RFR: 8355177: Speed up StringBuilder::append(char[]) via Unsafe::copyMemory [v8]
Raffaello Giulietti
rgiulietti at openjdk.org
Thu Jul 24 13:32:58 UTC 2025
On Fri, 27 Jun 2025 01:04:32 GMT, Shaojin Wen <swen at openjdk.org> wrote:
>> In BufferedReader.readLine and other similar scenarios, we need to use StringBuilder.append(char[]) to build the string.
>>
>> For these scenarios, we can Unsafe.copyMemory instead of the character copy of the char-by-char loop to improve the speed.
>>
>> @RogerRiggs completed the optimization when the encoder is LATIN1 in PR #24967. This PR continues to complete the optimization when the encoder is UTF16.
>
> Shaojin Wen has updated the pull request incrementally with one additional commit since the last revision:
>
> Update src/java.base/share/classes/java/lang/StringUTF16.java
>
> Co-authored-by: Chen Liang <liach at openjdk.org>
src/java.base/share/classes/java/lang/StringUTF16.java line 1490:
> 1488: val,
> 1489: Unsafe.ARRAY_BYTE_BASE_OFFSET + ((long) index << 1),
> 1490: (long) (end - off) << 1);
The documentation of `copyMemory()` is not super-clear about endianness.
But it seems to imply that in this case it behaves as if it were to copy `short`s, so endianness seems to be preserved.
The invocation of `copyMemory()` here implicitly assumes that `ARRAY_CHAR_INDEX_SCALE` and `ARRAY_BYTE_INDEX_SCALE` are 2 and 1, resp., which seems quite reasonable but not written in the stone.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/24773#discussion_r2228500325
More information about the core-libs-dev
mailing list