-
Enhancement
-
Resolution: Fixed
-
P4
-
1.4.1
-
b07
-
x86
-
windows_2000
-
Not verified
| Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
|---|---|---|---|---|---|---|
| JDK-2200255 | 7 | Tom Rodriguez | P4 | Closed | Fixed | b112 |
| JDK-2205832 | 6u25 | Tom Rodriguez | P4 | Closed | Fixed | b01 |
| JDK-2201466 | 6u23 | Tom Rodriguez | P4 | Resolved | Fixed | b02 |
| JDK-2200008 | 6u22m | Tom Rodriguez | P4 | Resolved | Fixed | b02 |
| JDK-2199986 | 6u21p | Tom Rodriguez | P4 | Resolved | Fixed | b04 |
| JDK-2200256 | hs20 | Tom Rodriguez | P4 | Closed | Fixed | b01 |
DESCRIPTION OF THE PROBLEM :
System.arraycopy() performs brilliantly on Intel/AMD CPUs as it maps right on down to a single REP MOVS instruction. Perhaps an order of magnitude faster than the same thing in a normal loop.
I'd like a System.arrayfill() method that could be built on a REP STOS operation.
(Today I'm initializing buffers for a b-tree. Need to blank-fill the text buffer - a pretty common thing to do. Index efficiency is critical to DB operations, so speed matters here.)
(Review ID: 180473)
======================================================================
- backported by
-
JDK-2199986 Optimize Arrays.fill(...)
-
- Resolved
-
-
-
- Resolved
-
-
-
- Resolved
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
- relates to
-
JDK-8307683 Loop Predication should not hoist range checks with trap on success projection by negating their condition
-
- Resolved
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-