StringCoding.hasNegatives
Andrew Haley
aph-open at littlepinkcloud.com
Fri Oct 1 13:59:03 UTC 2021
On 10/1/21 1:57 PM, Brett Okken wrote:
> I know java.lang.StringCoding.hasNegatives has a
> HotSpotIntrinsicCandidate annotation/implementation, but is there
> interest/value in a faster pure java implementation?
>
> Using Unsafe to read and compare 8 bytes at a time as a long is faster
> than the current simple implementation both in interpreter only mode
> and default hotspot compiler mode. I can provide jmh bencmark and
> results if this is worthwhile to pursue.
The current pure Java implementation does two things: it provides a fallback
for pure-interpreter JVMs and it provides the reader with a simple implementation.
I'm not at all sure we'd want a complex implementation.
Having said that, if I were looking at a faster pure Java version of
this logic, I'd look at MethodHandles.byteArrayViewVarHandle().
--
Andrew Haley (he/him)
Java Platform Lead Engineer
Red Hat UK Ltd. <https://www.redhat.com>
https://keybase.io/andrewhaley
EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
More information about the core-libs-dev
mailing list