6%) ``` For railsbench: ``` Top-6 invokeblock handler (100.0% of total 2,529,104): monomorphic_iseq: 834,452 (33.0%) megamorphic: 818,347 (32.4%) polymorphic: 632,273 (25.0%) monomorphic_ifunc: 224,243 ( 8.9%) monomorphic_other: 19,595 ( 0.8%) no_profiles: 194 ( 0.0%) ``` For shipit: ``` Top-6 invokeblock handler (100.0% of total 2,104,148): megamorphic: 1,269,889 (60.4%) polymorphic: 411,475 (19.6%) no_profiles: 173,367 ( 8.2%) monomorphic_other: 118,619 ( 5.6%) monomorphic_iseq: 84,891 ( 4.0%) monomorphic_ifunc: 45,907 ( 2.2%) ``` Seems like a monomorphic case for a specific ISEQ actually isn't a bad way of going about this, at least to start...
2025-10-28ZJIT: Specialize Array#pop for no argument case (#14933)Aiden Fox Ivey
Fixes https://github.com/Shopify/ruby/issues/814 This change specializes the case of calling `Array#pop` on a non frozen array with no arguments. `Array#pop` exists in the non-inlined C function list in the ZJIT SFR performance burndown list. If in the future it is helpful, this patch could be extended to support the case where an argument is provided, but this initial work seeks to elide the ruby frame normally pushed in the case of `Array#pop` without an argument.
2025-10-22ZJIT: Fetch Primitive.attr!(leaf) for InvokeBuiltinMax Bernstein
Fix https://github.com/Shopify/ruby/issues/670
2025-10-21YJIT: ZJIT: Extract common bindings to jit.c and remove unnamed enums.Alan Wu
The type name bindgen picks for anonymous enums creates desync issues on the bindgen CI checks.
2025-10-21ZJIT: Fix binding to `INVALID_SHAPE_ID` under `-std=c99 -pedantic`Alan Wu
``` /src/jit.c:19:5: error: ISO C restricts enumerator values to range of 'int' (4294967295 is too large) [-Werror,-Wpedantic] 19 | RB_INVALID_SHAPE_ID = INVALID_SHAPE_ID, | ^ ~~~~~~~~~~~~~~~~ ```
2025-10-20ZJIT: Implement expandarray (#14847)Max Bernstein
Only support the simple case: no splat or rest. lobsters before: <details> ``` ***ZJIT: Printing ZJIT statistics on exit*** Top-20 not inlined C methods (60.5% of total 11,039,954): Kernel#is_a?: 1,030,769 ( 9.3%) String#<<: 851,954 ( 7.7%) Hash#[]=: 742,941 ( 6.7%) Regexp#match?: 399,894 ( 3.6%) String#empty?: 353,775 ( 3.2%) Hash#key?: 349,147 ( 3.2%) String#start_with?: 334,961 ( 3.0%) Kernel#respond_to?: 316,528 ( 2.9%) ObjectSpace::WeakKeyMap#[]: 238,978 ( 2.2%) TrueClass#===: 235,771 ( 2.1%) FalseClass#===: 231,144 ( 2.1%) Array#include?: 211,385 ( 1.9%) Hash#fetch: 204,702 ( 1.9%) Kernel#block_given?: 181,797 ( 1.6%) Kernel#dup: 179,341 ( 1.6%) BasicObject#!=: 175,997 ( 1.6%) Class#new: 168,079 ( 1.5%) Kernel#kind_of?: 165,600 ( 1.5%) String#==: 157,735 ( 1.4%) Module#clock_gettime: 144,992 ( 1.3%) Top-20 not annotated C methods (61.4% of total 11,202,087): Kernel#is_a?: 1,212,660 (10.8%) String#<<: 851,954 ( 7.6%) Hash#[]=: 743,120 ( 6.6%) Regexp#match?: 399,894 ( 3.6%) String#empty?: 361,013 ( 3.2%) Hash#key?: 349,147 ( 3.1%) String#start_with?: 334,961 ( 3.0%) Kernel#respond_to?: 316,528 ( 2.8%) ObjectSpace::WeakKeyMap#[]: 238,978 ( 2.1%) TrueClass#===: 235,771 ( 2.1%) FalseClass#===: 231,144 ( 2.1%) Array#include?: 211,385 ( 1.9%) Hash#fetch: 204,702 ( 1.8%) Kernel#block_given?: 191,666 ( 1.7%) Kernel#dup: 179,348 ( 1.6%) BasicObject#!