Skip to content

Better uop coverage in the JIT optimizer #131798

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
brandtbucher opened this issue Mar 27, 2025 · 16 comments
Open

Better uop coverage in the JIT optimizer #131798

brandtbucher opened this issue Mar 27, 2025 · 16 comments
Assignees
@brandtbucher
Labels
interpreter-core (Objects, Python, Grammar, and Parser dirs) performance Performance or resource usage topic-JIT type-feature A feature request or enhancement

Comments

@brandtbucher
Copy link
Member

brandtbucher commented Mar 27, 2025 •
edited by bedevere-app bot
Loading

Out of 263 total uops, 155 of these are ignored by the tier two optimizer. These represent over half of all uops by dynamic execution count.

This issue will serve as a checklist for auditing these missing uops, and adding them where they make sense. At first glance, there's quite a bit of potential here... especially around ability to narrow known output types (like _CONTAINS_OP_SET), and the ability to narrow and remove guards on input types (like _BINARY_OP_SUBSCR_LIST_INT). As I'm going through, I'll cross out anything that doesn't seem like it makes sense to add.

First, here are the 53 missing uops that each represent at least 0.1% of all uops executed:

  • _SET_IP (12.1%)
  • _CHECK_VALIDITY (10.1%)
  • _CHECK_VALIDITY_AND_SET_IP (6.5%)
  • _CHECK_PERIODIC (3.1%)
  • _MAKE_WARM (2.8%)
  • _START_EXECUTOR (1.7%)
  • _GUARD_NOS_INT (1.5%)
  • _BINARY_OP_SUBSCR_LIST_INT (1.0%)
  • _CHECK_FUNCTION (1.0%)
  • _CHECK_MANAGED_OBJECT_HAS_VALUES (0.7%)
  • _ITER_CHECK_LIST (0.7%)
  • _CONTAINS_OP_SET (0.6%)
  • _FOR_ITER_TIER_TWO (0.6%)
  • _GUARD_NOT_EXHAUSTED_LIST (0.6%)
  • _ITER_NEXT_LIST_TIER_TWO (0.6%)
  • _SAVE_RETURN_OFFSET (0.6%)
  • _CALL_LEN (0.5%)
  • _CALL_LIST_APPEND (0.5%)
  • _POP_TOP (0.5%)
  • _RESUME_CHECK (0.5%)
  • _BINARY_OP_SUBSCR_STR_INT (0.4%)
  • _GUARD_DORV_VALUES_INST_ATTR_FROM_DICT (0.4%)
  • _GUARD_KEYS_VERSION (0.4%)
  • _BINARY_OP_SUBSCR_DICT (0.3%)
  • _CALL_BUILTIN_FAST (0.3%)
  • _CHECK_STACK_SPACE_OPERAND (0.3%)
  • _GET_ITER (0.3%)
  • _STORE_SUBSCR (0.3%)
  • _GUARD_NOT_EXHAUSTED_RANGE (0.2%)
  • _BINARY_SLICE (0.2%)
  • _BUILD_LIST (0.2%)
  • _CALL_BUILTIN_O (0.2%)
  • _CALL_NON_PY_GENERAL (0.2%)
  • _CHECK_IS_NOT_PY_CALLABLE (0.2%)
  • _GUARD_NOS_FLOAT (0.2%)
  • _ITER_CHECK_RANGE (0.2%)
  • _ITER_CHECK_TUPLE (0.2%)
  • _LOAD_DEREF (0.2%)
  • _STORE_SUBSCR_LIST_INT (0.2%)
  • _BINARY_OP_EXTEND (0.1%)
  • _CALL_ISINSTANCE (0.1%)
  • _CALL_METHOD_DESCRIPTOR_FAST (0.1%)
  • _CALL_METHOD_DESCRIPTOR_FAST_WITH_KEYWORDS (0.1%)
  • _CALL_METHOD_DESCRIPTOR_NOARGS (0.1%)
  • _CALL_TYPE_1 (0.1%)
  • _CHECK_ATTR_CLASS (0.1%)
  • _CONTAINS_OP_DICT (0.1%)
  • _GUARD_BINARY_OP_EXTEND (0.1%)
  • _GUARD_NOT_EXHAUSTED_TUPLE (0.1%)
  • _ITER_NEXT_TUPLE (0.1%)
  • _LIST_APPEND (0.1%)
  • _STORE_ATTR_SLOT (0.1%)
  • _STORE_SUBSCR_DICT (0.1%)

