@@ -116,12 +116,12 @@ after the main thread is running. The following objects are immortalized:
116116* :ref: `classes <classes >` (type objects)
117117
118118Because immortal objects are never deallocated, applications that create many
119- objects of these types may see increased memory usage. This is expected to be
120- addressed in the 3.14 release.
119+ objects of these types may see increased memory usage. Work to further reduce
120+ this overhead continued in the 3.14 release.
121121
122122Additionally, numeric and string literals in the code as well as strings
123- returned by :func: `sys.intern ` are also immortalized. This behavior is
124- expected to remain in the 3.14 free-threaded build.
123+ returned by :func: `sys.intern ` are also immortalized. This behavior remains
124+ in the 3.14 free-threaded build.
125125
126126
127127Frame objects
@@ -152,9 +152,10 @@ Programs that spend most of their time in C extensions or I/O will see
152152less of an impact. The largest impact is because the specializing adaptive
153153interpreter (:pep: `659 `) is disabled in the free-threaded build. We expect
154154to re-enable it in a thread-safe way in the 3.14 release. This overhead is
155- expected to be reduced in upcoming Python release. We are aiming for an
156- overhead of 10% or less on the pyperformance suite compared to the default
157- GIL-enabled build.
155+ expected to be reduced in upcoming Python release. This overhead **was reduced **
156+ in the 3.14 release. Reducing overhead further **remains an active development goal **,
157+ with an aim for 10% or less on the pyperformance suite compared to the default GIL-enabled
158+ build.
158159
159160
160161Behavioral changes
0 commit comments