=: 176,181 ( 1.6%) Class#new: 168,079 ( 1.5%) Kernel#kind_of?: 165,634 ( 1.5%) String#==: 163,667 ( 1.5%) Module#clock_gettime: 144,992 ( 1.3%) Top-2 not optimized method types for send (100.0% of total 72,318): cfunc: 48,055 (66.4%) iseq: 24,263 (33.6%) Top-6 not optimized method types for send_without_block (100.0% of total 4,523,682): iseq: 2,271,936 (50.2%) bmethod: 985,636 (21.8%) optimized: 949,703 (21.0%) alias: 310,747 ( 6.9%) null: 5,106 ( 0.1%) cfunc: 554 ( 0.0%) Top-13 not optimized instructions (100.0% of total 4,293,171): invokesuper: 2,373,404 (55.3%) invokeblock: 811,926 (18.9%) sendforward: 505,452 (11.8%) opt_eq: 451,754 (10.5%) opt_plus: 74,404 ( 1.7%) opt_minus: 36,228 ( 0.8%) opt_send_without_block: 21,792 ( 0.5%) opt_neq: 7,231 ( 0.2%) opt_mult: 6,752 ( 0.2%) opt_or: 3,753 ( 0.1%) opt_lt: 348 ( 0.0%) opt_ge: 91 ( 0.0%) opt_gt: 36 ( 0.0%) Top-9 send fallback reasons (100.0% of total 25,530,724): send_without_block_polymorphic: 9,722,491 (38.1%) send_no_profiles: 5,894,788 (23.1%) send_without_block_not_optimized_method_type: 4,523,682 (17.7%) not_optimized_instruction: 4,293,171 (16.8%) send_without_block_no_profiles: 998,746 ( 3.9%) send_not_optimized_method_type: 72,318 ( 0.3%) send_without_block_cfunc_array_variadic: 15,134 ( 0.1%) obj_to_string_not_string: 9,765 ( 0.0%) send_without_block_direct_too_many_args: 629 ( 0.0%) Top-9 unhandled YARV insns (100.0% of total 690,950): expandarray: 328,490 (47.5%) checkkeyword: 190,694 (27.6%) getclassvariable: 59,901 ( 8.7%) invokesuperforward: 49,503 ( 7.2%) getblockparam: 49,119 ( 7.1%) opt_duparray_send: 11,978 ( 1.7%) getconstant: 952 ( 0.1%) checkmatch: 290 ( 0.0%) once: 23 ( 0.0%) Top-3 compile error reasons (100.0% of total 3,718,636): register_spill_on_alloc: 3,418,255 (91.9%) register_spill_on_ccall: 182,018 ( 4.9%) exception_handler: 118,363 ( 3.2%) Top-14 side exit reasons (100.0% of total 10,860,385): compile_error: 3,718,636 (34.2%) guard_type_failure: 2,638,926 (24.3%) guard_shape_failure: 1,917,209 (17.7%) unhandled_yarv_insn: 690,950 ( 6.4%) block_param_proxy_not_iseq_or_ifunc: 535,789 ( 4.9%) unhandled_kwarg: 455,347 ( 4.2%) patchpoint: 370,476 ( 3.4%) unknown_newarray_send: 314,786 ( 2.9%) unhandled_splat: 122,071 ( 1.1%) unhandled_hir_insn: 76,397 ( 0.7%) block_param_proxy_modified: 19,193 ( 0.2%) obj_to_string_fallback: 566 ( 0.0%) guard_type_not_failure: 22 ( 0.0%) interrupt: 17 ( 0.0%) send_count: 62,244,604 dynamic_send_count: 25,530,724 (41.0%) optimized_send_count: 36,713,880 (59.0%) iseq_optimized_send_count: 18,587,512 (29.9%) inline_cfunc_optimized_send_count: 7,086,414 (11.4%) non_variadic_cfunc_optimized_send_count: 8,375,754 (13.5%) variadic_cfunc_optimized_send_count: 2,664,200 ( 4.3%) dynamic_getivar_count: 7,365,995 dynamic_setivar_count: 7,245,005 compiled_iseq_count: 4,796 failed_iseq_count: 447 compile_time: 814ms profile_time: 9ms gc_time: 9ms invalidation_time: 72ms vm_write_pc_count: 64,156,223 vm_write_sp_count: 62,812,449 vm_write_locals_count: 62,812,449 vm_write_stack_count: 62,812,449 vm_write_to_parent_iseq_local_count: 292,458 vm_read_from_parent_iseq_local_count: 6,599,701 code_region_bytes: 22,953,984 side_exit_count: 10,860,385 total_insn_count: 517,606,340 vm_insn_count: 162,979,530 zjit_insn_count: 354,626,810 ratio_in_zjit: 