And here are the 102 missing uops that are less than 0.1%. These are less important, but still may net us some wins on individual benchmarks:

  • _BINARY_OP_SUBSCR_CHECK_FUNC
  • _BINARY_OP_SUBSCR_TUPLE_INT
  • _BUILD_MAP
  • _BUILD_SET
  • _BUILD_SLICE
  • _BUILD_STRING
  • _CALL_BUILTIN_CLASS
  • _CALL_BUILTIN_FAST_WITH_KEYWORDS
  • _CALL_INTRINSIC_1
  • _CALL_INTRINSIC_2
  • _CALL_KW_NON_PY
  • _CALL_METHOD_DESCRIPTOR_O
  • _CALL_STR_1
  • _CALL_TUPLE_1
  • _CHECK_ATTR_METHOD_LAZY_DICT
  • _CHECK_EG_MATCH
  • _CHECK_EXC_MATCH
  • _CHECK_FUNCTION_VERSION_INLINE
  • _CHECK_FUNCTION_VERSION_KW
  • _CHECK_IS_NOT_PY_CALLABLE_KW
  • _CHECK_METHOD_VERSION
  • _CHECK_METHOD_VERSION_KW
  • _CHECK_PERIODIC_IF_NOT_YIELD_FROM
  • _CONVERT_VALUE
  • _COPY_FREE_VARS
  • _DELETE_ATTR
  • _DELETE_DEREF
  • _DELETE_FAST
  • _DELETE_GLOBAL
  • _DELETE_NAME
  • _DELETE_SUBSCR
  • _DEOPT
  • _DICT_MERGE
  • _DICT_UPDATE
  • _END_FOR
  • _END_SEND
  • _ERROR_POP_N
  • _EXIT_INIT_CHECK
  • _EXPAND_METHOD
  • _EXPAND_METHOD_KW
  • _FATAL_ERROR
  • _FORMAT_SIMPLE
  • _FORMAT_WITH_SPEC
  • _GET_AITER
  • _GET_ANEXT
  • _GET_AWAITABLE
  • _GET_LEN
  • _GET_YIELD_FROM_ITER
  • _GUARD_DORV_NO_DICT
  • _GUARD_GLOBALS_VERSION
  • _GUARD_TOS_FLOAT
  • _GUARD_TOS_INT
  • _GUARD_TYPE_VERSION_AND_LOCK
  • _IMPORT_FROM
  • _IMPORT_NAME
  • _IS_NONE
  • _LIST_EXTEND
  • _LOAD_ATTR_NONDESCRIPTOR_NO_DICT
  • _LOAD_ATTR_NONDESCRIPTOR_WITH_VALUES
  • _LOAD_BUILD_CLASS
  • _LOAD_COMMON_CONSTANT
  • _LOAD_FAST_LOAD_FAST
  • _LOAD_FROM_DICT_OR_DEREF
  • _LOAD_GLOBAL
  • _LOAD_GLOBAL_BUILTINS
  • _LOAD_GLOBAL_MODULE
  • _LOAD_LOCALS
  • _LOAD_NAME
  • _LOAD_SUPER_ATTR_ATTR
  • _LOAD_SUPER_ATTR_METHOD
  • _MAKE_CALLARGS_A_TUPLE
  • _MAKE_CELL
  • _MAKE_FUNCTION
  • _MAP_ADD
  • _MATCH_CLASS
  • _MATCH_KEYS
  • _MATCH_MAPPING
  • _MATCH_SEQUENCE
  • _MAYBE_EXPAND_METHOD_KW
  • _NOP
  • _POP_EXCEPT
  • _POP_TWO_LOAD_CONST_INLINE_BORROW
  • _PUSH_EXC_INFO
  • _PUSH_NULL_CONDITIONAL
  • _SETUP_ANNOTATIONS
  • _SET_ADD
  • _SET_FUNCTION_ATTRIBUTE
  • _SET_UPDATE
  • _STORE_ATTR
  • _STORE_ATTR_INSTANCE_VALUE
  • _STORE_ATTR_WITH_HINT
  • _STORE_DEREF
  • _STORE_FAST_LOAD_FAST
  • _STORE_FAST_STORE_FAST
  • _STORE_GLOBAL
  • _STORE_NAME
  • _STORE_SLICE
  • _TIER2_RESUME_CHECK
  • _UNARY_INVERT
  • _UNARY_NEGATIVE
  • _UNPACK_SEQUENCE_LIST
  • _WITH_EXCEPT_START

