RFR: 8318650: Optimized subword gather for x86 targets. [v3]

[Jatin Bhateja]

On Fri, 3 Nov 2023 00:22:55 GMT, Sandhya Viswanathan <sviswanathan at openjdk.org> wrote:

>> Jatin Bhateja has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Restricting masked sub-word gather to AVX512 target to align with integral gather support.
>
> src/hotspot/share/opto/vectornode.hpp line 886:
> 
>> 884:     init_class_id(Class_LoadVectorGather);
>> 885:     add_req(indices);
>> 886:     assert(req() == MemNode::ValueIn + 1, "match_edge expects that index input is in MemNode::ValueIn");
> 
> The assert message in "" is questionable. For subword types match_edge expects last input in MemNode::ValueIn+1. I think the assert needs to move in if/else below with correct message.

Its placed appropriately before adding offset edge, I have added other assertion to check for offset.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/16354#discussion_r1382574475