68.5% ``` </details> lobsters after: <details> ``` ***ZJIT: Printing ZJIT statistics on exit*** Top-20 not inlined C methods (59.9% of total 11,291,815): Kernel#is_a?: 1,046,269 ( 9.3%) String#<<: 851,954 ( 7.5%) Hash#[]=: 743,274 ( 6.6%) Regexp#match?: 399,894 ( 3.5%) String#empty?: 353,775 ( 3.1%) Hash#key?: 349,147 ( 3.1%) String#start_with?: 334,961 ( 3.0%) Kernel#respond_to?: 316,502 ( 2.8%) ObjectSpace::WeakKeyMap#[]: 238,978 ( 2.1%) TrueClass#===: 235,771 ( 2.1%) FalseClass#===: 231,144 ( 2.0%) String#sub!: 219,579 ( 1.9%) Array#include?: 211,385 ( 1.9%) Hash#fetch: 204,702 ( 1.8%) Kernel#block_given?: 181,797 ( 1.6%) Kernel#dup: 179,341 ( 1.6%) BasicObject#!=: 175,997 ( 1.6%) Class#new: 168,079 ( 1.5%) Kernel#kind_of?: 165,600 ( 1.5%) String#==: 157,742 ( 1.4%) Top-20 not annotated C methods (60.9% of total 11,466,928): Kernel#is_a?: 1,239,923 (10.8%) String#<<: 851,954 ( 7.4%) Hash#[]=: 743,453 ( 6.5%) Regexp#match?: 399,894 ( 3.5%) String#empty?: 361,013 ( 3.1%) Hash#key?: 349,147 ( 3.0%) String#start_with?: 334,961 ( 2.9%) Kernel#respond_to?: 316,502 ( 2.8%) ObjectSpace::WeakKeyMap#[]: 238,978 ( 2.1%) TrueClass#===: 235,771 ( 2.1%) FalseClass#===: 231,144 ( 2.0%) String#sub!: 219,579 ( 1.9%) Array#include?: 211,385 ( 1.8%) Hash#fetch: 204,702 ( 1.8%) Kernel#block_given?: 191,666 ( 1.7%) Kernel#dup: 179,348 ( 1.6%) BasicObject#!=: 176,181 ( 1.5%) Class#new: 168,079 ( 1.5%) Kernel#kind_of?: 165,634 ( 1.4%) String#==: 163,674 ( 1.4%) Top-2 not optimized method types for send (100.0% of total 72,318): cfunc: 48,055 (66.4%) iseq: 24,263 (33.6%) Top-6 not optimized method types for send_without_block (100.0% of total 4,524,016): iseq: 2,272,269 (50.2%) bmethod: 985,636 (21.8%) optimized: 949,704 (21.0%) alias: 310,747 ( 6.9%) null: 5,106 ( 0.1%) cfunc: 554 ( 0.0%) Top-13 not optimized instructions (100.0% of total 4,294,241): invokesuper: 2,375,446 (55.3%) invokeblock: 810,955 (18.9%) sendforward: 505,451 (11.8%) opt_eq: 451,754 (10.5%) opt_plus: 74,404 ( 1.7%) opt_minus: 36,228 ( 0.8%) opt_send_without_block: 21,792 ( 0.5%) opt_neq: 7,231 ( 0.2%) opt_mult: 6,752 ( 0.2%) opt_or: 3,753 ( 0.1%) opt_lt: 348 ( 0.0%) opt_ge: 91 ( 0.0%) opt_gt: 36 ( 0.0%) Top-9 send fallback reasons (100.0% of total 25,534,542): send_without_block_polymorphic: 9,723,469 (38.1%) send_no_profiles: 5,896,023 (23.1%) send_without_block_not_optimized_method_type: 4,524,016 (17.7%) not_optimized_instruction: 4,294,241 (16.8%) send_without_block_no_profiles: 998,947 ( 3.9%) send_not_optimized_method_type: 72,318 ( 0.3%) send_without_block_cfunc_array_variadic: 15,134 ( 0.1%) obj_to_string_not_string: 9,765 ( 0.0%) send_without_block_direct_too_many_args: 629 ( 0.0%) Top-8 unhandled YARV insns (100.0% of total 362,460): checkkeyword: 190,694 (52.6%) getclassvariable: 59,901 (16.5%) invokesuperforward: 49,503 (13.7%) getblockparam: 49,119 (13.6%) opt_duparray_send: 11,978 ( 3.3%) getconstant: 952 ( 0.3%) checkmatch: 290 ( 0.1%) once: 23 ( 0.0%) Top-3 compile error reasons (100.0% of total 3,798,744): register_spill_on_alloc: 3,495,669 (92.0%) register_spill_on_ccall: 184,712 ( 4.9%) exception_handler: 118,363 ( 3.1%) Top-15 side exit reasons (100.0% of total 10,637,319): compile_error: 3,798,744 (35.7%) guard_type_failure: 2,655,504 (25.0%) guard_shape_failure: 1,917,217 (18.0%) block_param_proxy_not_iseq_or_ifunc: 535,789 ( 5.0%) unhandled_kwarg: 455,492 ( 4.3%) patchpoint: 370,478 ( 3.5%) unhandled_yarv_insn: 362,460 ( 3.4%) unknown_newarray_send: 314,786 ( 3.0%) unhandled_splat: 122,071 ( 1.1%) unhandled_hir_insn: 83,066 ( 0.8%) block_param_proxy_modified: 19,193 ( 0.2%) guard_int_equals_failure: 1,914 ( 0.0%) obj_to_string_fallback: 566 ( 0.0%) guard_type_not_failure: 22 ( 0.0%) interrupt: 17 ( 0.0%) send_count: 62,495,067 dynamic_send_count: 25,534,542 (40.9%) optimized_send_count: 36,960,525 (59.1%) iseq_optimized_send_count: 18,582,072 (29.7%) inline_cfunc_optimized_send_count: 7,086,638 (11.3%) non_variadic_cfunc_optimized_send_count: 8,392,657 (13.4%) variadic_cfunc_optimized_send_count: 2,899,158 ( 4.6%) dynamic_getivar_count: 7,365,994 dynamic_setivar_count: 7,248,500 compiled_iseq_count: 4,780 failed_iseq_count: 463 compile_time: 816ms profile_time: 9ms gc_time: 11ms invalidation_time: 70ms vm_write_pc_count: 64,363,541 vm_write_sp_count: 63,022,221 vm_write_locals_count: 63,022,221 vm_write_stack_count: 63,022,221 vm_write_to_parent_iseq_local_count: 292,458 vm_read_from_parent_iseq_local_count: 6,850,977 code_region_bytes: 23,019,520 side_exit_count: 10,637,319 total_insn_count: 517,303,190 vm_insn_count: 160,562,103 zjit_insn_count: 356,741,087 ratio_in_zjit: 69.0% ``` </details> railsbench before: <details> ``` ***ZJIT: Printing ZJIT statistics on exit*** Top-20 not inlined C methods (66.1% of total 25,524,934): Hash#[]=: 1,700,237 ( 6.7%) String#getbyte: 1,572,123 ( 6.2%) String#<<: 1,494,022 ( 5.9%) Kernel#is_a?: 1,429,930 ( 5.6%) String#empty?: 1,370,323 ( 5.4%) Regexp#match?: 1,235,067 ( 4.8%) Kernel#respond_to?: 1,198,251 ( 4.7%) Hash#key?: 1,087,406 ( 4.3%) String#setbyte: 810,022 ( 3.2%) Integer#^: 766,624 ( 3.0%) Kernel#block_given?: 603,613 ( 2.4%) String#==: 590,409 ( 2.3%) Class#new: 506,216 ( 2.0%) Hash#delete: 455,288 ( 1.8%) BasicObject#!=: 428,771 ( 1.7%) Hash#fetch: 408,621 ( 1.6%) String#ascii_only?: 373,915 ( 1.5%) ObjectSpace::WeakKeyMap#[]: 287,957 ( 1.1%) NilClass#===: 277,244 ( 1.1%) Kernel#Array: 269,590 ( 1.1%) Top-20 not annotated C methods (66.8% of total 25,392,654): Hash#[]=: 1,700,416 ( 6.7%) String#getbyte: 1,572,123 ( 6.2%) Kernel#is_a?: 1,515,672 ( 6.0%) String#<<: 1,494,022 ( 5.9%) String#empty?: 1,370,478 ( 5.4%) Regexp#match?: 1,235,067 ( 4.9%) Kernel#respond_to?: 1,198,251 ( 4.7%) Hash#key?: 1,087,406 ( 4.3%) String#setbyte: 810,022 ( 3.2%) Integer#^: 766,624 ( 3.0%) Kernel#block_given?: 603,613 ( 2.4%) String#==: 601,115 ( 2.4%) Class#new: 506,216 ( 2.0%) Hash#delete: 455,288 ( 1.8%) BasicObject#!=: 428,876 ( 1.7%) Hash#fetch: 408,621 ( 1.6%) String#ascii_only?: 373,915 ( 1.5%) ObjectSpace::WeakKeyMap#[]: 287,957 ( 1.1%) NilClass#===: 277,244 ( 1.1%) Kernel#Array: 269,590 ( 1.1%) Top-2 not optimized method types for send (100.0% of total 186,159): iseq: 112,747 (60.6%) cfunc: 73,412 (39.4%) Top-6 not optimized method types for send_without_block (100.0% of total 8,142,248): iseq: 3,464,671 (42.6%) optimized: 2,632,884 (32.3%) bmethod: 1,290,701 (15.9%) alias: 706,020 ( 8.7%) null: 47,942 ( 0.6%) cfunc: 30 ( 0.0%) Top-11 not optimized instructions (100.0% of total 8,394,873): invokesuper: 