Linked PRs
@brandtbucher brandtbucher added 3.14 new features, bugs and security fixes interpreter-core (Objects, Python, Grammar, and Parser dirs) performance Performance or resource usage topic-JIT labels Mar 27, 2025
@brandtbucher brandtbucher self-assigned this Mar 27, 2025
@bedevere-app bedevere-app bot mentioned this issue Mar 27, 2025
Merged
@brandtbucher brandtbucher mentioned this issue Mar 28, 2025
Merged
brandtbucher pushed a commit that referenced this issue Mar 30, 2025
@fluhus
GH-131798: Remove type checks for _TO_BOOL_STR (GH-131816)
685fd74
@brandtbucher
Copy link
Member Author

brandtbucher commented Apr 1, 2025 •
edited
Loading

@diegorusso is going to add _CALL_LEN.

brandtbucher added a commit that referenced this issue Apr 1, 2025
@brandtbucher
GH-131798: Allow the JIT to remove more int/float/str guards (GH-131800)
1a9d4a1
@brandtbucher brandtbucher mentioned this issue Apr 3, 2025
Open
17 tasks
@brandtbucher
Copy link
Member Author

brandtbucher commented Apr 3, 2025 •
edited
Loading

@fluhus is going to add _BINARY_SLICE.

@brandtbucher
Copy link
Member Author

@Klaus117 is going to improve _TO_BOOL_INT.

@bedevere-app bedevere-app bot mentioned this issue Apr 3, 2025
Merged
brandtbucher pushed a commit that referenced this issue Apr 5, 2025
@tomasr8
GH-131798: Narrow the result of _CONTAINS_OP_SET to bool in the JIT (G…
85bc489 
…H-132057)
@picnixz picnixz added type-feature A feature request or enhancement and removed 3.14 new features, bugs and security fixes labels Apr 5, 2025
@bedevere-app bedevere-app bot mentioned this issue Apr 6, 2025
Merged
@brandtbucher
Copy link
Member Author

@Zheaoli is going to add _CONTAINS_OP_DICT.

@Zheaoli Zheaoli mentioned this issue Apr 8, 2025
Merged
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 8, 2025
@Zheaoli
pythongh-131798: Allow the JIT to remove an extra _TO_BOOL_BOOL for _…
9053639 
…CONTAINS_OP_DICT

Signed-off-by: Manjusaka <[email protected]>
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 8, 2025
@Zheaoli
pythongh-131798: Allow the JIT to remove an extra _TO_BOOL_BOOL for _…
adf36c3 
…CONTAINS_OP_DICT

