RFR: 8266951: Partial in-lining for vectorized mismatch operation using AVX512 masked instructions [v6]

Jatin Bhateja jbhateja at openjdk.java.net
Sun May 30 18:37:13 UTC 2021


On Fri, 28 May 2021 09:41:06 GMT, Andrew Haley <aph at openjdk.org> wrote:

>> Jatin Bhateja has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   8266951: Review comments resolution.
>
> src/hotspot/cpu/x86/x86.ad line 8074:
> 
>> 8072:   %}
>> 8073:   ins_pipe( pipe_slow );
>> 8074: %}
> 
> The semantics of `VectorCmpMasked` seem to be defined nowhere. It doesn't much matter where you put the comment, perhaps on the definition of the node itself, but a reader shouldn't have to study the code to figure out what `VectorCmpMasked` is supposed to do.

Hi @theRealAph , name VectorCmpMaskedNode is on the lines of other masked operations LoadVectorMaskedNode/StoreVectorMaskedNode.  Nomenclature used implicitly describe the semantics i.e. <Vector Operation>MaskedNode.  Appropriate comments appended.

-------------

PR: https://git.openjdk.java.net/jdk/pull/3999


More information about the hotspot-compiler-dev mailing list