5,602,274 (66.7%) invokeblock: 1,764,936 (21.0%) sendforward: 551,832 ( 6.6%) opt_eq: 441,959 ( 5.3%) opt_plus: 31,635 ( 0.4%) opt_send_without_block: 1,163 ( 0.0%) opt_lt: 372 ( 0.0%) opt_mult: 251 ( 0.0%) opt_ge: 193 ( 0.0%) opt_neq: 149 ( 0.0%) opt_or: 109 ( 0.0%) Top-8 send fallback reasons (100.0% of total 40,748,753): send_without_block_polymorphic: 12,933,923 (31.7%) send_no_profiles: 9,033,636 (22.2%) not_optimized_instruction: 8,394,873 (20.6%) send_without_block_not_optimized_method_type: 8,142,248 (20.0%) send_without_block_no_profiles: 1,839,228 ( 4.5%) send_without_block_cfunc_array_variadic: 215,046 ( 0.5%) send_not_optimized_method_type: 186,159 ( 0.5%) obj_to_string_not_string: 3,640 ( 0.0%) Top-9 unhandled YARV insns (100.0% of total 1,604,456): getclassvariable: 458,136 (28.6%) getblockparam: 455,921 (28.4%) checkkeyword: 265,425 (16.5%) invokesuperforward: 239,383 (14.9%) expandarray: 137,305 ( 8.6%) getconstant: 48,100 ( 3.0%) checkmatch: 149 ( 0.0%) once: 23 ( 0.0%) opt_duparray_send: 14 ( 0.0%) Top-3 compile error reasons (100.0% of total 5,570,130): register_spill_on_alloc: 4,994,130 (89.7%) exception_handler: 356,784 ( 6.4%) register_spill_on_ccall: 219,216 ( 3.9%) Top-13 side exit reasons (100.0% of total 12,412,181): compile_error: 5,570,130 (44.9%) unhandled_yarv_insn: 1,604,456 (12.9%) guard_shape_failure: 1,462,872 (11.8%) guard_type_failure: 845,891 ( 6.8%) block_param_proxy_not_iseq_or_ifunc: 765,968 ( 6.2%) unhandled_kwarg: 658,341 ( 5.3%) patchpoint: 504,437 ( 4.1%) unhandled_splat: 446,990 ( 3.6%) unknown_newarray_send: 332,740 ( 2.7%) unhandled_hir_insn: 160,205 ( 1.3%) block_param_proxy_modified: 59,589 ( 0.5%) obj_to_string_fallback: 553 ( 0.0%) interrupt: 9 ( 0.0%) send_count: 119,067,587 dynamic_send_count: 40,748,753 (34.2%) optimized_send_count: 78,318,834 (65.8%) iseq_optimized_send_count: 39,936,542 (33.5%) inline_cfunc_optimized_send_count: 12,857,358 (10.8%) non_variadic_cfunc_optimized_send_count: 19,722,584 (16.6%) variadic_cfunc_optimized_send_count: 5,802,350 ( 4.9%) dynamic_getivar_count: 10,980,323 dynamic_setivar_count: 12,962,726 compiled_iseq_count: 2,531 failed_iseq_count: 245 compile_time: 414ms profile_time: 21ms gc_time: 33ms invalidation_time: 5ms vm_write_pc_count: 129,093,714 vm_write_sp_count: 126,023,084 vm_write_locals_count: 126,023,084 vm_write_stack_count: 126,023,084 vm_write_to_parent_iseq_local_count: 385,461 vm_read_from_parent_iseq_local_count: 11,266,484 code_region_bytes: 12,156,928 side_exit_count: 12,412,181 total_insn_count: 866,780,158 vm_insn_count: 216,821,134 zjit_insn_count: 649,959,024 ratio_in_zjit: 75.0% ``` </details> railsbench after: <details> ``` ***ZJIT: Printing ZJIT statistics on exit*** Top-20 not inlined C methods (66.0% of total 25,597,895): Hash#[]=: 1,724,042 ( 6.7%) String#getbyte: 1,572,123 ( 6.1%) String#<<: 1,494,022 ( 5.8%) Kernel#is_a?: 1,429,946 ( 5.6%) String#empty?: 1,370,323 ( 5.4%) Regexp#match?: 1,235,067 ( 4.8%) Kernel#respond_to?: 1,198,251 ( 4.7%) Hash#key?: 1,087,406 ( 4.2%) String#setbyte: 810,022 ( 3.2%) Integer#^: 766,624 ( 3.0%) Kernel#block_given?: 603,613 ( 2.4%) String#==: 590,699 ( 2.3%) Class#new: 506,216 ( 2.0%) Hash#delete: 455,288 ( 1.8%) BasicObject#!