Signed-off-by: Manjusaka <[email protected]>
brandtbucher pushed a commit that referenced this issue Apr 8, 2025
@tomasr8
GH-131798: Narrow the result type of _BINARY_OP_SUBSCR_STR_INT to str…
71009cb 
… in the JIT (GH-132153)
@bedevere-app bedevere-app bot mentioned this issue Apr 8, 2025
Open
brandtbucher pushed a commit that referenced this issue Apr 8, 2025
@Zheaoli @tomasr8
GH-131798: Narrow the result of _CONTAINS_OP_DICT to bool in the JIT (G…
d753d8a 
…H-132269)

Co-authored-by: Tomas R. <[email protected]>
@Zheaoli
Copy link
Contributor

Zheaoli commented Apr 8, 2025

I think I can work on _BINARY_OP_SUBSCR_LIST_INT and _BINARY_OP_SUBSCR_DICT

@brandtbucher
Copy link
Member Author

I think I can work on _BINARY_OP_SUBSCR_LIST_INT and _BINARY_OP_SUBSCR_DICT

Sorry, I already have a branch to do the guards for these (and a couple others) that I was going to up in a minute! I'll tag you for review though.

@bedevere-app bedevere-app bot mentioned this issue Apr 8, 2025
Merged
brandtbucher added a commit that referenced this issue Apr 9, 2025
@brandtbucher
GH-131798: Remove JIT guards for dict, frozenset, list, set, and tuple (
20926c7 
GH-132289)
@brandtbucher
Copy link
Member Author

@tomasr8 is going to add _CALL_STR_1, _CALL_TUPLE_1, and _CALL_TYPE_1.

@Zheaoli, want to take _BUILD_LIST, _BUILD_MAP, _BUILD_SET, _BUILD_SLICE, and _BUILD_STRING? For all but _BUILD_SLICE and _BUILD_STRING, I think we can only set the type of the output (sym_new_type). For _BUILD_SLICE and _BUILD_STRING, we may be able to have a constant output if the items are constant (sym_is_const/sym_get_const/sym_new_const). Maybe one PR for the first three, and separate PRs for _BUILD_SLICE and _BUILD_STRING?

@brandtbucher
Copy link
Member Author

Yep, your understanding is correct!

This was referenced Apr 11, 2025
Merged
Merged
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 12, 2025
@Zheaoli
pythongh-131798: Use sym_new_type instead of sym_new_not_null for…
429a534 
… _BUILD_LIST, _BUILD_SET, _BUILD_MAP

Signed-off-by: Manjusaka <[email protected]>
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 12, 2025
@Zheaoli
pythongh-131798: Use sym_new_type instead of sym_new_not_null for…
d2809ff 
… _BUILD_LIST, _BUILD_SET, _BUILD_MAP

Signed-off-by: Manjusaka <[email protected]>
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 15, 2025
@Zheaoli
pythongh-131798: Add JIT path for _BUILD_STRING
6dc906f 
Signed-off-by: Manjusaka <[email protected]>
@bedevere-app bedevere-app bot mentioned this issue Apr 15, 2025
Merged
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 15, 2025
@Zheaoli
pythongh-131798: Add JIT path for _BUILD_STRING
11ea46e 
Signed-off-by: Manjusaka <[email protected]>
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 15, 2025
@Zheaoli
pythongh-131798: Add JIT path for _BUILD_STRING
ab00ff4 
Signed-off-by: Manjusaka <[email protected]>
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 15, 2025
@Zheaoli
pythongh-131798: Add JIT path for _BUILD_STRING
445ec7d 
Signed-off-by: Manjusaka <[email protected]>
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@fluhus @seehwan80
pythonGH-131798: Remove type checks for _TO_BOOL_STR (pythonGH-131816)
2cc01e1
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@brandtbucher @seehwan80
pythonGH-131798: Allow the JIT to remove more int/float/str guards (p…
4cd6063 
…ythonGH-131800)
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@tomasr8 @seehwan80
pythonGH-131798: Narrow the result of _CONTAINS_OP_SET to bool in the…
0c01fc8 
… JIT (pythonGH-132057)
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@tomasr8 @seehwan80
pythonGH-131798: Narrow the result type of _BINARY_OP_SUBSCR_STR_INT …
a94ad48 
…to str in the JIT (pythonGH-132153)
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@Zheaoli @tomasr8 @seehwan80
pythonGH-131798: Narrow the result of _CONTAINS_OP_DICT to bool in th…
18edba8 
…e JIT (pythonGH-132269)

