Skip to content

gh-141510: Optimize hash(frozendict) #144919

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

Merged
vstinner merged 2 commits into from
Feb 17, 2026
+52 −18
Merged

gh-141510: Optimize hash(frozendict) #144919

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b13add2
gh-141510: Optimize hash(frozendict)
vstinner Feb 17, 2026
390119d
Update comment
vstinner Feb 17, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions Lib/test/test_dict.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1775,6 +1775,15 @@ class MyFrozenDict(frozendict):
d = MyFrozenDict(x=1, y=2)
self.assertEqual(repr(d), "MyFrozenDict({'x': 1, 'y': 2})")

def test_hash(self):
# hash() doesn't rely on the items order
self.assertEqual(hash(frozendict(x=1, y=2)),
hash(frozendict(y=2, x=1)))

fd = frozendict(x=[1], y=[2])
with self.assertRaisesRegex(TypeError, "unhashable type: 'list'"):
hash(fd)


if __name__ == "__main__":
unittest.main()
56 changes: 39 additions & 17 deletions Objects/dictobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7869,33 +7869,55 @@ frozendict_repr(PyObject *self)
return res;
}

static Py_uhash_t
_shuffle_bits(Py_uhash_t h)
{
return ((h ^ 89869747UL) ^ (h << 16)) * 3644798167UL;
}

// Code copied from frozenset_hash()
static Py_hash_t
frozendict_hash(PyObject *op)
{
PyFrozenDictObject *self = _PyFrozenDictObject_CAST(op);
Py_hash_t hash = FT_ATOMIC_LOAD_SSIZE_RELAXED(self->ma_hash);
if (hash != -1) {
return hash;
Py_hash_t shash = FT_ATOMIC_LOAD_SSIZE_RELAXED(self->ma_hash);
Copy link
Member

@corona10 corona10 Feb 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why atomic operation is needed?

Copy link
Member Author

@vstinner vstinner Feb 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

PyFrozenDictObject.ma_hash is mutable and so it needs a lock to handle properly concurrent access.

Copy link
Member Author

@vstinner vstinner Feb 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's the same for the Unicode strings hash: the string is immutable, but the hash member is mutable. So unicode_hash() uses atomic operations to get and set the hash member.

Copy link
Member

@corona10 corona10 Feb 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah got it

Copy link
Contributor

@kumaraditya303 kumaraditya303 Feb 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah, the hash is "lazily" computed so we need to use relaxed atomics here.

if (shash != -1) {
return shash;
}

PyObject *items = _PyDictView_New(op, &PyDictItems_Type);
if (items == NULL) {
return -1;
}
PyObject *frozenset = PyFrozenSet_New(items);
Py_DECREF(items);
if (frozenset == NULL) {
return -1;
PyDictObject *mp = _PyAnyDict_CAST(op);
Py_uhash_t hash = 0;

PyObject *key, *value; // borrowed refs
Py_ssize_t pos = 0;
while (PyDict_Next(op, &pos, &key, &value)) {
Py_hash_t key_hash = PyObject_Hash(key);
if (key_hash == -1) {
return -1;
}
hash ^= _shuffle_bits(key_hash);

Py_hash_t value_hash = PyObject_Hash(value);
if (value_hash == -1) {
return -1;
}
hash ^= _shuffle_bits(value_hash);
}

hash = PyObject_Hash(frozenset);
Py_DECREF(frozenset);
if (hash == -1) {
return -1;
/* Factor in the number of active entries */
hash ^= ((Py_uhash_t)mp->ma_used + 1) * 1927868237UL;

/* Disperse patterns arising in nested frozendicts */
hash ^= (hash >> 11) ^ (hash >> 25);
hash = hash * 69069U + 907133923UL;

/* -1 is reserved as an error code */
if (hash == (Py_uhash_t)-1) {
hash = 590923713UL;
}

FT_ATOMIC_STORE_SSIZE_RELAXED(self->ma_hash, hash);
return hash;
FT_ATOMIC_STORE_SSIZE_RELAXED(self->ma_hash, (Py_hash_t)hash);
return (Py_hash_t)hash;
}


Expand Down
5 changes: 4 additions & 1 deletion Objects/setobject.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -964,7 +964,10 @@ _shuffle_bits(Py_uhash_t h)
This hash algorithm can be used on either a frozenset or a set.
When it is used on a set, it computes the hash value of the equivalent
frozenset without creating a new frozenset object. */
frozenset without creating a new frozenset object.
If you update this code, update also frozendict_hash() which copied this
code. */

static Py_hash_t
frozenset_hash_impl(PyObject *self)
Expand Down
Loading