=: 428,771 ( 1.7%) Hash#fetch: 408,621 ( 1.6%) String#ascii_only?: 373,915 ( 1.5%) ObjectSpace::WeakKeyMap#[]: 287,957 ( 1.1%) NilClass#===: 277,244 ( 1.1%) Kernel#Array: 269,590 ( 1.1%) Top-20 not annotated C methods (66.7% of total 25,465,615): Hash#[]=: 1,724,221 ( 6.8%) String#getbyte: 1,572,123 ( 6.2%) Kernel#is_a?: 1,515,688 ( 6.0%) String#<<: 1,494,022 ( 5.9%) String#empty?: 1,370,478 ( 5.4%) Regexp#match?: 1,235,067 ( 4.8%) Kernel#respond_to?: 1,198,251 ( 4.7%) Hash#key?: 1,087,406 ( 4.3%) String#setbyte: 810,022 ( 3.2%) Integer#^: 766,624 ( 3.0%) Kernel#block_given?: 603,613 ( 2.4%) String#==: 601,405 ( 2.4%) Class#new: 506,216 ( 2.0%) Hash#delete: 455,288 ( 1.8%) BasicObject#!=: 428,876 ( 1.7%) Hash#fetch: 408,621 ( 1.6%) String#ascii_only?: 373,915 ( 1.5%) ObjectSpace::WeakKeyMap#[]: 287,957 ( 1.1%) NilClass#===: 277,244 ( 1.1%) Kernel#Array: 269,590 ( 1.1%) Top-2 not optimized method types for send (100.0% of total 186,159): iseq: 112,747 (60.6%) cfunc: 73,412 (39.4%) Top-6 not optimized method types for send_without_block (100.0% of total 8,142,248): iseq: 3,464,671 (42.6%) optimized: 2,632,884 (32.3%) bmethod: 1,290,701 (15.9%) alias: 706,020 ( 8.7%) null: 47,942 ( 0.6%) cfunc: 30 ( 0.0%) Top-11 not optimized instructions (100.0% of total 8,442,456): invokesuper: 5,649,857 (66.9%) invokeblock: 1,764,936 (20.9%) sendforward: 551,832 ( 6.5%) opt_eq: 441,959 ( 5.2%) opt_plus: 31,635 ( 0.4%) opt_send_without_block: 1,163 ( 0.0%) opt_lt: 372 ( 0.0%) opt_mult: 251 ( 0.0%) opt_ge: 193 ( 0.0%) opt_neq: 149 ( 0.0%) opt_or: 109 ( 0.0%) Top-8 send fallback reasons (100.0% of total 40,796,314): send_without_block_polymorphic: 12,933,921 (31.7%) send_no_profiles: 9,033,616 (22.1%) not_optimized_instruction: 8,442,456 (20.7%) send_without_block_not_optimized_method_type: 8,142,248 (20.0%) send_without_block_no_profiles: 1,839,228 ( 4.5%) send_without_block_cfunc_array_variadic: 215,046 ( 0.5%) send_not_optimized_method_type: 186,159 ( 0.5%) obj_to_string_not_string: 3,640 ( 0.0%) Top-8 unhandled YARV insns (100.0% of total 1,467,151): getclassvariable: 458,136 (31.2%) getblockparam: 455,921 (31.1%) checkkeyword: 265,425 (18.1%) invokesuperforward: 239,383 (16.3%) getconstant: 48,100 ( 3.3%) checkmatch: 149 ( 0.0%) once: 23 ( 0.0%) opt_duparray_send: 14 ( 0.0%) Top-3 compile error reasons (100.0% of total 5,825,923): register_spill_on_alloc: 5,225,940 (89.7%) exception_handler: 356,784 ( 6.1%) register_spill_on_ccall: 243,199 ( 4.2%) Top-13 side exit reasons (100.0% of total 12,530,763): compile_error: 5,825,923 (46.5%) unhandled_yarv_insn: 1,467,151 (11.7%) guard_shape_failure: 1,462,876 (11.7%) guard_type_failure: 845,913 ( 6.8%) block_param_proxy_not_iseq_or_ifunc: 765,968 ( 6.1%) unhandled_kwarg: 658,341 ( 5.3%) patchpoint: 504,437 ( 4.0%) unhandled_splat: 446,990 ( 3.6%) unknown_newarray_send: 332,740 ( 2.7%) unhandled_hir_insn: 160,273 ( 1.3%) block_param_proxy_modified: 59,589 ( 0.5%) obj_to_string_fallback: 553 ( 0.0%) interrupt: 9 ( 0.0%) send_count: 119,163,569 dynamic_send_count: 40,796,314 (34.2%) optimized_send_count: 78,367,255 (65.8%) iseq_optimized_send_count: 39,911,967 (33.5%) inline_cfunc_optimized_send_count: 12,857,393 (10.8%) non_variadic_cfunc_optimized_send_count: 19,770,401 (16.6%) variadic_cfunc_optimized_send_count: 