-
Enhancement
-
Resolution: Fixed
-
P4
-
20
-
b22
G1 feeds the various number sequences used for prediction initial values.
These initial values are derived from now non-existent very old machines, so startup is very slow, and the first few garbage collections basically always use a too small young gen.
Improve here, possibly without depending too much on always-to-be-updated start values.
These initial values are derived from now non-existent very old machines, so startup is very slow, and the first few garbage collections basically always use a too small young gen.
Improve here, possibly without depending too much on always-to-be-updated start values.
- relates to
-
JDK-8343086 [BACKOUT] JDK-8295269 G1: Improve slow startup due to predictor initialization
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
- links to
-
Commit(master)
openjdk/jdk/c202a2f7
-
Review(master)
openjdk/jdk/21518