Co-authored-by: Tomas R. <[email protected]>
seehwan pushed a commit to seehwan/cpython that referenced this issue Apr 16, 2025
@brandtbucher @seehwan80
pythonGH-131798: Remove JIT guards for dict, frozenset, list, set, an…
cd4602d 
…d tuple (pythonGH-132289)
Zheaoli added a commit to Zheaoli/cpython that referenced this issue Apr 16, 2025
@Zheaoli
pythongh-131798: Use sym_new_type instead of sym_new_not_null for…
9be0d7f 
… _BUILD_LIST, _BUILD_SET, _BUILD_MAP

Signed-off-by: Manjusaka <[email protected]>
Fidget-Spinner pushed a commit that referenced this issue Apr 16, 2025
@Zheaoli
gh-131798: Use sym_new_type instead of sym_new_not_null for `_BUI…
b9e88ff 
…LD_LIST`, `_BUILD_SLICE`, and `_BUILD_MAP` (GH-132434)

---------

Signed-off-by: Manjusaka <[email protected]>
brandtbucher pushed a commit that referenced this issue Apr 22, 2025
@tomasr8
GH-131798: JIT: Split CALL_TYPE_1 into several uops (GH-132419)
a6a3dbb
This was referenced Apr 23, 2025
Merged
Merged
brandtbucher pushed a commit that referenced this issue Apr 24, 2025
@tomasr8
GH-131798: Split up and optimize CALL_STR_1 in the JIT (GH-132849)
0a387b3
brandtbucher pushed a commit that referenced this issue Apr 24, 2025
@tomasr8
GH-131798: Split up and optimize CALL_TUPLE_1 in the JIT (GH-132851)
08e3389
diegorusso added a commit to diegorusso/cpython that referenced this issue Apr 25, 2025
@diegorusso
pythongh-131798: JIT: Narrow the return type of _CALL_LEN to int
ee947e3 
Reduce unnecessary guards whenever `len()` is called and used
in arithmetic operations.
@bedevere-app bedevere-app bot mentioned this issue Apr 25, 2025
Merged
@Fidget-Spinner Fidget-Spinner mentioned this issue Apr 25, 2025
Open
Fidget-Spinner pushed a commit that referenced this issue Apr 25, 2025
@diegorusso @tekknolagi
gh-131798: JIT: Narrow the return type of _CALL_LEN to int (#132940)
4c20f46 
Reduce unnecessary guards whenever `len()` is called and used
after.

Co-authored-by: Max Bernstein <[email protected]>
@bedevere-app bedevere-app bot mentioned this issue Apr 26, 2025
Merged
Fidget-Spinner pushed a commit that referenced this issue Apr 26, 2025
@tomasr8
gh-131798: JIT: Propagate the result in _BINARY_OP_SUBSCR_TUPLE_INT (…
5e96e4f 
…GH-133003)
Fidget-Spinner pushed a commit that referenced this issue Apr 27, 2025
@Zheaoli
gh-131798: JIT - Use sym_new_type instead of sym_new_not_null for…
614d792 
… _BUILD_STRING, _BUILD_SET (GH-132564)

Signed-off-by: Manjusaka <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@brandtbucher brandtbucher

Labels
interpreter-core (Objects, Python, Grammar, and Parser dirs) performance Performance or resource usage topic-JIT type-feature A feature request or enhancement
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
6 participants
@Zheaoli @picnixz @runzh-crypto @splasky @Funnyyanne @brandtbucher