RFR: 8266951: Partial in-lining for vectorized mismatch operation using AVX512 masked instructions [v4]

[Jatin Bhateja]

On Mon, 24 May 2021 11:07:16 GMT, Vladimir Ivanov <vlivanov at openjdk.org> wrote:

>> Jatin Bhateja has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   8266951: Removing the changes to existing benchmark since a separate benchmark has been added to partial in-lining.
>
> src/hotspot/share/opto/library_call.cpp line 5225:
> 
>> 5223:   BasicType prim_types[] = {T_BYTE, T_SHORT, T_INT, T_LONG};
>> 5224:   BasicType vec_basictype = prim_types[scale_val];
>> 5225:   const Type* vec_type = Type::get_const_basic_type(vec_basictype);
> 
> It's not a vector type, but the element type. 
> 
> Also, should `VectorMaskGenNode` just accept element basic type instead?

VectorMaskGenNode is of LONG type, explicit lane element type associated with VectorMaskGenNode is used to convert Store/LoadVectorMasked to Store/LoadVector for all true masks. Existing node definition expects a Type instead of a BasicType. I can change the name to vec_lane_type if that's acceptable.

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

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