Skip to content

Specialize long tail of binary operations using a table. #100239

Description

@markshannon

There is a desire to specialize the remaining binary operations (including binary subscript).
However adding more and more specialized instructions is likely to make performance worse.

This idea is to have a lookup table of types pairs and function pointers. This is less efficient than inlining the code, but more extensible.

A single instruction can then support up to 256 specializations.
This will only work for immutable classes.

structtable_entry {
PyTypeObject*left;
PyTypeObject*left;
binaryfunc*func;
};
TARGET(BINARY_OP_TABLE) {
PyObject*lhs=SECOND();
PyObject*rhs=TOP();
Cache*cache=GET_CACHE();
structtable_entry*entry=&THE_TABLE[cache->table_index];
DEOPT_IF(Py_TYPE(lhs) !=entry->left);
DEOPT_IF(Py_TYPE(rhs) !=entry->right);
PyObject*res=entry->func(lhs, rhs);
if (res==NULL) {
goto error;
}
STACK_SHRINK(1);
Py_DECREF(lhs);
Py_DECREF(rhs);
SET_TOP(res);
DISPATCH();
}

An ancillary mapping of (left, right) -> index will be needed for efficient specialization.

It is probably worth keeping the most common operations int + int, float + float, etc. inline.

We can replace BINARY_SUBSCR with BINARY_OP ([]) to allow effective specialization of BINARY_SUBSCR
E.g. subscripting array.array[int] can be handled with the registration mechanism described below.

Registering binary functions at runtime

faster-cpython/ideas#162

Linked PRs

Metadata

Metadata

Assignees

Labels

interpreter-core(Objects, Python, Grammar, and Parser dirs)performancePerformance or resource usagetype-featureA feature request or enhancement

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions