-
Enhancement
-
Resolution: Fixed
-
P4
-
11, 16
-
b24
-
aarch64
-
linux
Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
---|---|---|---|---|---|---|
JDK-8263733 | 13.0.7 | Sergey Nazarkin | P4 | Resolved | Fixed | b04 |
JDK-8266006 | 11.0.12 | Dong Bo | P4 | Resolved | Fixed | b01 |
Base64.encodeBlock are not intrinsified on aarch64, which has been done on x86_64.
We can implement it with LD3/TBL instruction, a basic idea: http://0x80.pl/articles/base64-simd-neon.html#encoding-quadwords.
The performance improvements shown there are promising.
The usage of TBL instruction: https://developer.arm.com/documentation/dui0801/g/A64-SIMD-Vector-Instructions/TBL--vector-?lang=en.
We can implement it with LD3/TBL instruction, a basic idea: http://0x80.pl/articles/base64-simd-neon.html#encoding-quadwords.
The performance improvements shown there are promising.
The usage of TBL instruction: https://developer.arm.com/documentation/dui0801/g/A64-SIMD-Vector-Instructions/TBL--vector-?lang=en.
- backported by
-
JDK-8263733 AArch64: Implement Base64.encodeBlock accelerator/intrinsic
- Resolved
-
JDK-8266006 AArch64: Implement Base64.encodeBlock accelerator/intrinsic
- Resolved
- links to
-
Commit openjdk/jdk13u-dev/ebea17d0
-
Commit openjdk/jdk/8638cd9a
-
Review openjdk/jdk13u-dev/149
-
Review openjdk/jdk/992
(1 links to)