[
https://issues.apache.org/jira/browse/LUCENE-9636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Feng Guo updated LUCENE-9636:
-----------------------------
Description:
In decode6(), decode7(), decode14(), decode15(), decode24() longs always `&` a
same mask and do some shift. By printing assemble language, i find that JIT did
not optimize them with SIMD instructions. But when we extract all `&`
operations and do them first, JIT will use SIMD optimize on them.
was:
In `decode6()` `decode7()` `decode14()` `decode15()` `decode24`, longs always
`&` a same mask and do some shift. By printing assemble language, i find that
JIT did not optimize them with SIMD instructions. But when we extract all `&`
operations and do them first, JIT will use SIMD optimize on them.
> Exact and operation to get a SIMD optimize
> ------------------------------------------
>
> Key: LUCENE-9636
> URL: https://issues.apache.org/jira/browse/LUCENE-9636
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/codecs
> Reporter: Feng Guo
> Priority: Trivial
> Time Spent: 10m
> Remaining Estimate: 0h
>
> In decode6(), decode7(), decode14(), decode15(), decode24() longs always `&`
> a same mask and do some shift. By printing assemble language, i find that JIT
> did not optimize them with SIMD instructions. But when we extract all `&`
> operations and do them first, JIT will use SIMD optimize on them.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