5,827,494 ( 4.9%) dynamic_getivar_count: 10,980,323 dynamic_setivar_count: 12,986,381 compiled_iseq_count: 2,523 failed_iseq_count: 252 compile_time: 420ms profile_time: 21ms gc_time: 30ms invalidation_time: 4ms vm_write_pc_count: 128,973,665 vm_write_sp_count: 125,926,968 vm_write_locals_count: 125,926,968 vm_write_stack_count: 125,926,968 vm_write_to_parent_iseq_local_count: 385,752 vm_read_from_parent_iseq_local_count: 11,267,766 code_region_bytes: 12,189,696 side_exit_count: 12,530,763 total_insn_count: 866,667,490 vm_insn_count: 217,813,201 zjit_insn_count: 648,854,289 ratio_in_zjit: 74.9% ``` </details>
2025-10-16ZJIT: Inline String#getbyte (#14842)Max Bernstein
2025-10-15ZJIT: Use rb_gc_disable() over rb_gc_disable_no_rest()Alan Wu
no_rest() trips an assert inside the GC when we allocate with the GC disabled this way: (gc_continue) ../src/gc/default/default.c:2029 (newobj_cache_miss+0x128) [0x105040048] ../src/gc/default/default.c:2370 (rb_gc_impl_new_obj+0x7c) [0x105036374] ../src/gc/default/default.c:2482 (newobj_of) ../src/gc.c:995 (rb_method_entry_alloc+0x40) [0x1051e6c64] ../src/vm_method.c:1102 (rb_method_entry_complement_defined_class) ../src/vm_method.c:1180 (prepare_callable_method_entry+0x14c) [0x1051e87b8] ../src/vm_method.c:1728 (callable_method_entry_or_negative+0x1e8) [0x1051e809c] ../src/vm_method.c:1874 It's tries to continue the GC because it was out of space. Looks like it's not safe to allocate new objects after using rb_gc_disable_no_rest(); existing usages use it for malloc calls.
2025-10-15ZJIT: Never yield to the GC while compilingAlan Wu
This fixes a reliable "ZJIT saw a dead object" repro on my machine, and should fix the flaky ones on CI. The code for disabling the GC is the same as the code in newobj_of(). See: https://github.com/ruby/ruby/actions/runs/18511676257/job/52753782036
2025-10-14ZJIT: Include GC object dump when seeing dead objectsAlan Wu
Strictly more info than just the builtin_type from `assert_ne!`. Old: assertion `left != right` failed: ZJIT should only see live objects left: 0 right: 0 New: ZJIT saw a dead object. T_type=0, out-of-heap:0x0000000110d4bb40 Also, the new `VALUE::obj_info` is more flexible for print debugging than the dump_info() it replaces. It now allows you to use it as part of a `format!` string instead of always printing to stderr for you.
2025-10-10ZJIT: Add Insn:: ArrayArefFixnum to accelerate Array#[] (#14717)Aiden Fox Ivey
* ZJIT: Add Insn:: ArrayArefFixnum to accelerate Array#[] * ZJIT: Use result from GuardType in ArrayArefFixnum * ZJIT: Unbox index for aref_fixnum * ZJIT: Change condition and add ArrayArefFixnum test * ZJIT: Fix ArrayArefFixnum display for InsnPrinter * ZJIT: Change insta test
2025-10-09ZJIT: Name enum for bindgen (#14802)Max Bernstein
Relying on having the same compiler version and behavior across platforms is brittle, as Kokubun points out. Instead, name the enum so we don't have to rely on gensym stability. Fix https://github.com/Shopify/ruby/issues/787
2025-09-30ZJIT: Add --zjit-trace-exits (#14640)Aiden Fox Ivey
Add side exit tracing functionality for ZJIT
2025-09-19ZJIT: Expand the list of safe allocatorsMax Bernstein
It's not just the default allocator; other allocators are also leaf.
2025-09-18ZJIT: Support variadic C calls (#14575)Stan Lo
* ZJIT: Support variadic C calls This reduces the `dynamic_send_count` in `liquid-render` by ~21% * ZJIT: Reuse gen_push_frame * ZJIT: Avoid optimizing variadic C call when tracing is enabled
2025-09-17ZJIT: Prevent custom allocator in ObjectAllocClassMax Bernstein
2025-09-17ZJIT: Const-fold IsMethodCfuncMax Bernstein
2025-09-16ZJIT: Print local names in FrameState (#14571)Max Bernstein
2025-09-12ZJIT: Share more code with YJIT in jit.c (#14520)Takashi Kokubun
* ZJIT: Share more code with YJIT in jit.c * Fix ZJIT references to JIT
2025-09-11ZJIT, YJIT: Drop "// From xxx.h" comments in bindgen (#14519)Takashi Kokubun
2025-09-08ZJIT: Fix backtraces on opt_new (#14461)Takashi Kokubun
2025-09-05ZJIT: Invalidate local variables on EP escape (#14448)Takashi Kokubun
2025-09-04ZJIT: Add patchpoint for TracePoint (#14420)Stan Lo
ZJIT: Add patchpoint for TracePoint activation Co-authored-by: Takashi Kokubun <takashikkbn@gmail.com> Co-authored-by: Alan Wu <XrXr@users.noreply.github.com>
2025-09-02ZJIT: Clear jit entry from iseqs after TracePoint activation (#14407)Stan Lo
ZJIT: Remove JITed code after TracePoint is enabled
2025-08-29Add rb_jit_vm_unlock and share it in ZJIT and YJITStan Lo
2025-08-29Add rb_jit_vm_lock_then_barrier and share it in ZJIT and YJITStan Lo
2025-08-29Add rb_jit_multi_ractor_p and share it in ZJIT and YJITStan Lo
2025-08-29ZJIT: Specialize monomorphic GetIvar (#14388)Max Bernstein
Specialize monomorphic `GetIvar` into: * `GuardType(HeapObject)` * `GuardShape` * `LoadIvarEmbedded` or `LoadIvarExtended` This requires profiling self for `getinstancevariable` (it's not on the operand stack). This also optimizes `GetIvar`s that happen as a result of inlining `attr_reader` and `attr_accessor`. Also move some (newly) shared JIT helpers into jit.c.
2025-08-28ZJIT: Generate code for ArrayExtendMax Bernstein
2025-08-28ZJIT: Generate code for DefinedIvarMax Bernstein
2025-08-27ZJIT: Implement side exit stats (#14357)Takashi Kokubun
2025-08-21Remove unused SPECIAL_CONST_SHAPE_IDÉtienne Barrié
Its usage was removed in 306d50811dd060d876d1eb364a0d5e6106f5e4f1.
2025-08-19ZJIT: Compile toregexp (#14200)Daniel Colson
`toregexp` is fairly similar to `concatstrings`, so this commit extracts a helper for pushing and popping operands on the native stack. There's probably opportunity to move some of this into lir (e.g. Alan suggested a push_many that could use STP on ARM to push 2 at a time), but I might save that for another day.
2025-08-05ZJIT: Profile type+shape distributions (#13901)Max Bernstein
ZJIT uses the interpreter to take type profiles of what objects pass through the code. It stores a compressed record of the history per opcode for the opcodes we select. Before this change, we re-used the HIR Type data-structure, a shallow type lattice, to store historical type information. This was quick for bringup but is quite lossy as profiles go: we get one bit per built-in type seen, and if we see a non-built-in type in addition, we end up with BasicObject. Not very helpful. Additionally, it does not give us any notion of cardinality: how many of each type did we see? This change brings with it a much more interesting slice of type history: a histogram. A Distribution holds a record of the top-N (where N is fixed at Ruby compile-time) `(Class, ShapeId)` pairs and their counts. It also holds an *other* count in case we see more than N pairs. Using this distribution, we can make more informed decisions about when we should use type information. We can determine if we are strictly monomorphic, very nearly monomorphic, or something else. Maybe the call-site is polymorphic, so we should have a polymorphic inline cache. Exciting stuff. I also plumb this new distribution into the HIR part of the compilation pipeline.
2025-07-29ZJIT: Add --zjit-stats (#14034)Takashi Kokubun
2025-07-16ZJIT: Add missing write barrier in profiling (GH-13922)Alan Wu
Fixes `TestZJIT::test_require_rubygems`. It was crashing locally due to false collection of a live object. See <https://alanwu.space/post/write-barrier/>. Co-authored-by: Max Bernstein <max@bernsteinbear.com> Co-authored-by: Takashi Kokubun <takashi.kokubun@shopify.com> Co-authored-by: Stan Lo <stan.lo@shopify.com>
2025-07-16ZJIT: Profile each instruction at most num_profiles times (#13903)Takashi Kokubun
* ZJIT: Profile each instruction at most num_profiles times * Use saturating_add for num_profiles
generated by cgit v1.2.3 (git 2.25.1) at 2026-01-15 21:09:02 +0000